This item is only available as the following downloads:
i PRE RESTORATION ECOLOGICAL ASSESSMENT OF BOWMAN CREEK, AN URBAN TRIBUTARY OF THE ST. JOSEPH RIVER IN SOUTH BEND, INDIANA BY SARAH A. MCMANUS A Thesis Submitted to the Division of Natural Sciences New College of Florida in partial fulfi llment of the requirements for the degree Bachelor of Arts Under the sponsorship of Dr. Elzie McCord, Jr. May 2011 Sarasota, FL
ii Acknowledgements Thanks to: Dr. McCord, for sponsoring this thesis, and Dr. Hart and Dr. Beulig, for being on my ba ccalaureate committee; Patrick Shirey and Dr. Gary Lamberti, for w orking with me on this project, Mike Brueseke, for helping me in the field and lab; Dr. Chaloner, Peter Levi Dave, Janine, Mia, Sarah, and Ursula for tackling my endless questions and frust ra tions regarding fancy equipment; Shaun, for showing me the poetry of South Bend; Bob and Betty Urbanski, for the wonderful food and conversations ; E llen for defending her stretch of the stream ; and NSF for supporting this research via IGERT training gr ant #050449. I am also deeply grateful to everyone else who has encouraged my interest in stream ecology, especially my parents, who let me play in my grandfather's excellent backyard creek to my heart's content and didn't mind too much when I filled all the tupperware with crayfish. Thanks also to Matt Landon, Jason Robinson and all the fine folks at United Mountain Defense for two summers of counting bugs to stop the blasting. Finally, thank you to all my friends and classmates at New College who've gon e swamp stomping, pond slogging and bugpicking with me! Especially Jesse Wheeler and Trevor Caughlin your enthusiasm for ecology and dedication to scientific rigor got me off to a great start at New College, and I can only hope to pass that on to others
iii The authors first recorded stream survey, circa 1995 (age 6)
iv Page 2 of the author's first recorded stream survey (1995). Observations were taken from a 70m section of an urban tributary to Hanks Branch, in the Lower Yadkin watershed (Thomasville, NC, centered on 3553'46"N, 80 5'5"W )
v Table of Contents List of Abbreviations .................................................................................................... vii List of Figures ............................................................................................................. viii List of Tables ................................................................................................................. xi Abstract ........................................................................................................................ x i i Chapter 1: Introduction ................................................................................................1 1.1 Stream ecology in a landscape context ........................................................1 1.2 Stream ecology in a human context .............................................................5 1.2.1 Streams and urban environments .............................................................8 1.2.2 The urban stream syndrome .................................................................. 11 1.2.3 Culverts and buried streams .................................................................. 21 1.2.4 Urban stream restoration and daylighting buried streams ................... 22 1.2.5 Restoration and ecosystem health: community definitions and goals ..... 28 1.3 Stream ecology in a scientific context ........................................................ 33 1.3.1 Models and limitations .......................................................................... 33 1.3.2 Nutrient and energy flow in stream ecosystems ..................................... 35 1.3.3 The river continuum concept and stream metabolism ............................ 47 1.4 Diversity ind ices and monitoring regimes ................................................. 56 1.4.1 Overview of benthic macroinvertebrate biomonitoring .......................... 56 1.4.2 What are benthic macroinverteb rates? ................................................... 57 1.4.3 Why are benthic macroinvertebrates used to measure water quality? ..... 58 1.4.4 How are benthic macroinvertebrates us ed to measure water quality? ..... 61 1.4.5 How are the data collected? ................................................................... 65 1.4.6 How are the invertebrates identified? .................................................... 71 1.4.7 How are the data analyzed? ................................................................... 76 1.4.8 How are the data used? ......................................................................... 79 1.4.9 Be nthic macroinvertebrates in urban streams ........................................ 80
vi 1.4.10 Summary ........................................................................................... 83 Chapter 2: Project design ............................................................................................ 84 2.1.1 Overview .............................................................................................. 84 2.1.2 Site history and descriptions .................................................................. 84 2.1.3 Project goals, questions, and hypothe ses ............................................... 90 2.1.4 Timeline ............................................................................................... 93 2.1.5 Study sites ............................................................................................. 94 Chapter 3: Materia ls and methods ............................................................................. 96 3.1 Field protocols ............................................................................................ 96 3.2 Lab materials and protocols ...................................................................... 99 Chapter 4: Results ..................................................................................................... 105 4.1 Physical data ............................................................................................ 105 4.1.1 Equipment issues ................................................................................ 105 4.1.2 Sonde data by site ............................................................................... 109 4.2 Chemical data .......................................................................................... 115 4.2.1 Nutrient data. ...................................................................................... 116 4.2.2 Dissolved Organic Carbon data ........................................................... 121 4.3 Biological data .......................................................................................... 122 4.3.1 Biofilm: chlorophyll a den sity and ashfree dry mass .......................... 122 4.3.2 Benthic macroinvertebrate community analysis ................................... 125 Chapter 5: Conclusions and discussion ..................................................................... 136 5.1 Discussion of results and potential sources of error ................................ 136 5.2 Future research directions ....................................................................... 142 5.3 Implications for restoration ..................................................................... 144 Chapter 6: Works Cited ............................................................................................ 145 Chapter 7: Appendices .............................................................................................. 160 7.1 Appendix I: Detailed study site information ........................................... 160 7.2 Appendix II: Invertebrate counts ............................................................ 172
vi i List of A bbreviations AFDM : Ash Free Dry Mass AI : Autotrophic Index CPOM : Coarse Particulate Organic Matter CR: Community Respiration DAYTON: Sample site at the Dayton St. crossing DO : Dissolved Oxygen DOC: Dissolved Organic Carbon EI : Effective Imperviousness EPA: United States Environmental Protection Agency FBI: Family Biotic Index FFG : Functional Feeding Group FHG : Functional Habitat Group FPOM: Fine Particulate Organic Mattter GOLF : Sample site at the Studebaker Golf Course GPP: Gross Primary Productivity LDO : Luminescent Dissolved Oxygen a method of measuring dissolved oxygen LOCST: Sample site at the Locust Rd. crossing LWE : Sample site near the Lincolnway East crossing MAIN : Sample site at the Main St. crossing NIPSCO : Northern Indiana Public Service Company a facility near the LWE site P/R: Ratio of photosynthesis to respiration PAR : Photosynthetically Active Radiation RAV: Sample site at Ravina Park RRCULV: Sample site at the railroad crossing culvert near Lincolnway East SRP : Soluble Reactive Phosphorus T I: Total Imperviousness URB : Sample site behind the Urbanskis house USACE: United States Army Corps of Engineers USGS: United States Geological Survey
viii List of Figures Figure 1.1: Lower Mississippi River, early stream channels at half century intervals. Army Corps of Engineers (1938), from http://lmvmapping.erdc.usace.army.mil/ ..3 Figure 1.2: Hierarchical organization of stream habitats, illustrating how smaller units are repeating elements nested within larger units (Frissell et al. 1986) ................4 Figure 1.3: Bridge abutment scour. USGS, from http://pubs.usgs.gov/fs/FS 02496/ .......9 Figure 1.4: Fat buildup in London sewers. Thames Water 2010, from http://www.thameswater.co.uk/cps/rde/xchg/corp/hs.xsl/9092.htm ..................... 10 Figure 1.5: Atherosclerosis in a human artery, from http://www.healthtree.com/articles/coronary artery disease/anatomy/ ................ 11 Figure 1.6: Conceptual model of the urban stream syndrome, from Wenger et al. 2009. EI = Effective imperviousness ............................................. 14 Figure 1.7: Effects of urbanization (e.g. stream channel burial and piping) on catchment hydrology, from Roy et al. (2009a) ............................................... 16 Figure 1.8: MAIN site on Bowman Creek, a) on 7/14/10 before the storm and b) on 7/16/10 after the strom, showing flood debris accumulation ................ 19 Figure 1.9: Channel alteration, trash buildup and debris dam at LWE site on Bowman Creek ........................................................................... 20 Figure 1.10: Inappropriately overengineered restoration, from Palmer (2008) ................ 24 Figure 1.11: Nutrient spiraling, from Minshall et al. (1983) ........................................... 37 Figure 1.12: Another representation of nutrient spiraling, from Allan and Casti llo (2007) ........................................................................... 38 Figure 1.13: Energy flow through a stream ecosystem, from Odum (1956) .................... 41 Figure 1.14: The river continuum concept, from Vannote et al. (1980) ........................... 48 Figure 1.15: Diel respiration curve, showing GPP and CR (from Marzolf et al. 1994) .... 52 Figure 1.16: Dissolved oxygen tolerance (from http://www.water research.net/Watershed/dissolvedoxygen.htm) .............. 53 Figure 2.1: a) Juday Creek (above) and Bowman Creek (below) are tributaries in b) the St. J oseph River watershed. Maps from Google Earth and USGS ............. 86 Figure 2.2: Combined sewer outflow, from City of South Bend (2010) .......................... 88
ix Figure 3.1: Hach Hydrolab equipment and accessories. .................................................. 96 Figure 3.2: Odyssey photosynthetic irradience loggers and mounting setup. .................. 97 Figure 3.3: Collecting macroinvertebrates with a Surber sampler. .................................. 99 Figure 3.4: The Lachat QuikChem 8500, in all its great and terrible glory ................. 100 Figure 3.5: Vacuum manifold with filter cups .............................................................. 101 Figure 3.6: Chlorophyll a extracted from biofilm samples (shown after analysis) ......... 102 Figure 3.7: Chlorophyll a analysis using flurometer ..................................................... 102 Figure 3.8: Processing invertebrate samples ................................................................. 104 Figure 3.9: Invertebrate sample in sorting tray ............................................................. 104 Figure 4.1: Promising results from the first test. LDO = Luminescent Dissolved Oxygen ........................................................... 107 Figure 4.2: Further testing revealed that the promising results were an anomaly caused by equipment malfunction .................................................. 108 Figure 4.3: MAIN and DAYTON LDO, July 14 July 16 ............................................ 111 Figure 4.4: MAIN and DAYTON temperature, July 14 16 ......................................... 112 Figure 4.5: LOCST, URB and GOLF LDO, July 30 Aug 2 ........................................ 113 Figure 4.6: LOCST, URB and GOLF Temperature, July 30 Aug 2 ............................. 113 Figure 4.7: LOCST, URB and GOLF pH, July 30 Aug 2 ............................................ 114 Figure 4.8: Photosynthetically Active Radiation, in microEinsteins per second per square meter ............................................................................................... 115 Figure 4.9: Ammonium in Bowman Creek ................................................................... 117 Figure 4.10: Nitrate in Bowman Creek ......................................................................... 118 Figure 4.11: Soluble Reactive Phosphorus in Bowman Creek ...................................... 119 Figure 4.12: Soluble Reactive Phosphorus in Bowman Creek, showing outliers ........... 120 Figure 4.13: Soluble Reactive Phosphorus in Bowman Creek, outliers removed .......... 120 Figure 4.14: Dissolved Organic Carbon in Bowman Creek .......................................... 121 Figure 4.15: Chlorophyll a density in Bowman Creek .................................................. 123 Figure 4.16: Ash Free Dry Mass in Bowman Creek biofilm ......................................... 123 Figure 4.17: Autotrophic Index in Bowman Creek Biofilm (normal range is from Steinman et al. 2007) ............................................................................ 124
x Figure 4.18: Invertebrate abundance ............................................................................ 125 Figure 4.19: Total invertebrate taxa richness per site .................................................... 126 Figure 4.20: a) Hydropsyche and b) Cheumatopsyche (Trichoptera: Hydropsychidae) .. 128 Figure 4.21: a) Black fly larva (Diptera: Simuliidae) b) Peltodytes lengi (Coleoptera: Haliplidae) ..................................................... 128 Figure 4.22: Dryopidae: a) Helichus striatus ; Elmidae: b) Optioservus fastiditus ......... 129 Figure 4.23: Sperchopsis tesselatus (Coleoptera: Hy drophilidae) ................................. 129 Figure 4.24: Comparisons of the relative abundance of different functional feeding groups ................................................................................. 131 Figure 4.25: P/R Ratio, based on functional feeding group analysis ............................. 132 Figure 4.26: CPOM/FPOM Ratio, based on functional feeding group analysis ............. 133 Figure 4.27: Family Biotic Index ................................................................................. 135 Figure 7.1: a) LOCST site, a ditch next to a construction area. b) Map from Google Earth ............................................................................... 160 Figure 7.2: URB site, in a wooded area. b) Map from Google Earth ............................. 162 Figure 7.3: a) MAIN site, above the first major culvert. b) Map from Google Earth ..... 164 Figure 7.4: a) GOLF site, a golf course between two culverts. b) Map from Google Earth. .............................................................................. 165 Figure 7.5: a) DAYTON site, the outflow of a long culvert. b) Ma p from Google Earth ............................................................................... 166 Figure 7.6: a) RAV site, in Ravina Park. b) Map from Google Earth ............................ 168 Figure 7.7: a) NIPSCO culvert outf low. b) Map from Google Earth. c) LWE site ......... 170
xi List of Tables Table 1.1: Functional Feeding Groups, from Merritt et al. (2002) .................................. 74 Table 1.2: Functional Habitat Groups, from Merrit et al. (2001) ..................................... 75 Table 1.3: Traits from USGS meta database (Poff et al. 2006) ........................................ 76 Table 1.4: Ecological metrics for benthic macroinvertebrates ......................................... 78 Table 2.1: Site information ............................................................................................. 95 Table 4.1: Water quality values based on Family Biotic Index. Values from Hilsenhoff (1988); method reported in Carter et al. (2007) ....................... 134 Table 4.2: Water quality values based on mean Fam ily Biotic Index (FBI). SEM = standard error of the mean .................................................................... 135 Table 7.1: Raw macroinvertebrate data for sites LOCST, URB, and MAIN .................. 172 Table 7.2: Raw macroinvertebrate data for sites GOLF, RAV, and LWE ....................... 174
xii PRE RESTORATION ECOLOGICAL ASSESSMENT OF BOWMAN CREEK, AN URBAN TRIBUTARY OF THE ST. JOSEPH RIVER IN SOUTH BEND, INDIANA Sarah A. McManus New College of Florida, 2011 ABSTRACT Bowman Creek is one of the most ecologically impaired tributaries of the St. Joseph River, a large river in the Lake Michigan drainage system It has been rerouted underground through pipes and culverts for more than half of its downstream length. T he C ity of South Bend is planning to restore portions of the stream to a more natural aboveground condition, with goals of improving hydrological, ecological, education and aesthetic values. In this study I i nvestigated the effects of culverts on the physical, chemical and biologica l characteristics of the stream ecosystem Deployable dataloggers were used to characterize diel respiration curves. W ater samples were analyzed for ammonium, nitrate, soluble react ive phosphorus and dissolved organic carbon B iofilm samples were analyzed for chlorophyll a concentra tion and ashfree dry mass. B enthic macroinvertebrate samples were used to calculate indices of biotic integrity. The effect of culverts on stream ecosyst em function could not be determined from this study However, collected baseline data can be used to assess the progress of restoration efforts. Further
xiii research and continued long term monitoring is needed to characterize the dynamics of underground urban streams and evaluate the restoration of Bowman Creek. ______________________________ Dr. Elzie McCord, Jr. Division of Natural Sciences
xiv As for men, those myriad little detached ponds with their own swarming cor puscular life, what were they but a way that water has of going about beyond the reach of rivers? Loren Eiseley We are but whirlpools in a river of ever flowing water. We are not stuff that abides, but patte rns that perpetuate themselves. Norbert Wiener
1 Chapter 1: Introduction 1.1 Stream ecology in a landscape context The course of flowing water is shaped by the s urrounding terrain, and in turn, water shapes the terrain through which it flows (Stanford 2007) Lotic systems such as creeks, streams, and rivers form an integral part of the hydrologic cycle, connecting uplands to lowlands through the net aboveground movement of water. This cycle starts w hen water absorbs thermal (kinetic) energy from the sun and evaporates. The warm we t air rises from the sea, lakes, and soil, joining the water vapor exhaled and t ranspired by animals and plants. E ventually water condenses and falls onto the land as pre cipitation. As the water falls, some of its kinetic energy of motion acts directly on the surface of the land. Plants can absorb part of this force, absorbing water like a sponge and holding the soil with their roots and mycorrhizal associates but bare earth may be displaced downhill. Why downhill? It appears obvious, but it is important to know that this water has not lost all of the energy it a bsorbed from the sun on its way up into the atmosphere. If water falls on any land above sea level, it still has pote ntial energy of position. Differences in elevation create an energy gradient (Ba ll 2009) and it is this gradient from high to low energy which creates the flow of water so central to stream ecology. Streams are thermodynamically open systems dissipative dynamical systems in two important ways. First, they represent a flow of me chanical energy. The force of moving water, pulled downhill by gravity, splits boulders into rocks, grinds pebbles to sand, and reshapes the land by eroding away mountains and carving canyons as it carries sediment towards the sea (Stanford 2007) Streams also catalyze another form of
2 thermodynamic dissipation: the transport and breakdown of chemical energy in the form of leaves, woody debris and other nutrient sources that are washed from the land into the water (Vannote et al. 1980) These outside, or al lochthonous, sources of organic matter may dwarf the in stream, or autochthonous, fixation of carbon by aquatic plants and algae, especially for small streams in forested watersheds (Benfield 2007) Streams transport and transform matter and energy across a vast range of spatial and temporal scales. Moving water shapes and reshapes the land over geologic time, wearing down the highlands and redistributing stones and sediment (Howard, Dietrich, and Seidl 1994) Some river systems experience yearly flood ing due to snowmelt or monsoon rains, depositing nutrient laden mud onto the banks of the floodplains downstream. Over shorter intervals, large amounts of organic matter are constantly being broken down and reassimilated by aquatic communities. For example t he nutrients and chemical energy contained in a leaf that falls into a mountain stream may pass through a vast array of organisms, from bacterial biofilms to flat headed scraper mayflies to hel lgrammites to trout, spreading out through the food web. The flow of nutrients shapes communities and is shaped by them Vannote et al. (1980) developed the river continuum concept to explain this process, stating that : Downstream communities are fashioned to capitalize on upstream processing inefficiencies. F lowing water can also transform the shape of its own path. Rivers change their course over time a phenomenon which can be traced using satellite imagery and other techniques as shown in Fig. 1.1 (USACE 1938) The hydrography of a particular section of a river or stream (its physical characteristics such as width, depth, and flow) can also
3 change from day to day and hour to hour in response to rainfall events or other changes (Baker et al. 2004) As Heraclitus said, You cannot step twice into the same riv er, for other waters are continually flowing in (Sedley 2003) Poff et al. (1997) discuss the importance of understanding consistent yearly patterns in flow variation (the natural flow regime) in the context of long term monitoring and restoration proje cts. Figure 1. 1: Lower Mississippi River, early stream channels at half century intervals. Army Corps of Engineers (1938) from http://lmvmapping.erdc.usace.army.mil/ The physical characteristics of str eams also change longitudinally Over large spatial scales, mountain headwaters combine to form piedmont rivers, which then flow
4 through undulating coastal floodplain l andscapes on the way to the sea. O n shorter scales many smaller streams are characterized by riffle run pool sequences (Stanford 2007) Riffles are shallow, fast moving, rocky areas that exhibit a relatively quick drop in elevation compared to runs, which are straighter, deeper sections of flowing water, while pools are areas of slower movi ng or nearly still water (Jowett 1993) Sediment is generally eroded from riffles, transported through runs, and deposited in areas of slower moving current, such as the inside bank of curved stream sections and the bottom of pools. Streams can be understo od as nested hierarchical systems of habitats and microhabitats (Fig. 1.2) each supporting different assemblages of species that are adapted to the characteristic nutrient dynamics and flow regimes of these areas (Frissell et al. 1986) Figure 1. 2: Hierarchical organization of stream habitats, illustrating how smaller units are repeating elements nested within larger units ( Frissell et al. 1986)
5 1.2 Stream ecology in a human context Human settlements also influence the shape of streams and streams influence patterns of human settlement Many of the earliest civilizatio ns developed along river floodplains, as these sites provide readily accessible fresh water food, and fertile soil. Rivers can also act as transportation corridors for boats and barges, facilitating the spread of settlement inland along these ready made routes. Human communities alter the characteristics of the land surrounding the stream, changing the composition of the solutes carried by th e water and the paths by which it flows I ncreased agriculture in the drainage basin (the area of land drained by a particular stream, also called a catchment basin or watershed) may increase sediment runoff due to erosion from plowed fields. Fertilizers o r pesticides may lea ch into the stream water altering the nutrient load and potentially adversely affecting biological communities. The removal of riparian vegetation (trees and plants growing along river banks) reduces the buffer between terrestrial and aquatic ecosystems, allowing more stormwater runoff (along with any suspended sediments and chemicals) to flow directly into streams (Groffman et al. 2003) C onstruction of paved roads, parking lots, and buildings in the watershed also reduces the amount o f stormwater that is absorbed into the ground, often causing flooding. I mpervious surfaces do little to slow the flow of runoff, increasing the force that these waters exert on hillsides and riverbanks an d leading to increased erosion. Rushing stormwater, channeled into streams through pipes and culverts, can cause scouring flows and flash flood s that tear through aquatic ecosystems, disrupting habitat and sweeping organic matter and organisms downstream (see Section 1.2.2).
6 T hese constant disturbances reduce the streams ability to process nutrients effectively. Due to overapplication of fertilizers, mishandling of wastewater, and impairment of stream ecosystems, the unused excess of nutrients builds and builds, growing with each juncture as creeks join rivers and empty into receiving waters such as oceans or lakes. There, the overabundance of nitrates and phosphates can feed vast unsustainable algae blooms, leading to eutrophication. Algal respir ation and decomposition deplet e the available oxygen in the water, creating dead zones that are unsuitable for fish and other oxygen breathing organisms. Many of the worlds inhabited coastlines have begun experiencing seasonal or persistent anoxic conditions and the affected areas have spread exponentially since the 1960s (Diaz and Rosenberg 2008) In the United States, the largest and most well known dead zones occur in the Gulf of Mexico and t he Chesapeake Bay resulting in fish kills and adverse effects on fisheries (Diaz and R osenberg 2008) Recent studies have highlighted the urgent need to reexamine our role in altering global nutrient dynamics, especially the phosphorus cycle. Through phosphate mining, fertilizer application, wastewater discharge, and stream ecosystem alterations, we have drastically changed the way phosphorus moves from the continents to the sea and back again (Rockstrm et al. 2009) According to Carpenter and Bennett (2011) c urrent conditions exceed all planetary boundaries or upper tolerable limits, for phosphate discharge to the oceans. These authors consider ocean eutrophication and potential phosphate shortages to be serious problems, requiring a definite shift towards greater recycling of phosphorus thorough th e human food system. Ozaki et al. (2 011) present an
7 even grimmer possibility. Based on models of prehistoric ocean biogeochemical cycles they conclude that elevated riverine flux of reactive [phosphorus] is the most important factor for triggering global anoxic events via a positive feedba ck loop among ocean anoxia, phosphorus regeneration, and surface biological productivity. These global anoxic events have been associated with mass extinctions throughout Earths history ( in conjunction with high atmospheric CO2 levels, elevated temperatu res, decreased oxygen solubility, and alterations in thermohaline circulation or the ocean conveyor belt) (Meyer and Kump 2008) These ancient anoxic periods also led to the accumulation of sediments rich in phosphates and organic matter, forming some of the very rock layers from which we are now mining and extracting phosphates and fossil fuels (Ingall and Jahnke 1997) Carpenter and Bennett (2011) do not consider the oceans to be at risk of widespread anoxia in the foreseeable future, because we may n ot be able to sustain our unsustainable nutrient inputs over a long enough timescale to cause a catastrophic shift in total oceanic phosphorus concentrations. We may run out of easily extractable phosphate resources first, according to Filippelli (2008) R ockstrm et al. (2009) propose setting certain maximum boundaries for nutrient runoff and other potentially destabilizing human inputs, and urge caution due to uncertainties in modeling. Research into tipping points and the complex dynamics of humanecos ystem interactions is ongoing (Walker and Meyers 2004) but the good news is that these changes can go either way For example, a massive hypoxic zone in the Black Sea resolved itself after the collapse of the Soviet Union due to a loss of fertilizer subs idies and a corresponding sharp decrease in
8 nutrient loading from the surrounding regions (Diaz and Rosenberg 2008) Resilience in the face of unprecedented environmental challenges will undoubtedly require flexibility, innovation, and initiative in both t op down (international, governmental, regulatory) and bottom up (local, grassroots, self organized) responses. Even something so small as an examination of ecosystem function in urban streams may contribute to the search for solutions on a larger scale. As Richard Feynman (1967) so eloquently put it, Nature uses only the longest threads to weave her patterns, so that each small piece of her fabric reveals the organization of the entire tapestry. 1.2.1 Streams and urban environments As described in Section 1.1 streams are naturally dynamic and constantly changing. Conflicts arise when human communities and hu man made structures on riverbanks and floodplains are not as adaptable. Figure 1.3 shows a dangerously unstable bridge. T he surrounding soil has been scour ed away, exposing the road to potential c ollapse as a result of undermining erosion .
9 Figure 1. 3: Bridge abutment scour. USGS, f rom http://pubs.usgs.gov/fs/F S 02496/ It may appear that an obvious solution would be to avoid constructing permanent settlements or hardened structures on floodplains or riverbanks but human communities have developed around order and structures In industrialized nations, acces s to fresh water is considered so ubiquitous that it goes unnoticed by the vast majority of the population, at least until the water runs out or becomes contaminated (Carroll 2008) Streams are l ike the veins of the continents and F irst World nations have constructed an incredibly branched vascular system to meet their water needs. C apillary network s of pipes perfuse nearly every subdivision and home, bringing fresh water to every bathroom and kitchen and channeling wastewater into septic systems and munic ipal sewers. However, as a city expands, its water requirements soar and ever more flow is diverted from surrounding surface and groundwater reserves through systems of canals and culverts. This process is somewhat reminiscent of tumor angiogenesis, in whi ch blood
10 flow is rerouted and disproportionately concentrated in a specific region. To extend the cardiovascular metaphor, urban sewer systems can sometimes manifest conditions similar to atherosclerosis. If waste fats from commercial and residential gray water make their way into the drains, these fatty plaques can clog pipes as well as human arteries. B lockage s can create serious problems as seen in Figure s 1.4 and 1.5: Figure 1. 4: Fat buildup in Londo n sewers. Thames Water 2010, from http://www.thameswater.co.uk/cps/rde/xchg/corp/hs.xsl/9092.htm
11 Figure 1. 5: Atherosclerosis in a human artery from http://www.healthtree.com/articles/coronary artery diseas e/anatomy/ 1.2.2 The urban stream syndrome U rban planners and civil engineers often respond to community concerns regarding flooding and erosion with straightforward tactics that emphasize the efficient delivery of results. If there is too much water, make it drain faster. If the riverbank is crumbling, patch it or pave it. In many cases, municipalities opt for a twoin one solution: channelization of streams and hardening of streambeds. This process can involve straightening and deepening stream channel s lin ing channel s and banks with concrete or riprap (rock or rubble), or rerouting the stream entirely through a culvert or pipe. Unfortunately, while these measures allow water to drain faster from roads and parking lots, they often result in increased hydrolo gical problems downstream (such as flash floods and scouring) as well as decreased habitat value and impaired ecosystem function (Walsh et al. 2005b ) The science of urban stream ecology is an emerging and rapidly developing field. The first major revie w paper on this subject was published by Paul and Meyer (2001)
12 The 1st Symposium on Urban Stream Ecology (SUSE) was held in Melbourne, Australia in 2003, and some of the research presented at the conference was published in Vol. 24, Issue 3 of the Journal of the North American Benthological Society (J NABS). Several important synthesis papers followed in the next two years. Meyer et al. (2005) described recent investigations into urban ecosystem function, and Walsh et al. (2005b ) introduced the term urban stream syndrome to describe common characteristics of stream responses to urbanization. The 2nd Symposium on Urban Stream Ecology was held in Salt Lake City, Utah in 2008, bringing together 118 attendees from 8 countries to discuss and compare 30 resear ch presentations. Papers from this conference were published in J NABS 28(3) and are summarized in the introduction to the issue by Roy et al. (2009b ) In this issue, Wenger et al. (2009) discuss ed Twenty six key research questions in urban stream ecology and describe d a conceptual model for understanding the complex interactions between interconnected stressors and response variables in this system. Carter et al. (2009) also offer ed a very useful analysis of the experimental designs and analytical method s used in previous urban stream ecosystem research and suggest ed techniques and directions for further study. In the same year (though not in the same publication), Brown et al. (2009) presented the results of an impressive large scale study of stream resp onses to urban gradients in nine metropolitan areas across the U.S. conducted by the USGS as part of the National Water Quality Assessment (NAWQA) program. Previous efforts to synthesize generalizable conclusions from research conducted by various teams had been hampered by differences in study design s and methods. Brown et al. (2009) note d that the variability in stream ecosystem response to
13 urbanization between these prior studies "probably is partially the result of real differences in stream process es among environmental settings. The USGS project aimed to address these difficulties by conducting a standardized, coordinated sampling and analysis program. A list of further USGS NAWQA publications on urban streams is available at: http://water.usgs.gov/nawqa/urban/html/publications.html As described by Walsh et al. (2005b ) the s ymptoms of the urban stream syndrome include a flashier hydrograph, elevated concentrations of nutrients and contaminants, altered channel morphology, and reduced biotic richn ess, with increased dominance of tolerant species Wenger et al. (2009) provide d a conceptual model for addressing the most prominent linkages between certain management decisions and top level effects on stream function, although this simplified graphica l representation does not show interactions between the biotic com ponents of the stream ecosystem (Fig. 1.6) Another conceptual model for understanding the urban stream syndrome (and other stressors affecting freshwater lotic systems) is the Causal Analy sis/Diagnosis Decision Information System (CADDIS) developed by the U.S. Environmental Protection Agency (U.S. EPA 2010) Named for the pollution sensitive invertebrate taxa (caddisflies), this online resource provides a step by step diagnostic framework f or evaluating stream impairments and choosing appropriate monitoring and management responses. Extensive documentation and links to further information are included throughout.
14 Figure 1. 6: Conceptual mo del of the urban stream syndrome, from Wenger et al. 2009. EI = Effective imperviousness
15 Flashy hydrography is the first characteristic that Walsh et al. (2005 b ) used to describe urban streams, and this single parameter is of great importance in understanding these systems. Hydrographic changes integrate a complex assemblage of stressor factors and are also critical in influencing habitat characteristics and ecosystem responses. The term flashy hydrography refers to the high variability of water velocit y and transport volume that often result s from increased direct stormwater input to streams (Baker et al. 2004) In extreme cases, this can result in flash flooding (hence the name). If the stream discharge (the volume per cross sectional area per unit t ime) is plotted against time, the resulting graph (a hydrograph) will show sharp spikes after rain events, representing bursts of very high flows that force large amounts of water down the stream channel in a short time, putting g reat stress on the syste m (Baker et al. 2004) Streams that flow through more natural forested catchments receive rainfall more gradually because much of the water soaks into the land and percolates into the stream through underground seepage (Walsh et al. 2009) This process results in a slower increase, lower peak, and longer tapering off of higher flows following precipitation events (Walsh et al. 2009) Urbanization changes the way water flows th rough the entire drainage basin (Roy et al. 2009a ) As shown in Fig. 1.7, s tr eam burial and direct connectivity to stormwater systems can lead to loss of ephemeral and intermittent stream habitats consolidating multiple small catchments into fewer, larger units (Roy et al. 2009a ). H ydraulically effective stormwater connections i ncrease the percent effective impervious cover of the area drained by the stream, a metric which has been found to be more useful in predicting detrimental urban impacts than % Total Impervious Cover alone (Walsh et al. 2009)
16 Figure 1. 7: Effects of urbanization (e.g. stream channel burial and piping) on catchment hydrology, from Roy et al. (2009a ) Water that is drained from impervious surfaces such as buildings, roads, and parking lots and channeled into p ipes bypasses the filtration function of terrestrial vegetation and riparian buffers, allowing no chance for nutrient uptake or retenti on of heavy metals, pesticides, or other toxins befo re the water enters the stream (Walsh 2004) The water and nutrients are not available for use by terrestrial plants and ecosystems, and the flow of runoff is not slowed in the stormwater pipes. Fast moving stormwater causes scouring flows in connected streams, dislodging any movable objects in the stream channel (sediment, woody debris, leafpacks, benthic algae, macroinvertebrates and other organisms) and flushing them downstream (Borchardt and Statzner 1990; Borchardt 1993) A certain amount of disturbance and scouring is a natural component of stream ecosystems, reducing embeddedness and improving terrestrial soil fertility in floodplains downstream (Resh et al. 1988; Allan and Castillo 2007) However increased frequency
17 and magnitude of disruption result in less effective retention of the woody debris and organic matter that form the basis of many stream habitats and food webs, leading to impaired ecosystem function (Aldridge et al. 2009) If stream microhabitats and attachment sites for invertebrates and algae are constantly disturbed, in stream processing and uptake of nutrients is decreased and trophic complexity falls (meaning the stream cannot support as ma ny fish and other vertebrates) (Helms 2008) Direct stormwater inputs and habitat disruption lower the ability/capacity of the stream to utilize and assimilate n utrients (Walsh et al. 2005a ) Excess nutrient levels in Midwestern streams are already some of the highest in the nation (Caskey et al. 2010) and whatever stream ecosystems cannot process is simply washed further downstream (see Sections 1.3.2 and 1.3.3) As headwater streams become more and more impaired and less able to use these nutrients in situ ever increasing concentrations are exported and the mismatch between processing capacity and nutrie nt load grows at each river junction. By the time the largest rivers empty into receiving waters such as lakes, bays, or oceans the problem may become so unmanageable that massive overloads of nutrient s produce unsustainably high amounts of algae that blo om, die, and rot. This process, known as eutrophication, consumes much of the available oxygen in wide swaths of these waters, leading to large and persistent hypoxic or anoxic dead zones where no fish can survive (Diaz and Rosenb e rg 2008). In the speci fic system being studied, Bowman Creek flows into the St. Joseph River which empties into Lake Michigan Eutrophication in the Great Lakes was recognized as a severe problem in the 1960s and 1970s (Schelske et al. 1983; U.S. EPA
18 1995) Current nutrient co ncentrations appear to be decreasing, partially due to the filtering feeding actions of invasive zebra and quagga mussels (Fahnenstiel et al. 2010; Mida et al. 2010) R ecent studies suggest that the rapidly growing populations of these exotic mussels could tip the nutrient balance too far in the other direction, removing unprecedented amounts of algae from the lakes and potentially destabilizing Great Lakes food webs and fisheries (Fahnenstiel et al. 2010) Other urban runoff inputs (including PCBs, mercury dioxins and excess sediment) remain pollutants of concern (U.S. EPA 1995) Bowman Creek certainly shows many of the typical characteristics of the urban stream syndrome. We were not able to obtain direct hydr ographic data for Bowman Creek, although the city is in the process of installing stream gauges at several locations (Gilot 2009) Even without hydrographs, the results of short, sharp increases in flow following rainstorms are readily observable. Figure 1.8a shows the MAIN sample site on Bowman Cre ek on July 14, 2010. Figure 1.8b shows the site on July 16 after a storm. F loodwaters filled the culvert to capacity and overflowed onto the surrounding banks, filling the grate with debris.
19 Figure 1. 8: MAIN site on Bowman Creek, a) on 7/14/10 before the storm and b) on 7/16/10 after the strom, showing f lood debris accumulation a) b)
20 Many sites in the urban areas of Bowman Creek also show extensive channel alteration. Trash buil dup is also a recurring pro blem. Garbage may be dumped directly into the creek, or it may wash into the stream along with other debris during high flow events as shown in Figure 1.9. Figure 1. 9: Channel alteration, t rash buildup and debris dam at LWE site on Bowman Creek
21 1.2.3 Culverts and buried streams Two of the twenty six key research questions in urban stream ecology as identified by Wenger et al. (2009) concern culverts and stream burial. Question 10 asks, What are the character istics of structure and function within piped and concretelined streams, especially with regard to biogeochemical processing? and Question 11 asks, How do piped and concrete lined streams affect ecosystem structure and function in downstream reaches? While many reviews and conceptual models of urban stream ecology have recognized culvert flow and stream burial as stresso rs a ssociated with urbanization ( Wenger et al. 2009), little direct field research has been conducted to investigate the mechanism by which these alterations affect stream ecosystems. The most cited papers dealing with buried streams were written by Elmore and Kaushal (2008) and Roy et al. (2009a ) These studies focused on mapping and modeling techniques. Elmore and Kaushal (2008) used hydrological modeling, aerial photography and GIS to identify where streams should be and compared these results to the current aboveground extent of waterways in Baltimore, MD. They calculated that 66% of streams within the city limits and 19% of stream s elsewhere in the catchment had been buried during urban development (Elmore and Kaushal 2008). Roy et al. (2009a ) conducted f ield assessments (groundtruthing) of ephemeral, intermittent a nd perennial stream channel origins in Hamilton County nea r Cinc innati, OH. They found that the standard 1:24,000 scale US Geological Survey maps underestimated stream channel length by 78% on average, contributing to the under reporting of headwater stream burial (Roy et al. 2009a ). Based
22 on ArcH ydro models, Roy et al. (2009a ) estimate d that the effects of urbanization have resulted in the loss of 93% of ephemeral stream channel length and 46% of intermittent channel length in Hamilton County. In urban areas, burial of headwater streams caused drainage network simplifi cation and consolidated the catchments once drained by smaller streams, resulting in a 22% increase in total perennial stream length compared with the modeled historical condition as shown in Figure 1.7 ( Roy et al. 2009a ). The conclusions of these studies have been influential in drawing attention to the m agnitude and extent of previously unrecognized urban headwater stream burial; however, both teams emphasize d the need for further study. Studies of s treams that flow through pipes for significant portion s of the middle of their length have not been reported in the literature. These systems are distinct from entirely buried headwater streams, fully piped stormwater systems, and otherwise unburied streams that receive piped stormwater input s Bowman Creek i s somewhat intermediate b etween these systems. Portions of t he headwaters are channelized while the middle section of the stream flows unimpeded through wooded areas. The downstream portion, in the South Bend urban area, flows through enclosed culverts fo r over half its length. 1.2.4 Urban stream restoration and daylighting buried streams Research gaps remain in the science of urban stream ecology, and the study of stream restoration is no exception. Timesensitive management decisions must often be m ade bas ed on incomplete data (Wenger et al. 2009). The available literature on urban stream restoration can be divided into several categories: daylighting methods specific
23 in stream restoration techniques whole catchment restoration models and broader appro aches to designing and managing urban ecosystems Pinkham (2000) provides an extensive overview of various daylighting projects involving the restoration of buried strea ms to an aboveground condition, addressing one of the main immediate causes of impai rment in these waterways. Aldridge et al. (2009) discuss using leaf litter addition to mitigate the lack of coarse particulate organic matter (CPOM) in urban streams with scouring flows Francis and Hoggart (2008) advocate the modification of existing rive r walls and other artificial structures to increase habitat complexity (by adding ledges and other features) in order to provide attachment surfaces for plants and other organisms. Many other common stream channel and bank restoration techniques, from rock weirs to rootwad revetments, may also be applic able in certain urban settings. However, a growing number of researchers and managers have found that stream restoration efforts will not be sustainable or successful in the long run unless the focus is shif ted from the stream channel to the entire catchment being drained (K.E. Anderson et al. 2006; Palmer 2008; Walsh et al. 2009) Streams reflect the state of their surroundings (Vannote et al. 1980) and treating the symptoms of the urban stream syndrome wit hout addressing the underlying causes may prove to be wasted effort (Suren and McMurtrie 2005) For example re grading and replant ing of heavily incised and eroding stream banks will be ineffective if high discharge stormwater flows continue to sweep away the new plantings especially if the chosen engineering approach was inappropriate for the stream being restored Figure 1.10 from Palmer (2008) shows an example of an
24 unsuccessful, overengineered stream restoration project. Palmer (2008) and Palmer et al (2005; 2010) present particularly insightful critiques of the shortcomings in stream restoration and propose guidelines for improving the integration of theory and practice in restoration ecology. Figure 1. 10: Inappropriately overengineered restoration, from Palmer (2008)
25 Palmer (2008) and Palmer et al (2005; 2010) emphasize the importance of scrutinizing and testing previously unexamined assumptions in restoration ecology. They critique the limited pe rspective of projects that focus on stream channel restoration without addressing catchment scale problems. Unless it is considered in the context of the surrounding environment, a simple, physical increase in the habitat heterogeneity of the stream channe l is unlikely to produce lasting, self sustaining increases in ecosystem integrity (Moerke et al. 2004; Palmer et al. 2010; Walsh et al. 2009). The assumption that improving in stream habitat structure will be sufficient to promote biodiversity has been re ferred to as the Field of Dreams hypothesis: if you build it, they will come (Palmer et al. 1997) This frame of mind is also similar to the fallacy that Feynman called Cargo Cult Science, in reference to the practices that developed in certain Pacific island regions following World War II (Feynman and Leighton 1997). Residents of some islands built imitation runways out of local materials, in the hopes of attracting planes laden with useful goods as had occurred during the war. As described by Feynman: They're doing everything right. The form is perfect. It looks exactly the way it looked before. But it doesn't work. No airplanes land. So I call these things cargo cult science, because they follow all the apparent precepts and forms of scientific inves tigation, but they're missing something essential because the planes don't land (Feynman and Leighton 1997). To avoid this cargo cult science, Palmer (2008), Palmer et al. (2005; 2010) and others emphasize that current restoration practices need greater input from current restoration t heory and that greater long term and fine grained assessments of restoration outcomes are essential for testing and refining the underlying theories.
26 An emerging consensus from the research community is pointing toward in situ retention and utilization of stormwater as a key aspect of wholecatchment stream restoration. Walsh et al. (2005a ; 2009) found that t he single aspect of urbanization most consistently linked with stream ecosystem degradation is not the percentage of land in the drainage basin covered with impervious surfaces (%Total Imperviousness, or %TI), but in stead the percent effective imperviousness (%EI) This metric takes into account the extent to which impervious surfaces are connected to streams; for examp le, roads draining into stormwater culverts that flow into streams would be directly connected. A building with no drainage system, far from streams, would not be directly connected because water flowing off the roof would have to percolate through the gro und before reaching a stream. Walsh et al. (2009) suggest a primary objective of zero directly connected i mperviousness for catchments where the ecological objective is to protect stream ecosystems, because the direct connection between impervious surface s and streams is a severe stressor to stream ecosystems. There are many different techniques for managing stormwater without sending it directly downstream. First, precipitation can be captured and utilized where it falls. Green roofs (Carter and Jackson 2007) rain storage tanks and rainfed gardens are several potential options (Walsh et al. 2009). Second, impervious surfaces can be redesigned to be pervious, allowing water to soak through to the so il and recharge the groundwater (Rushton 2001; Tennis 2004) Third, unavoidable stormwater runoff can be channeled and retained in swales and created wetlands for nutrient uptake and filtration by plants, fungi and microbes (Collins et al. 2010)
27 The history of s tream restoration projects in South Bend, India na (the area studied in this thesis research ) provide s an excellent illustration of these changes in the goals and practices of restoration Portions of Juday Creek were restored in 1997 as part of the permitting process during the construction of a golf c ourse (Moerke et al. 2004). This project focused on in stream restoration techniques, and while biotic recovery was relatively rapid, long term persistence is uncertain because of continued sedimentation at a watershed scale (Moerke et al. 2004). Moerke et al. (2004) conclude that In many instances, reach scale restorations may be ineffective in the face of basin wide degradation. In contrast, current plans for the restoration of Bowman Creek are firmly anchored in the context of a catchment wide stormw ater management program aimed at meeting federal water quality standards in the St. Joseph River (City of South Bend 2010). The city plans to focus on low impact development strategies and infiltration techniques, such as rain gardens, constructed wetlands bioswales, and riparian buffers (City of South Bend 2010). Gary Gilot, the Director of Public Works, described this rationale: Low impact development techniques are based on the premise that storm water is a resource, not a waste to be quickly transport ed a nd disposed of somewhere else. Instead of conveying and managing/treating storm water in large, costly, end of pipe facilities often located at the bottom of drainage areas, low impact development addresses storm water through small, cost effective lan dscape features (City of South Bend 2010). However, current proposed projects would result in the direct discharge of stormwater to Bowman Creek at an additional two sites (City of South Bend 2010). This action would be strongly discouraged under the guid elines developed by Palmer et al.
28 (2005; 2010) and Walsh et al. (2009), if the goal is to improve ecosystem function in the stream. A truly integrated approach to stream restoration will require an even wider perspective that not only encompasses the ma nagement of the catchment drained by the stream, but also addresses these issues in the context of the surrounding urban ecosystem and human community. Whole ecosystem approaches to urban planning and water management can be found in several books most no tably Water Centric Sustainable Communities: Planning, Retrofitting, and Building the N ext Urban Environment Novotn et al. (2010) Other relevant publications include Novotn and P. Brown (2007) Feyen (2008) and Berkowitz et al. (2002) Bejan and Loren te (2008) present an innovative approach to effective design, including civil engineering and water management applications, based on an understanding of material and energy flows in natural systems McDonough and Braungart (2002) and Benyus (2002) offer f urther discussions of biomimicry and suggestions for sustainable industrial systems. 1.2.5 Restoration and ecosystem health: community definitions and goals B efore beginning any restoration project the goals must be clearly defined and prioritized. Important questions include: What other factors are being considered in addition to stream ecosystem health?; How do we define stream ecosystem health ? and Who decides what definition to use? Human communities living alongside the stream are not m ade up of obj ective observers. E ach person or group may have different interest s in the stream, and may assign different values to various aspects of the watercourse and riparian habitat
29 Restoration projects are often designed to fulfill many different goals and c ommunication between scientists, policymakers, and community members is essential but sometimes overlooked (Rhoads et al. 1999; Walsh et al. 2005a,b ; Palmer 2008) As currently implemented, watershed restoration is largely an engineering driven process ( Palmer 2008). Projects that are designed solely from a civil engineering perspective often prioritize the reduction of flooding and erosion to protect public and private infrastructure (Novotny et al. 2001) In f ocusing on the symptoms, this approach may f ail to address underlying landuse factors that are contributing to recurrent hydrological and ecologic al problems (Novotny et al. 2001; Palmer 2008). Community organizations may have different views of the characteristics that would constitute a success ful restoration project. Local governments have a responsibility to their constituents. In making stream restoration decisions they must consider balancing budgets, addressing public concerns, and meeting state and federal mandates (City of South Bend 201 0) Many communities rely on drinking water drawn directly from rivers and lakes, and contamination or hydrological alterations also have the potential to affect groundwater and nearby wells (Niemczynowicz 1999) I ndustrial concerns may arise if the river is large enough to generate hydroelectric power, cool a power plant, or serve as a shipping channel (Novotny et al. 2001) Rivers and streams that flow through agricultural lands present their own unique set of usage concerns and challenges, such as livest ock access, crop irrigation, and runoff of fertilizer, livestock waste, pesticides, and topsoil (Gordon et al. 2008) Individual residents also express varying opinions about urban water features.
30 Adams et al. (1984) surveyed several hundred Columbia, MD residents to ascertain their views on urban ponds and basins that were constructed for stormwater mitigation. The most common concern associated with ponds was appearance ( weedy/full of algae/unattractive ), as cited by 37% of respondents (Adams et al. 1984). Other perceived potential nuisances or hazards included risks to children (23% of respondents ), mosquitoes and other bothersome insects (23%) muddy water (22%), trash/debris (19%) and offensive odor (10%) (Adams et al. 1984). Residents were also con cerned that ponds may attract undesirable wildlife; write in comments listed raccoons (14%), muskrats (11%), and snakes (9%) as particularly undesirable However, more respondents viewed the urban ponds positively as an overall benefit to the community (47%), than negatively as a nuisance or hazard (13%). Ponds were considered to have both beneficial and nuisance aspects by 36% of residents surveyed. Among respondents, 94% agreed that wetlands add to the beauty, diversity, and quality of the human living e nvironment and 98% enjoyed viewing birds and other wildlife species (especially ducks) at the ponds. Among homeowners surveyed, 75% felt that permanent bodies of water added to real estate values (Adams et al. 1984). zgner and Kendle (2006) found that residents of Sheffield, UK reported a similar mix of perceived benefits and concerns relating to u rban naturalistic spaces. T he authors also provide a review of other studies on this topic Community comments and concerns about Bowman Creek include many of the issues discussed by Adams et al. (1984) In a local news article about the planned restoration, South Bend resident Tyrone Brelsford described his perceptions of Ravina Park, a small urban green space that spans a section of Bowman Creek: This is
31 somewhere you can relax and enjoy nature. You can bring the kids here and play in the park. But, I wouldnt bring my kids down near this river (WSBT 2010) The same article also included a quote from resident Taryn Jones: It worries me a lot with all the new diseases that are coming out. Our creek needs to be clean. But, every time I walk by, theres like an odor. Its too beautiful to be smelling like that. Public opinions on aesthetics and risk can influence specific components of stream restoration p olicy. In a survey of students from nine countries, Pigay et al. (2005) found widespread cultural barriers to the introduction of wood to rivers in the course of restoration. Fallen trees and branches provide important instream habitats, but respondents from France, India, Italy, Poland, Russia, Spain and Texas (USA) rated photographs of stream channels with wood as less aesthetic and more dangerous than riverscapes without wood (Pigay et al. 2005) Results from surveys in Germany and Oregon (USA) showed the opposite trend, and responses from Swedish students showed no significant preference between the two conditions (Pigay et al. 2005). Urban and rural residents may value riparian ecosystems differently. Schrader (1995) conducted a survey of agricultural landowners in Dickinson County, Kansas, focusing specifically on perceptions of streamland and riparian areas. This study found that respondents perceive the highest importance of streamland to be that of maintaining environmental quality (Schrader 1 995) Wildlife habitat protection, visual appearance and stream water quality protection were ranked as more important uses for riparian areas than cropland or livestock production (Schrader 1995). The use of streams for active recreation activities of hu nting, fishing and boating was ranked equally with income -
32 generating agricultural activities and much higher than passive recreation (Schrader 1995) This result shows a departure from the largely passive recreational activities often reported in urban ar eas ( Adams et al. 1984; zgner and Kendle 2006) Walsh et al. (2005b ) note that urban stream attributes with limited ecological values, such as mowed grass riparian zones or paved streamside paths, may have amenity values for some urban communities and that Sometimes, value placed in such altered, unnatural environments can be a product of people not missing what they never had Walsh et al. (2005b ) suggest that concerned stream ecologists should take a more active role in community education and Louv (2008) considers increased experiential interaction with natural environments to be highly beneficial for children and society. Purcell et al. (2007) describe d an urban stream restoration project in Berkeley, CA that involved a fairly successful collabor ation between local high school and undergraduate students in invasive species removal, native plant revegation, and continued site monitoring. Wenger et al. (2009), Palmer (2008) and Rhoads et al. (1999) emphasize that communication between restoration theorists and practitioners is most effective when the exchange of information is mutual Academic ecologists need to become more aware of the informational needs of policymakers, managers, and community members As described by Palmer (2008): A significa nt amount of fundamental ecological knowledge dealing with issues such as system dynamics, state changes, context dependency of ecological response, and diversity is both under used by managers and practitioners and under developed by ecologists for use in real world applications. The e cological community is continuously discussing and refining the theoretical frameworks and
33 protocols for defining, assessing and restoring ecosystem health (see Sections 1.2.4 and 1.3), and many aspects of our understanding of stream ecosystem function remain incomplete. Wenger et al. (2009) point out that while some of these questions might never be answered fully from a scientific perspective,  streams neverthele ss must be managed now. The consensus best management goals and practices that emerge from academic debates need to be communicated clearly and effectively to policymakers. If advances in ecological understanding are unheard, mis applied or disregarded t he practical benefits of the research may be lost (Palmer 2008; Wenger et al 2009). 1.3 Stream ecology in a scientific context 1.3.1 Models and limitations Aquatic ecologists working on restoration project s may feel that they are speaking for the stream to the best of their ability much a s Dr. Seuss' Lorax said, I speak for the trees, for the trees have no tongues (Geisel 1971) Because it is not possible to interrogate the stream ecosystem directly about its health and wellness scientists must come to an un derstanding of the whole by systematically investigating its parts and piecing together bits of data to form an entire picture. The state of an ecosystem at any given time is a snapshot of a dynamically evolving process, a self orchestrating cacophony of r elated rates tumbling gracefully through time and space (Kay 1991) As described by Heisenberg (1958) : What we observe is not nature itself, but nature exposed to our method of questioning. Sometimes our results may appear as incomplete and inconsistent as those of the blind men who came to wildly different conclusions about the nature of the whole elephant by the separate investigation of its parts. We can
34 only strive to integr ate our fragmentary data into a full and accurate description One method that ecologists use to evaluate ecosystem health or habitat value is to compare the stream of interest to a reference condition. Ideally, this would involve collecting long term data on comparable streams (of a similar size, in the same geographic region) th at are as undisturbed by readily visible human intervention as possible, then comparing various aspects of the healthy streams to those in the study site. To make these comparisons ecologists often calculate an index value that is based on the diversity of invertebrate species, weighted by their sensitivity to pollution. The ratios between different invertebrate guilds can also be used to provide more detailed information on habitat and food web dynamics (Poff et al. 2006) Section 1.4 provides more information on these aspects of monitoring. T he notion of stability in ecosystems is a tricky concept, as these complex systems are constantly in flux, and vary over time and space. Streams, in particular, often have low resist ance to outside disturbance due to low levels of stored biomass, but high resilience, or ability to recover from disruptions, due to high throughput flow of matter and energy (DeAngelis 1980; Webster et al. 1983) When attempting to isolate and analyze a single parameter (e.g. productivity) as a proxy for ecosystem function it may be possible to pick out patterns in the data that can be used to predict how different variables will affect this process. Models can be used to predict the outcome of a given initial condition or perturbation of the system. These results can be divided into four types, as described by Wolfram (2002) in his studies of cellular automata. These four classes are: 1) a stable final state; 2) repeating patterns or oscillations; 3) u npredictable,
35 stochastic states; or 4) states with complex features (Wolfram 2002) Ecologists use different types of studies to answer questions about ecosystem function. Ecological assessments describe the current state of a system and long term monit oring provides insights on changes over time Comparisons between systems (for example, different streams or differe nt subsites on the same stream) are often used to prioritize or assess management or regulatory actions. Experimental manipulations and befo re after control impact studies attempt to isolate the effect of a pa rticular variable on the system. However, the theoretical and statistical foundations of spacefor time substitutions and pseudoreplication are currently under debate (Pickett 1989; Oksanen 2001; Carter et al. 2009; Miao et al. 2009) Other research approaches synthesize many types of information to create a model that can be used to predict future states of the system given particular inputs. 1.3.2 Nutrient and energy flow in stream ecosystems The main questions that ecologists ask about stream metabolism and ecosystem function include: What are the inputs and outputs of the system? What are the limiting factors or reactants? How quickly is matter being cycled through the system? For stream sy stems, the most influential factors can be divided into several categories. Physical habitat characteristics (flow, attachment surfaces, pH, temperature, pressure, etc.) are important for all organisms, and light is an additional requirement of photosynthe tic plants, algae and cyanob acteria. The availability of chemical nutrients also plays a key role in regulating stream metabolism. These compounds include usable nitrogen and phosphorus organic carbon (for heterotrophs), and dissolved oxygen (for aerobic
36 organisms). Additional micronutrients are also necessary for plant and animal growth, and other solutes can certainly affect aquatic ecosystems, but these are the most important. The nutrient spiraling concept (Webster and Patten 1979) and the river conti nuum concept (Vannote et al. 1980) are two of the main foundational models used to underst and stream ecosystem function. Flowing water the defining feature of lotic systems, is central to both ideas. Flow creates an additional axis, a directionality or ve ctor in stream ecosystems that is not present in lentic (still) waters such as ponds and lakes. In a pond, processes such as nutrient uptake and release vary over time but co occur in the same location In a stream, these temporally variable processes are also spread out spatially, over a downstream distance that is determined by the vel ocity and volume of moving water Nutrient and energy cycles in stationary terrestrial or lentic ecosystems ar e often represented by simple circle s or networks, but to und erstand these processes in streams, it is necessary to take that circular cycle and stretch it out orthogonally (like a Slinky toy) in the downstream direction. This is the essence of the nutrient spiraling concept as shown in Fig 1.11 (Minshall et al 1983)
37 Figure 1. 11 : Nutrient spiraling, from Minshall et al. (1983) In Figure 1. 12, a diagram from Allan and Castillo (2007) provides a different look at the same concept, emphasizing the fact that th e nutrient spiral is a model rather than an explicit description of the movement of nutrients through stream systems Studies of nutrient dynamics focus on describing the average distance traveled by a nutrient particle in the water column before it beco mes incorporated into living systems (the uptake length). Researchers can also measure the average distance that an assimilated nutrient particle travels within the biota before being rereleased into the water (the turnover length) (Allan and Casti llo 2007). These lengths can then be divided by the velocity of the current (or, typically, the discharge) and converted into a rate. The discharge, or
38 volume rate of flow, is the water velocity multiplied by the cros s sectional area of the stream. The s pecific experimental procedures and calculations are significantly more complicated. One method involves releasing a known amount of the nutrient of interest, along with a conservative (nonbiologically active) tracer, and monitoring the change in concentr ations over time at a station downstream (Webster and Valett 2007) The conservative tracer is used to control for nonbiolo gical effects such as dilution. Another technique uses the addition of stable isotopes (such as 15N) to monitor uptake (Tank et al. 2007) An in depth discussion of the theory and practice of nutrient uptake research can be found in Stream Solute Workshop (1990) Ensign and Doyle (2006) provide a review and quantitative synthesis of the available nutrient spiraling literature. Figure 1. 12: Another representation of nutrient spiraling, from Allan and Castillo (2007)
39 The dynamics of nitrogen, phosphorus, and carbon receive the most attention in studies of stream ecosystem function. Thes e elements are necessary component s of all known living systems. Reactions involving hydrogen and oxygen, the defining components of aquatic ecosystems, are also important regulators and indicators of stream metabolism (as measured by pH and dissolved oxygen concentrations). At any given point along the stream, the biotic community receive s inputs of energy from upstr eam and from outside the stream and transforms these resources into interesting patterns and processes. The products may be retained in a give n section of the stream for a certain amount of time, but eventually exit the community and flow downstream or out of the stream. While individual molecules and units of solar energy eventually pass through, if more arrive to take their pl ace the overall p attern may be sustained in place indefinitely. As Woese (2004) described it: Imagine a child playing in a woodland stream, poking a stick into an eddy in the flowing current, thereby disrupting it. But the eddy quickly reforms. The child disperses it agai n. Again it reforms, and the fascinating game goes on. There you have it! Organisms are resilient patterns in a turbulent flow patterns in an energy flow.  It is becoming increasingly clear that to understand living systems in any deep sense, we must co me to see them not materialistically, as machines, but as stable, complex, dynamic organization. Where do these inputs come from, what happens to them in the stream, and where do they go afterwards? The energy that powers the system ultimately comes fr om the sun, whether directly (as photons absorbed by instream autotrophs) or indirectly (as the chemical bonds in compounds fixed by terrestrial photosynthesis, or in the flow of
40 moving water). Energy leaves the system in concentrated forms (biomass), par tially concentrated forms (debris and certain metabolic waste products), and diffuse forms (w aste heat). These products can also be thought of as useful, somewhat useful and not particularly useful for use in fur ther metabolic transformations. Bi ologically impor t ant nutrients also enter and leave the system in concentrated forms, partially c oncentrated and diffuse forms. These states can also be described along a gradient of high energy, bioavailable, and reactive to low energy, unavailable, an d nonreactive. Figure 1.13 (Odum 1956) shows a schematic representation of the inputs and outputs of a stream ecosystem. The ratios of these different forms of energy and matter, and the rates of conversion among them, are of primary importance in underst anding stream metabolism.
41 Figure 1. 13: Energy flow through a stream ecosystem, from Odum (1956) Every organism needs nitrogen to produce proteins and nucleic acids, the building blocks and genetic instruc tions of the cell Nitrogen makes up 78% of the Earths atmosphere, but these molecules are not readily available for metabolic use because they are locked together in a triple bonded diatomic form, N2. Breaking this bond requires a great deal of energy an d/or very specific conditions (such as those found in the active s ite of bacterial nitrogenases) (Kstner and Blchl 2007) However, nitrogen containing compounds can be metabolically useful means of storing energy because they release so
42 much energy upon decaying to their stable state. L ike pushing a rock up a hill, this energy can be used later to do useful work. Stream ecosystems receive inputs of ammonium and nitrate from upstream and from terrestrial sources, along with nitrogen containing compounds i n organic debris. Nitrogen fixing bacteria, including some cyanobacteria, also capture some nitrogen directly from atmospheric N2. These organisms catalyze the reaction of N2 and hydrogen to form ammonium (NH4 +) in a process called ammonification or nitro gen fixation The ammonium is then available for use by algae and plants, or it may be converted into nitrite (NO2 -) and then nitrate (NO3 -) by a different set of bacteria. This process is known as nitrification. Using solar energy and atmospheric carbon, photosynthetic organisms incorporate the NH4 + and/or NO3 into proteins and other biomolecules. These organic compounds are then available for use by heterotrophs, which consume plant matter to assimilate the nitrogenous compounds they need to survive and grow. Another group of bacteria can also strip successive oxygen molecules from dissolved nitrate, converting it into nitrite and back into dinitrogen (N2). This dissolved N2 can then equilibrate with atmospheric N2 concentrations. Aquatic insects, fish, and amphibians excrete nitrogenous waste s (the byproduct s of protein and nucleic acid breakdown) in the form of ammonia ( NH3) (Allan and Castillo 2007; Gullan and Cranston 2010) In water, ammonia exists in equilibrium with ionized ammonium. At higher pH le v els, the balance shifts towards more toxic ammonia (U.S. EPA 2009) The n itrogen that enters the stream ecosystem may be incorporated into biomass, temporarily stored in sediments, re released into the atmosphere, or exported downstream (in organic compou nds, ammonium or nitrate)
43 (Tank et al. 2007) Prior to human industrial innovations, nitrogenfixing bacteria and cyanobacteria were the main sources of biologically available nitrogen. The invention of the Haber Bosch proces s i n 1909 dramatically changed the global nitrogen cycle (Smil 2004) The commercial fixation of nitrogen into ammonium nitrate is an energyintensive process that relies on fossil fuels as a source of hydrogen (Smil 2004) Through industrial fertilizer manufacturing and increased cul tivation of legumes, w e have more than doubled the natural background rate of nitrogen fixation, an increase that Rockstrm et al. (2009) consider dangerously unsustainable. Fertilizer runoff and wastewater inputs overload stream ecosystems with more nitra tes than they can process effectively. If the nitrogen cannot be converted into biomass or released into the atmosphere, the excess is washed downstream. There, nitrates and phosphates can lead to harmful algal blooms and anoxic conditions (Diaz and Rosenb erg 2008) Phosphorus is an other essential component of life on Earth; it is present in the lipid bilayer that encloses every cell, in the scaffolding that supports our DNA and our skeletons, and in ATP, the energy currency that powers our most basic cell ular reactions. Once it is weathered from soil substrata and incorporated into plants, phosphorus may move quickly through trophic interactions among living organisms, but over time its net movement is inexorably toward the sea, never to return until geolo gic processes reincorporate it into benthic sediments and uplift the ocean floor into new highlands (Filippelli 2008) The global cycle of phosphorus is the slowest of all the essential nutrients, because while nitrogen and carbon circulate freely in the atmosphere (as N2 and
44 CO2), phosphorus is constrained to remain in a solid state (Allan and Castillo 2007) Phosphorus only moves from place to place when it is carried by flowing water, whether in the form of streams or the internal aqueous environments of busy animals. (A small amount of phosphorus also blows from the land to the sea in airborne dust [Carpenter and Bennett 2011]). Humans, by far the busiest of the animals, now move more phosphorus against the current from coastlines to uplands than untold eons of salmon and seabirds (Carpenter and Bennett 2011) Excess phosphates enter streams mainly through fertilizer runoff (Carpenter and Bennett 2011) In excess, what should be a blessing becomes a curse and the consequences of extensive eutrophication may be severe indeed (Rockstrm et al. 2009) Habitat degradation and loss of ecosystem function in urban streams exacerbates the problem of nutrient overloading. If higher amounts of nutrients are coming in to the stream from urban sources, and lower amo unts are being removed by biological processes, more nutrients will be exported downstream (Meyer et al. 2005) H eadwater streams can be critical locations for processing N (Royer et al. 2004) These small streams often exhibit greater efficiency in nitr ogen processing due to high biological activity combined with increased sediment/water contact time (Tank et al. 2007) Larger streams often have longer nutrient spiraling lengths and decreased efficiency of denitrification (conversion of nitrates to N2) (Royer et al. 2004), though this trend may be equally applicable in all cases (Ensign and Doyle 2006) Studies in urban streams have consistently reported reduced nutrient uptake as compared to reference conditions (Walsh et al. 2005b ) Meyer et al. (2005) conducted a nutrient addition study in streams along an
45 urban gradient near Atlanta, GA. They found that NH4 and soluble reactive phosphorus ( SRP ) uptake velocities (the rate at which a nutrient moves [downward] through the water column toward the benthos ) decreased with increasing urban intensity (Meyer et al. 2005). Th eir results were likely more conservative than the actual conditions Eleven of the 21 urban experiments were excluded from calculations because the change in nutrient concentrations ove r the sample reach was too smal l to measure accurately (Meyer et al. 2005). This outcome strongly suggested that the uptake velocity was even lower in these urban streams than in the other sites studied, indicating that the spiraling length was actually lo nger than the stream section itself (Meyer et al. 2005). The net concentrations of the added nutrients were hardly affected by their travel through these sections of urban streams. Meyer et al. (2005) suggest ed that the low nutrient uptake velocities obser ved in highly urbanized streams appear to result from reduced biotic demand for nutrients due in part to reduced retention of organic matter. Direct connectivity to stormwater pipes increases the effective imperviousness of the catchment and the flashi ness of the streams hydography (Walsh et al. 2009) Flashiness increases streambed scouring, leading to decreased retention of organic matter and lowered habitat complexity, as well as physically displacin g algae and macroinvertebrates (Borchardt 1993; Al dridge et al. 2009) Reduced retention of organic matter leads to lower biotic demand for nutrients (Meyer et al. 2005). Lower biotic uptake decreases the rate of nutrient removal (e.g. the conversion of nitrates to N2, which would then diffuse out of the stream), leading to increased export of nutrients downstream (Meyer et al. 2005). Urbanization may also a ffect nutrient processing in debris dams, streambank
46 sediments and in the hyporheic zone ( the porous groundwater exchange region underneath the main stream channel ) Bernhardt et al. (2008) offer an extensive review of current research in nitrogen processing in urban streams, as well as strategies for remediation. Wenger et al. (2009) provide a shorter summary of these and other recent findings on nut rient dynamics in the context of urban stream ecology. Microbial denitrification (the removal of oxygen atoms from nitrate to form N2) is dependent on anaerobic conditions (Allan and Castillo 2007) and the availability of carbon for heterotrophic bacteria (Groffman et al. 2005) Groffman et al. (2005) found that organic debris dams (branches and leafpacks) and gravel bars that are rich in organic matter have the potential to function as hot spots of denitrification in urban streams. The denitrification potential of these structures actually increases in response to higher stream NO3 concentrations, providing an important regulator y feedback mechanism (Groffman et al. 2005). However, this protective capacity is lost if scouring stormwater flows prevent debris dam retention in urban streams, or if a lack of leaf litter prevent s debris dam establishment (Groffman et al. 2005). Aldridge et al. (2009) found that microbial activity in leafpacks and debris dams is also important in phosphorus retention. Kausha l et al. (2008) found that denitrification rates were substantially higher in the hyporheic zones of riparian areas with low, hydrologically connected streambanks designed to promote flooding and dissipation of erosive force for storm water management as opposed to restored high nonconnected banks or unrestored reaches. Mulholland et al. (1997) reported that hyporheic metabolism can also contribute significantly to phosphorus uptake in streams, depending on the comparative size of the transient st orage
47 zone. 1.3.3 The river continuum concept and stream metabolism The river continuum concept is fundamentally linked to stream metabolism and the dynamics of carbon flow through stream ecosystems. Like the nutrient spiraling concept, this model can also be understood in terms of stretching time dependent stationary ecosystem process es into spatially dependent flowing ecosystem process es In this case, the factors of interest are successional trophic and metabolic interactions including the breakdown of ter restrial leaf litter, changes in macroinvertebrate community composition, and the overall balance between production and respiration (autotrophy vs. heterotrophy). From headwaters to large rivers, each section of the drainage network displays a different s uite of characteristics and exerts different influences on the communities downstream. Fig ure 1.14 gives an overview of the upstream to downstream changes in community composition and ecosystem function that are predicted by the river continuum concept (Vannote et al. 1980).
48 Figure 1. 14: The river continuum concept, from Vannote et al. (1980)
49 Headwater streams receive inputs of organic material leaves, twigs, and other coarse particulate organic m atter or CPOM from trees overhanging the stream and debris washing in from the drainage basin. These allochthonous, or external, inputs represent sources of carbon, nutrients, and energy that were fixed from the atmosphere, soil and sun by terrestri al plants. Before these resources can be made available for stream organisms, they must first be broken down into usable bits. Shredder invertebrates are able to exploit this niche and are consequently more common in headwater streams than in larger rive rs downstream. Collector or gatherer invertebrates take advantage of the second step in the leaf breakdown process, feeding on the fine particulate organic matter or FPOM that the shredders leave behind. Organisms in this functional feeding group als o make up a large proportion of headwater insect communities. Small streams that are shaded by overhanging trees may not receive enough light to support a great deal of in stream photosynthesis (by algae and macrophytes), so the overall ratio of gross primary production (GPP) to ecosystem respiration (ER) is often less than 1 (Benfield 2007) Streams with a GPP/ER ratio (also called a P/R ratio) < 1 are described as heterotrophic. Trave ling downstream, stream order and width increase and other characterist ics change accordingly. Downstream reaches receive the exports of upstream reaches, as terrestrial nutrient inputs are broken down into finer and finer particles and reassimilated into aquatic organisms of increasing size and trophic level. In comparison t o headwater streams, midsize streams and rivers contain similar proportions of collector invertebrates gathering FPOM, but host fewer shredders and more scrapers. These scraper organisms
50 feed on the microbial mats that cling to the surfaces of submerged rocks, logs and leaves. These microbial mats are largely composed of algae, bacteria, fungi, and meiofauna  that exist within a mucilaginous, polysaccharide matrix, and the assemblage is referred to as periphyton, aufwuchs, or biofilm (Steinman and Mu lholland 2007) Because medium size streams are wider, they receive more sunlight and can support more in stream photosynthesis. If the P/R ratio is greater than 1, the stream is described as autotrophic. Large rivers have a greater cross sectional area a nd transport much higher volumes of water per unit time than smaller streams. Their wide channels often receive ample sunlight, promoting photosynthesis by phytoplankton. However, the surface areato volume ratio of large river channels is lower than that of headwater streams, providing proportionally fewer attachment surfaces for aquatic organisms. Additionally, areas of deep or turbid (murky) water may block sunlight from reaching the substrate, preventing the growth of macrophytes and algae. Large rivers collect the nutrients, sediments and other debris that wash in from connecting streams, supporting large populations of filter feeding invertebrates. These rivers tend to be heterotrophic with a net P/R ratio < 1. What does this production/respiration r atio mean, and how is it measured? Primary productivity is the rate of formation of organic matter from inorganic carbon by photosynthesizing organisms, and respiration releases this stored energy by breaking down organic compounds and re forming (Bott 2 007) The simplified net equation for photosynthesis is: 6 CO2 + 12 H2O 2 + C6H12O6 + 6 H2O Re spiration can be understood as the reverse of this process. Wholestream metabolism
51 can be monitored by tracking changes in diss olved O2 or CO2 concentrations pH or uptake of added [14C] bicarbonate (Odum 1956; Marzolf et al 1994; Bott 2007) Dissolved oxygen (DO) is the most commonly studied parameter Measurements can be carried out in chambers that enclose or cover a section of stream substrate, or in the open water of the stream channel (Marzolf et al. 1994; Bott 2007) Chamber measurements provide more controlled conditions but may not accurately reflect all aspects of wholestream metabolism. Open channel measurements are more representative but introduce additional variables that must be taken into account. The most im portant of these is the reaeration constant, or the change in diss olved oxygen concentration due to nonbiological causes such as turbulent flow or diffusion from the atmosphere. This parameter can be determined by bubbling a volatile gas such as propane in to the stream and measuring the rate at which it diffuses out, or by taking various measurements of stream hydrology and using a mathematical model ( Marzolf et al. 1994). This thesis research used deployable dataloggers to collect open channel DO measurements, with the intention of calculating gross primary productivity (GPP), ecosystem respiration (ER) and reaeration (k) through the use of a model developed by Drs. Mike Grace and Mike Harper at Monash University, Melbourne, Australia. This protocol is use d by the Tank Lab at the University of Notre Dame. Grace and Imberger (2006) discuss these techniques and models in great detail When instream dissolved oxygen concentrations are tracked over several days, repeated daily fluctuation s are often observed. Pa tterns of daily DO variation are known as diel respiration curves. All aerobic organisms consume oxygen via respiration
52 throughout the day. During daylight hours, photosynthetic organisms release additional O2 into the system (though this curve is not c entered on noon due to the lag in processing). Figure 1.15 shows the change in DO over a 24 hour period : Figure 1. 15: Diel respiration curve, showing GPP and CR (from Marzolf et al. 1994) In Figure 1.15 (Marzolf et al. 1994) the community respiration (CR24) is the integral of the background rate of oxygen consumption, not including oxygen flux due to reaeration. The authors estimate the daytime respiration by drawing a line through the nighttime respirat ion data. The gross primary productivity (GPP) is the difference
53 between the observed increase in DO and the line estimating daytime respiration. T he authors calculated that the community respiration rate exceeded the rate of gross primary production for t he 24hr period shown in Fig. 1.15 (CR24 = 19.69 g O2m2s1 and GPP = 20.17 g O2m2s1). See Section 4.1.2 (p. 109) for diel respiration curves from Juday Creek and Bowman Creek. Dissolved ox ygen concentrations are also important for the health and survival of aerobic aquatic organisms ( Fig. 1.16) : Figure 1. 16: D issolved oxygen tolerance (from http://www.water research.net/ Watershed /dissolvedo xygen.htm) Daily fluctuations in pH can also be observed in stream ecosystems, and these are closely linked to the diel respiration curves (Maberly 1996) When respiration exceeds primary productivity, CO2 concentrations increase and pH d ecreases, accord ing to the equilibrium: CO2 + H2O H2CO3 HCO3 + H+ (Maberly 1996) The same system is responsible for the bicarbonate buffer that helps maintain a steady plasma pH in human blood. (See Figure 4. 7 on p. 114 for diel pH curves from B owman Creek.) As respiration
54 increases, aquatic animals excrete more ammonia along with CO2 In an interesting and fortunate self regulatory mechanism, the increasing acidity shifts the NH3/NH4 + equilibrium toward the less toxic ammonium. P/R ratios can also be estimated by analyzing biological samples. Biofilm samples can provide information about the P/R ratio by comparing the chlorophyll a content to the total amount of organic matter (the ash free dry mass, or AFDM). The ratio of AFDM to chl a (both in mg/m2) is the Autotrophic Index (AI). Higher values generally indicate heterotrophic conditions, while lower values indicate autotrophic conditions. According to Steinman et al. (2007) normal AI values range from 50 to 200. Macroinvertebrate communi ty composition can also be used to describe the metabolic activity of a stream reach. The ratio of scraper invertebrates to ( shredders + collectors) provides another estimate of the P/R ratio For this method, a P/R ratio > 0.75 suggests an autotrophic str eam with a net production of organic matter, and a ratio < 0.75 suggests a heterotrophic stream with net consumption of organic matter (Merritt and Cummins 2007) See Table 1. 1 in S ectio n 1.4.6 for mor e information and see Section 4.3.2 for P/R data from Bowman Creek macroinvertebrate analyses Benthic macroinvertebrates can be used to estimate secondary productivity, though this labor intensive process requires measuring l arge numbers of specimens and using length weight regression tables to calculate the invertebrate biomass (Kohlhepp and Hellenthal 1992; Lamberti and Berg 1995; Benke and Huryn 2007; Entrekin et al. 2007) An understanding of carbon dynamics is essential for understanding stream ecosystem function (Tank et al. 2010) As described in Allan and Castillo (2007): A high
55 rate of utilization [rate of conversion of organic C to CO2] relative to transport indicates that organic matter is contributing to stream m etabolism and the stream ecosystem is efficient in its processing of organic C inputs. The opposite result indicates that most organic matter is exported downstream and the stream ecosystem is inefficient. The lability and quality (availability and energy density) of in stream carbon sources can have important regulatory effects on nutrient processing (Strauss and Lamberti 2000; Groffman et al. 2005; Meyer et al. 2005; Aldridge et al. 2009) As described by Groffman et al. 2005: The balance between N sour ce and sink processes in streams is strongly influenced by C fluxes. Carbon is the energy source that drives microbial immobilization of inorganic N as well as most denitrification activity Groffman et al. (2005) Meyer et al. (2005) and Strauss and Lamb erti (2000) discuss the complex effects of organic matter on nitrification and denitrification. Meyer et al. (2005) and Aldridge et al. (2009) found that organic matter availability increases phosphorus uptake. Urbanization can influence various aspects o f stream ecosystem function through many different mechanisms. Controlled experimental manipulations are chal lenging, because researchers using this approach must find a balance between regulating unwanted sources of variation and accurately representing t he dynamic nature of the system being studied. Pringle and Triska (2007) describe this paradox: The more control exerted by the investigator, the more replicable the result, but the less applicable to natural systems. On the other hand, less controlled fi eld experiments and observations may accurately describe the current periphyton community at a particular site, but yield little insight into what factors control community development. Several studies have
56 address ed these difficulties by creating physica l model stream systems in which to test the explanatory power of theoretical models (Culp and Baird 2007) Mulholland et al. (1995) and Cardinale (2011) present particularly interesting examples of this type of investigation. Mulholland et al. (1995) teste d the implications of the river continuum concept, the nutrient spiraling concept the serial discontinuity concept and the patch dynamics concept by monitoring biofilm communities in constructed laboratory stream segments. Biofilm communities along a long itudinal gradient responded to changes in nutrient dynamics caused by upstream communities (Mulholland et al. 1995). Downstream communities contained higher percentages of cyano bacteria and exhibited increases in nutrient recycling within segments (Mulholland et al. 1995). Cardinale (2011) manipulated biofilm biodiversity and habitat heterogeneity in 150 experimental stream mesocosms. D iverse algal communities in complex habitats removed nitrate from the stream more effectively than low diversity biofilms, a result that Cardinale (2011) attributed to niche partitioning. Th ese type s of experiment s help test the foundations and assumptions of stream ecology, providing important linkages between theory and practice. Biodiversity is often used as an indicator of ecosystem health, but it is important to check that this theoretical framework is sound. 1.4 Diversity indices and monitoring regimes 1.4.1 Overview of benthic macroinvertebrate biomonitoring Standardized sampling methods for analyzing freshwater lotic benthic ma croinvertebrate communities provide invaluable data for understanding and quantifying the ecological effects of habitat disturbances (especially sedimentation and
57 toxic pollution) on stream habitats. Rapid Bioassessment Protocols (RBPs) focus on efficient data collection but lack the in depth data produced by more thorough methods. Some sampling programs involve community or student participation, promoting outreach and education. The development of these protocols is dependent on an understanding of the re gi onal, ecological and taxonomic context of the system being studied, and the results are usually used to make stream manag ement decisions 1.4.2 What are benthic macroinvertebrates? Benthic macroinvertebrates are the unsung, polysyllabic heroes of aquatic eco logy. These freshwater creatures include bottom dwelling crustaceans, mollusks, annelids, nematodes, mites and insects, though the last group is unquestionably the most rigorously studied. Adult t rue bugs (Hemiptera) and beetles (Coleoptera), along with th eir immatures and the immature stages of mayflies (Ephemeroptera), stoneflies (Plecoptera), caddisflies (Trichoptera), dragonflies (Odonata: Anisoptera), damselflies (Odonata: Zygoptera), dobsonflies and fishflies (Megaloptera: Corydalidae), alderflies (Me galoptera: Sialidae), and flies (Diptera), especially the midges (Diptera: Chironomidae) are some of the more commonly encountered groups (Merritt et al. 2008; McCafferty 1983; Lehmkuhl 1979) The technical definition of a benthic macroinvertebrate can be explained by splitting the term into its constituent parts. These organisms are members of the benthos, or bottom dwelling aquatic organisms, for at least part of their life cycle. They are macro invertebrates, meaning that they are retained by mesh si zes micrometers (Rosenberg and Resh 1992). Tiny invertebrates or meiofauna, such as
58 hydras and copepods are excluded from this definition. Finally, these protocols focus on invertebrates While larval fish and amphibians of similar size may share the same habitat, these animals are studied using different methods. 1.4.3 Why are benthic macroinvertebrates used to measure water quality? Benthic macroinvertebrates are the stream ecologist's version of the proverbial canary in the coal mine. Just as miners once carried caged canaries into tunnels to monitor levels of dangerous gases, scientists can examine communities of aquatic invertebrates to estimate the ecological health of a stream. Canaries succumb to carbon monoxide poisoning at concentrations below the level of human toxicity. I n a similar fashion, larval mayflies, stoneflies and caddisflies are more sensitive than other groups to sedimentation and pesticide runoff. Pollution intolerant aquatic insects are therefore likely to be among the firs t taxa to disappear in impacted streams. Carter et al. (2007) point out that, Unfortunately, there is circularity in the use of tolerance values in that they are based on where the organisms are found and then applied in an assessment based on an organism s distribution. An inferentially stronger approach would be to base tolerance values on empirically deriv ed laboratory and field testing. Dang et al. (2009) propose that differences in nutrient to carbon composition rati os among macroinvertebrate taxa co uld provide an explanatory mechanism for observed va riations in pollution tolerance. Benthic macroinvertebrate biomonitoring can also be compared to forensic entomology. If authorities suspect that a human has been killed by poison, forensic entomologist s may be called in to perform a close examination of the insects present at
59 the scene. By analyzing the growth stage and community structure of the insects feeding on the corpse, the scientists can often determine the approximate time of death. B y testing insect tissues and exuviae (shed exoskeletons) for foreign chemicals, they may be able to detect the type of toxin administered. If authorities suspect that a stream has been 'poisoned,' ecologists can apply similar sampling techniques and statistical anal yses to draw conclusions about the cause and magnitude of the habitat degradation, and make management and mitigation recommendations (Unlike rivers, however, cadavers offer few possibilities for remediation.) Benthic macroinvertebrates are uniquely suit ed to biomonitoring for several reasons. In general, assessment programs that use living organisms to study habitat degradation are much more effective than chemical testing alone, because these methods will show the actual ecological impact of stressors and can reveal episodic as well as cumulative pollution and habitat alteration (Barbour et al. 1996) Chemical tests of a single sample of stream water provide only a snapshot of the water conditions at that moment, potentially missing a catastrophic even t that happened a few days before (Whiting and Clifford 1983) or underestimating the combined, synergistic effects of small concentrations of toxins in the water over time (Schwarzenbach et al. 2006) However, a standardized macroinvertebrate sample from the same stream (when compared to prior samples from the site or appropriate reference conditions) would be expected to show deviations from the predicted community structure, indicating stress on the system. For example, if a coal sludge impoundment pond burst its dam and overflowed into a river weeks or months prior to sampling, the sediment and toxins might
60 have washed downstream, but irregularities in the aquatic communities would remain and could still be detected (e.g. if a large portion of the sensit ive stonefly larvae population had been smothered or poisoned, fewer stoneflies than expected would be collected). These differences can be analyzed statistically and used to create indices that quantify the magnitude of the disturbance. Though different monitoring programs also use other taxa to measure water quality (e.g. bacteria, protists, fungi, algae [mainly diatoms], macrophytes [larger plants] and especially fish ) macroinvertebrate sampling offers several advantages (Cairns and Pratt 1993) Comp ared to microscopic taxa, insects are easier for non e xperts to identify in the field (Rosenberg and Resh 1993). Macroinvertebrates generally have longer lifespans than microbes and meiofauna, allowing them to act as continuous monitors of the water they inhabit (Rosenberg and Resh 1993) As compared to larger organisms like fish, invertebrates are generally less mobile, especially as larvae (increasing the likelihood that a sample will reflect the conditions of the immediate area, without as much recruit ment from other areas following a major disturbance). Insects are typically more abundant than fish (Allan and Castillo 2007) making it faster and easier to catch enough individuals, in a standardized way, to obtain statistically significant results. Vert ebrate research also requires more stringent permits (including IACUC permits), and some rare fish species are under harvest restrictions (see Plotnikoff and Wiseman 2001) whereas fewer invertebrates are covered by these types of protections (with the exception of some freshwater mussels).
61 Howe ver, as Rosenberg and Resh (1993 ) point out, there are difficulties inherent in using benthic macroinvertebrates as indicators of water quality. Most importantly, monitoring protocols must control for confounding fa ctors (besides pollutants and habitat alteration) that could affect insect communities. T hese variables will be discussed later in more detail, but include natural habitat differences, seasonal changes, regional variations, and insect movement Taxonomic r esolution and volunteer training must also be taken into consideration when designing protocols for benthic macroinvertebrate identification. 1.4.4 How are benthic macroinvertebrates used to measure water quality? With no prior knowledge or preparation, wading into a random body of water, turning over a rock, and plucking out a squirming thing with lots of legs will tell absolutely nothing about the ecological condition of that stream. Biomonitoring data is only useful in context and for this reason, it is esse ntial that benthic macroinvertebrate sampling is done in a thoughtful, organized way, as part of a broader program. There are different types of biomonitoring programs for various uses, and there are many sampling protocols that have been developed for spe cific conditions and regions. Historically, early frameworks for connecting macroinvertebrate diversity and abundance to pollution levels were largely qualitative and focused on the use of indicator species. The Saprobien system developed in Europe by Kol kwitz and Marsson in 1909 related lake pollution to organic matter contamination, and paved the way for the development of the concept of indicator species (which display varying degrees of pollution tolerance), as summarized by Cairns and Pratt (1993) and Carter et al. (2007) Reynoldson et al. (1997) also discussed the evolution of the use of comparative reference
62 conditions in stream monitoring programs. In recent large scale work, if a group such as a state environmental agency decides to develop a new benthic macroinvertebrate sampling program, the first prerequisite is the availability of accurate, in depth, and specific information about the species in the area, including their pollution tolerance, ecological roles and traits. Next, the agency will co nduct a study of the least impacted (most pristine) rivers and streams in the area, to develop statistical models of how a normal assemblage of insects is expected to look. This data set is called the reference condition and must take into account ecoreg ion, season, stream size, in stream habitat, and other potential confounding factors. After several years of background data collection (at least), impacted streams can begin to be tested and compared to similar reference streams. However, earlier authors did not have access to much of the centralized taxonomic and ecological information that is currently available, such as the impressive USGS database of lotic invertebrate traits (Vieira et al. 2006) Therefore, much of their work focused on methods such a s upstream downstream comparisons, or taking samples from different sides of lakes, to attempt to determine effects of habitat disturbance (Reynoldson et al. 1997) These procedures are still used today, depending on the scale and goals of the project be ing conducted. Sampling regimes can compare a potentially impacted site to a set of reference conditions (spatial comparison i n different water bodies), or to another site upstream of the disturbance on the same stream (spatial comparison within the same w ater body), or to the same site before and after the disturbance (temporal comparison). Overviews of 1980s 1990s research (as cited by Reynoldson et al. 1997) showed that
63 63% of lotic (flowing water) and 26% of lentic (still water) studies used the first t ype of comparison, 15% of lotic and 28% of lentic studies used the second type, and 22% of lotic and 46% of lentic studies used the third type. The temporal comparison is especially useful in Before After Control Impact (BACI) studies, in which the change in habitat function prior to a disturbance is compared to the post impact condition (such as before and after dam construction or habitat restoration projects). Compliance biomonitoring is another type of study, typically involving regulatory agencies moni toring conditions downstream of a point source pollutant, such as a discharge pipe from a f actory (Rosenberg and Resh 1993). The rest of this section will focus on the specifics of data collection and analysis for the first type of large scale, long term biomonitoring programs described above (with a focus on No rth American programs). The methods used to collect and analyze data for this type of program can generally be divided into two types: Rapid Bioassessment Protocols (also known as RBPs or the Community Assessment Approach), which are designed for quicker, easier evaluations, versus the moreintensive protocols designed to collect more in depth data (Lenat and Barbour 1994) In general, the overall goal of both types of protocols is to distinguish be tween acceptable and unacceptable levels of habitat degradation in order to decide whether or not management actions are required This objective is accomplished through collecting standardized data, comparing results to appropriate reference conditions, and using weighted analysis metrics to calculate an index value as the final result. An index value is a single number on a scale describing ecosystem health, w ith ranges designated as Very Poor, Poor, Fairly Poor, Fair, Good,
64 Very Good Excellent etc., and it can be used to prioritize decisions about the need for management, regulatory decisions or remediation. However, protocols arrive at this number in different ways, and the amount of additional information collected varies enormously. Rapid Bioassess ment Protocol s aim to expend the minimum amount of effort required to get reproducible, scientifically valid results and m ost RBPs are designed to go from field collections (3 to 5 sites) to a report in five working days (Lenat and Barbour 1994). This is usually the standard in situations involving limited funding, time, personnel, training (e.g., programs involving citizen and student volunteers), a lack of background regional taxonomic data, and/or limited availability of expert taxonomists. The most widely used RBP was developed by the U.S. E nvironmental Protection Agency and is designed to be broadly app licable across geographic areas (Barbour et al. 1999) Lenat and Barbour (1994) describe four ways in which RBPs reduce th e amount of effort and reso urces required for ecological assessment. RBPs may limit the number of organisms collect ed, by sampling to a target number ( such as picking out the first 100 organisms from randomly selected subsamples of the whole sample ). These protocols may also limit t he type of organisms collected (such as focusing on sensitive EPT taxa only Ephemeroptera, Plecoptera and Trichopt era). These methods may also use qualitative, semiqualitative, or ordinal categories for counting organisms ( by rating abundance on a numeri cal scale or as rare, common or abundant etc.) Finally, RBPS may also reduce the effort involved with data collection and processing by identifying organisms only to order or family as opposed to genus or species.
65 There are many other protocols besides RBPs. Generally, each federal or state agency, or university research team, that uses biomonitoring will make its own regional variations or adapt the protocol of a neighboring area. Notable programs have been developed for use in Wisconsin (Hilsenhoff 1982), North Carolina (with the North Carolina Biotic Index), Maine, Ohio, Tennessee, Kentucky, Virginia, and California, among others, and the U.S. Forest Service also has region specific protocols. More specifics of the data collection process, and some v ariations among these different strategies, will be discussed in the next section. 1.4.5 How are the data collected? The choice of specific sampling methods can have a profound effect on the results and accuracy of a benthic macroinvertebrate sampling program The most important decisions involve the selection of the study area and within stream sampling sites, the target number and type of organisms to be collected, the tools to be used, and the people who will use them. When selecting a suitable study area, it is important to take into account the inherent limitations of aquatic insect sampling protocols. The vast majority of the research that has been done on the subject has focused on shallow (wadeable) freshwater streams, with little available data on very large rivers ( th order), lentic habitats like lakes and ponds, and transitional areas between lotic and lentic habitats (Merrit et al. 2002; Lenat and Barbour 1994). The main problems with larger rivers and slow moving or still waters are technical in nature I t may be difficult or dangerous to collect representative samples from deep, fast moving water, and it is also somewhat
66 harder to collect areaquantified samples in slow moving waters because there is no current to sweep dislodged invertebrates into nets. The size of a stream also has an effect on the richness of taxa it can be expected to support. M any protocols specify a minimum stream width because, for example, streams less than 3m wide have been found to have 20% lower expected EPT (Ephemer optera, Plecoptera, Trichoptera) richness than larger streams, but richness may also decrease in the largest rivers as a function of current speed (Lenat and Barbour 1994). Sampling protocols typically have a correction factor for stream order. Stream orde r is an important measure of relative size in lotic systems. First order streams are small headwaters, and two first order streams can join to form a second order stream. Two second order streams can join to form a third order stream, and so forth Anothe r important factor affecting benthic macroinvertebrate richness and abundance is the time of year when a sample is collected. Taxa such as mayflies, which emerge en masse as adults to breed during a specific time period, are much more difficult to find as aquatic larvae when the majority of the population is flying, or newly deposited as eggs. Many protocols (e.g., North Carolina, Ohio, Maine and Washington) have a recommended seasonal sampling window, which varies according to the regional climate. T he Nor th Carolina Biotic Index is based on sampling from June thro ugh Septem ber, and Lenat and Barbour (1994) recommended adjusting the final 0 10 scale of ecological stream health during winter and spring, by adding 0.2 points to streams in the Piedmon t and Coa stal Plains ecoregions and 0.5 po ints to colder mountain streams. The Washington State protocol has a slightly different sampling period, July 1st through
67 October 15th, in order to allow adequate time for habitat stabilization after spring snowmelt floods (Plotnikoff and Wiseman 2001). Marchal (2005) pointed out that macroinvertebrates of higher pollution tolerance are typically more common during warmer months (when dissolved oxygen values may be lower), and Jason Robinson (an entomology graduate student a t the University of Tennessee Knoxville), offered the somewhat contrasting view that, in general, sampling Appalachian streams in early spring would yield the highest taxa richness, while late summer sampling would yield the lowest richness estimates (Ro binson, pers. comm., 2008). Regardless of the specific temporal changes in taxa richness and abundance, it is important to collect sampling data throughout the year in order to provide a baseline reference condition. Within stream site selection is another factor that varies between sampling protocols. The goal is to minimize habitat differences between the test stream sampling sites and the sites sampled in the reference streams. Stream microhabitats vary between high gr adient and low gradient streams and also among streams with different substrate types (gravel, cobbles, sand, mud, etc.). Riffles, pools, snags, root mats, submerged macrophytes and hyporheic gravel beds are just a few examples of different in stream microhabitats (Frissell et al. 1986) I n vertebrate abundance and capture rates are subject to unavoidable stochastic variation (Lenat and Barbour 199 4). Due to this underlying noise in the data, invertebratederived measurements of stream health may vary by 20 30% among replicates from identic al habitats (Lenat and Barbour 1994 ). Unintentional sampling bias could introduce additional variability, so standardized site selection methods are used to minimize this type of error. Protocols are generally divided into
68 single habitat and multihabitat p rocedures. S ingle habitat procedures almost always sample shallow, erosional, fast flowing riffles M ultihabitat methods may involve separate or composited samples from representative habitats, the most productive habitat, any high current habitat with structure, or the least stressed habitat (Lenat and Barbour 1994). One EPA multihabitat protocol also indicates that composited samples should be taken from different habitats in accordance with the relative proportion of that habitat in the stream (N erbonne et al. 2008) Lenat and Barbour (199 4) also cite a study by Plafkin et al. (1989), which found that 100count single habitat riffle samples included only 3568% of the taxa collected in multiple habitats in the same streams. Once the study area and instream sites have been selected, the actual physical methods and tools to be used for invertebrate collection must be determined Hauer and Lamberti (2007) Allan and Castillo (2007) McCafferty (1981), and Lehmkuhl (1979) offer good overviews of the different types of equipment that are used for collecting aquatic insect samples. Nets are one of the most common tools used in stream ecology. Types of nets include sweep nets (D frame nets, A frame nets, and dipnets), as well as rectangular kicknets. Sp ecific procedures vary among protocols but often involve a set number of sweeps in a specific range of habitats, areadelimited searches, or time delimited visual encounter surveys. Tools such as Surber samplers and Hess samplers restrict the sample site to a region of known surface area on the substrate. Macroinvertebrate density and abundance can be quantified if areadelimited samples are taken. T his can be accomplished more generally with other kinds of nets by disturbing a certain area upstream of the net through kicking, handwashing of rocks, or the use of a
69 brush to dislodge particularly well attached insects. One of the few viable methods for large, unwadeable rivers is the use of artificial substrates. These samplers come in two main varieties, th e basket type and the multiplate or Hester Dendy sampler. The use of basket type substrates involves bundling up a collection of brick fragments, soaked wood, and other suitable material into a manageable size, perhaps using pou ltry netting to hold it tog ether It is placed on a river or lake bottom for a number of days until it is colonized by the surrounding insect community. The sampler substrate is then retrieved with a rope or wire previously attached, often collecting rare an d unusual taxa (Lehmku hl 1979). Hester Dendy collectors are more easily standardized and can be adjusted to fit the size of the habitat being sampled by changing the number, area or spacing of plates that are attached to a long central eyebolt (FL DEP 2006) One nearly ubiquit ous piece of field equipment is shallow white pans, because dark colored, cryptic insects stand out so well against the br ight background (McCafferty 1983 ; Lehmkuhl 1979). Some protocols also use ice cube trays for initial field sorting of invertebrates in to different taxonomic groups (Robinson n.d.) Many monitoring agencies enlist the help of volunteers, students, community members, citizen scientists and other nonprofessionals in conducting field samples of benthic macroinvertebrates. Nerbonne et al. (2008) states that participation in monitoring, can be an engaging activity that teaches local residents about effects of anthropogenic pollution on aquatic communities, and tested the effects of volunteer training through an experiment in Minnesota stre ams. The study by Nerbonne et al. (2008) cites results from Reynoldson et al. (1986) in which data collected by volunteer high school students was
70 found to be more variable than those collected by professional biologists, but the final results showed simil ar trends in water quality. Additionally, as volunteers in a separate study received more training, the results from their samples more closely resembled data from fisheries biologists (DiStefano 1999, as cited in Nerbonne et al. 2008). Nerbonne et al. (20 08) also offered interesting perspectives on why volunteers might skew the results from an EPA multihabitat sampling protocol: "volunteers (perhaps fearful of what might be lurking in the vegetation or woody debris) attempt to follow the EPA multihabitat protocol, but sample mostly in the riffle and run areas" or (perhaps excited about the potential for finding crayfish and other larger invertebrates) [...] sample mostly in wood y debris and vegetated banks." However, any programs utilizing non expert help will generally follow simple procedures often incorporating cross checks and periodic validation tests to ensure the accuracy of the data (Lenat and Barbour 199 4). Once samples have been collected, invertebrates are identified in the field or preserved for lab identification. Sorting in the field offers the advantages of using less battered specimens that have not degraded in color Living, moving invertebrates may also be easier to distinguish among debris. However, lab identification provides a more co ntrolled setting, with dissecting microscopes and entomological refer e nces, allowing greater inclusion and more precise identification of young instars and smaller taxa (especially midge larvae). Preservation in 95% to 70% ethanol is important if specimens are to be kept in the same container for any appreciable length of time, as confining live predators in jars with prey results in sample degradation (Plotnikoff and Wiseman 2001) (Watersnipes, a type of larval dipteran, can be especially ferocious in the se situations,
71 from personal observations.) Maintenance of preserved reference collections of specimens can also help future identification efforts and taxonomic studies. 1.4.6 How are the invertebrates identified? Some of the most debated procedures in any str eam protocol take place after all the samples have been collected. Many papers have been published dealing with taxonomic resolution (the taxonomic level to which the majority of the specimens are identified, for example or der, family, genus or species) (L enat and Barbour 1994; Bailey et al. 2001; Lenat and Resh 2001; Merritt et al. 2002; Arscott et al. 2006; Herbst and Silldorff 2006; Carter et al. 2007) Briefly, most authorities believe that identification to order or family level is justified when: Reso urces are limited (these could include funding, time, personnel, and availability of taxonomic experts) Genus or species level taxonomy is poorly known, or is in dispute There is little within family diversity; e.g., in areas that are recently glaciated, long anthropogenically disturbed, have low habitat diversity or naturally low richness (Lenat and Resh 2001) [T]hat the objective is to communicate, and that taxonomy should not be a secret code for a handful of conspirators (Lehmkuhl 1979) Selecting co arser taxononomic levels may increase perceived community similarity but provides a common currency for conversations among scientists and nonscientists (Bournard et al. 1996)
72 However, many sources say that identification to genus or species is more appro priate when: Assemblages of indicator species are being used to estimate pollution levels (tolerance levels often vary within f amilies) (Lenat and Barbour 1994) Physiology, toxicological response, population dynamics or secondary production of a particular species or group of species is used (Bailey et al. 2001) The study is conducted in taxonomically rich areas that have undergone high adaptive radiation, leading to greater diversity within famili es (Lenat and Resh 2001; Bailey et al. 2001) Attempting to differentiate between similar sites along a disturbance gradient where the expected difference is small (Arscott et al. 2006) Identifying high quality water r esources (Lenat and Barbour 1994 ) [I]t also satisfies the soul of the naturalist that lives within most benthologists to be able to name the species [...] (Bailey et al. 2001) Compromises do exist between these extremes. Id entification to mixed taxonomic levels is often referred to as using the best available taxonomy or identifying to the lowest practicable level (Lenat and Resh 2001). Non insect taxa such as oligochaete worms, as well as some insects such as chironomid midge larvae are often groups that are left at higher taxonomic levels because they are difficult to identify. However, the high abundance, wide distribution, and variable pollution tolerance of the Chironomidae can be used to great effect by experts in midge identification (Armitage et al. 1995) Bailey et al. (2001) propose a two tiered approach for bioassessment studies, using f amily level
73 identification for multivariate analyses or index calculation, and species identification for a short list of indicator taxa that are appropriate for a particular study. Several studies have provided quantitative analysis of the effort involv ed in identifying specimens to higher taxonomic res olution. Lenat and Barbour (1994) found that with the EPA RPB family level protocol, the reduction in processing time was mostly for personnel with limited experience, as taxonomists with 3 4 years experie nce could identify most organisms to the species or genus level almost as quickly as the family level. Hilsenhoff (1982) found that combined field and lab time to arrive at a genus based biotic index was only an hour per stream, and identifying to species level took just 21 minutes longer on average. Herbst and Silldorff (2006) compared three different sampling protocols and found that the most rigorous one, the CA SNARL (University of California Sierra Nevada Aquatic Research Laboratory), took 1.5 3 times as long to process in the lab as the other two. Perhaps the most pointed observation comes from Bailey et al. (2001): Skilled taxonomic technicians often argue that the incremental costs to identify a group of invertebrates to genus rather than family are quite small, but the more pertinent question is whether real cost savings are possible by reducing the need for highly skilled taxonomic experts. In addition to their taxonomic classification, invertebrates can also be sorted by their pollution toleranc e (indicator status), habitat preference (Functional Habitat Group, or FHG), and trophic level (or Functional Feeding Group, FFG), often as part of the multimetric analyse s that will be discussed in the next section. Unfortunately, little research has been done on the question of how to incorporate invasive species in to
74 benthic macroinvertebrate analyses Merrit et al. (2002) state that The entire issue of using alien species in calculating these [functional group] ratios needs further assessment, especia lly as they are currently usually excluded from calculations entirely. F unctional feeding groups (FFGs) divide benthic macroinvertebrates into ecologically meaningful guilds based on their food preferences. Although these characteristics may vary by ins t ar stage and food availability, FFG d esignations are often used in calculating ratios to analyze habitat characteristics. Cummins et al. (1973) is credited with the first classification of insects by feedin g group, and Merrit et al. (2007) designated the f ollowing groups (Table 1.1) : Table 1. 1: Functional Feeding Groups, from Merritt et al. (2002) Shredders Feed on CPOM (Coarse Particulate Organic matter, >1mm diameter, either live aquatic macrophyte tissue or coarse terrestrial plant litter) Scrapers Harvest periphyton and associated particulate material from substrate surfaces, primarily live plant stems Filtering and gathering collectors Detritivores; c ollect FPOM (Fine Particulate Organic Matter, <1 mm di ameter) Plant piercers Imbibe cell fluids from filamentous macroalgae and vascular hydrophytes Predators Capture live prey Merrit et al. (2001) also assigned benthic macroinvertebrates to Functional Habit Groups (FHGs), as shown in Table 1.2 (these agr ee closely with those from McCafferty 1981, though McCafferty presented very different FFGs):
75 Table 1. 2: Functional Habitat Groups, from Merrit et al. (2001) Clingers Have morpho behavioral adaptations to res ist dislodgement Climbers Have opposable legs or suction devices used in moving vertically up and down plant stems Sprawlers Have adaptations for staying on top of fine sediments, or are supported by floating plant leaves with no special adaptations Bur rowers Live in the sediment interstices or in constructed tubes within the sediment Swimmers Move in short bursts between resting locations, usually on submerged portions of aquatic vascular plants Skaters/divers Are associated with the surface film, inc luding some that regularly dive beneath the surface Planktonic forms Are suspended in the water column with little ability to move significant distances by their own locomotion S ome RBP protocols use FFGs and F FHs to simplify invertebrate classification usually by grouping families that share similar ecological traits. In contrast, Poff et al. (2006) studied these traits and others to design a more detailed analytical approach. Poff et al. (2006) used a comprehensive USGS metadatabase of North American invertebrates which listed the specific trait states displayed by each species for each of the traits studied. These traits are shown in Table 1.3, and each trait has multiple possible states that are not listed in Table 1.3 :
76 Table 1. 3: Traits from USGS meta database (Poff et al. 2006) Life history Voltinism (generations/yr) Development Synchronization of emergence Adult life span Adult ability to exit Ability to survive dessication Mobility Female dispersal Adult flying strengh Occurrence in drift Maximum crawling rate Swimming ability Morphology Attachment Armoring Shape Respiration Size at maturity Ecology Rheophily (depositional vs erosional vs both) Thermal preference Habit (burrow, climb, sprawl, cling, swim, skate) Trophic habit (collector gatherer, collector filterer, herbivore [scraper, piercer, and shredder], predator [piercer and engulfer], shredder [detritivore]) The use of such sophisticated and detailed analyses can increase the value of genus and species level identification data collected in biomonitoring samples. Specific knowledge of exactly how each different ecological guild of insects is impacted by habitat degradation can help pinpoint the causes of disturbance and assist in addressing these issues. 1.4.7 How are the data analyzed? Common metrics that can be calculated from raw abundance and identification data include community structure metrics, such as the taxa richness (# of species, S), similarity coefficients, and EPT taxa richness ( which has been shown to be less variable than total taxa richness year to year). These community structure metrics can be qualitative (e.g. the Jaccard Index) or quantitative (e.g. the Pinkham Pearson Index). Maine protocols and the EPA RBP use the Communi ty Loss Index developed by Courtemanch and Davies (1987) while Arkansas and North Carolina protocols use
77 variations on the Common Taxa I ndex and Common Dominants Index (Lenat and Barbour 1994). Other metrics require additional information from natural hi story studies or databases These analyses include community balance metrics (such as evenness) and diversity indices (such as the Shannon Diversity Index) that take into account species richness and evenness. Other metrics include: % contribution of the d ominant taxon, EPT abundance/Total abundance, EPT abundance/Chironomid abundance, Hydropsychidae abundance/Total Trichoptera, Baetidae abundance/Total Ephemeroptera, Tanytarsini abundance/Total Chironomidae, and others (Lenat and Barbour 1994). Tolerance m etrics often require species or genus specific information, and are usually based on the percentage of intolerant species in the sample, by richness or abundance. Ecological metrics based on feed ing groups are also very common, as described by Merritt and Cummins (2007) and Robinson (n.d) (Table 1.4):
78 Table 1. 4: Ecological metrics for benthic macroinvertebrates Production Respiration Ratio (P/R) Ratio of Scrapers to ( Shredders and Collectors ) P/R > 0.75 suggests an autotrophic stream (net production of organic matter) P/R < 0.75 suggests a heterotrophic stream (net consumption of organic matter) Allochthonous (external) Input Index (CPOM/FPOM) Ratio of Shredders to ( Collectors and Filterers ) Heterotrophic systems relying primarily on leaf and wood inputs (shredder streams) will have spring and summer values > 0.25 (autumn and winter values > 0.5 ) Streams which have little riparian vegetation may have fewer organisms which rely on allocthono us materials Fine Particulate Organic Matter Transport (TFPOM/BFPOM) Ratio of Filterers to Gatherers 'Natural values in small to medium sized streams should be > 0.5 Enrichment of streams with FPOM (fine particulate organic matter) may result in be nthic storage of these materials, potentially contributing to an increased BOD (biological oxygen demand) and diminished habitat value Substrate (Channel) Stability Index Ratio of ( Scrapers and Filterers ) to ( Shredders and Gatherers ) This metric evalu ates the availability of stable surfaces and substrates for benthic macroinvertebrate habitat (values should be > 0.5 ) Top Down Control Ratio of Predators to all other groups If predators are controlling macroinvertebrate distributions, value would be greater than 0.15 (Normal range 0.10 0.20)
79 More advanced, complex methods of multivariate statistical analysis include ordination such as correspondence analysis and classification techniques such as clustering (Gotelli and Ellison 2004) Brown et al. (2009) use d nonmetric multidimensional scaling, a type of ordination, to analyze macroinvertebrate assemblage response to urban gradients. Maloney et al. (2009) compared the ability of five different modeling methods to predict the benthic macroinvertebrate Index of Biotic Integrity from land use and land cover variables and landscape measures. Their research focused on Maryland streams and involved the use of classification and regression trees, conditional inference trees, random forests (RF), conditi onal random forests (cRF) and ordinal logistic regression (Maloney et al. 2009). Ramrez et al. (2009) use d cluster analysis to partition urban, agricultural and forested stream sites in Puerto Rico based on their macroinvertebrate assemblages. Caskey et al. (2010) use d a variety of different statistical ana lysis techniques to examine breakpoints in the relationships between macroinvertebrate and fish community response metrics and various urban stream stressor variables. Carter et al. (2007) compare d vario us multimetric and multivariate approaches for analyzing benthic macroinvertebrate biomonitoring data. Gotelli and Ellison (2004) provide detailed information about the theory and practice of ecological statistics. 1.4.8 How are the data used? B enth ic macroinvertebrate community data can be used for a wide range of applications. Statistical analyses that create indices of ecosystem health are especially useful for making management and regulatory decisions about hydrological projects,
80 remediation and possible legal action against wastedischarging industries that violate permit specifications. More complex multivariate analyses based on long term data sets with specimens identified to genus or species are excellent sources of information for answering ecological questions about trophic interactions (including food availability for fish, an important value for fisheries biologists and fishermen) and stream nutrient dynamic s. 1.4.9 Benthic macroinvertebrates in urban streams Throughout the available literature, t he most consistent ly reported effects of urbanization on benthic macroinvertebrates are changes in community composition. As described by Walsh et al. ( 2005b ) : Assemblages of highly degraded streams within urban catchments are numerically dominated by a few spec ies of oligochaetes (typically tubificids, lumbriculids, and naidids) and chironomids. We know of no studies where any other pattern has been reported. Lambert i and Berg (1995) found consistent results in South Bend, Indiana, reporting that, chironomids accounted for up to 80% of the total aquatic insect secon dary production in Juday Creek prior to restoration. The relative abundance of macroinvertebrates belonging to shredder functional feeding groups often decreases with urbanization (Walsh et al. 2005b ) This specific reduction is likely associated with decreases in leaf pack input and retention (Borchardt 1993; Aldridge et al. 2009) Excess sedimentation increases the embeddedness of cobble (rocks) in the stream substrate filling in interstitial stre ambed habitat s and negatively affecting riffle invertebrate communities (Lamberti Berg 1995; Roy et al. 2003) Urbanization may also affect dispersal, drift and seasonal variation in
81 macroinvertebrate assemblages. Smith et al. (2009) examined urban effect s on the dispers al of adult macroinvertebrates, and suggest that anthropogenic activities have the potential to prevent the completion of aquatic insect life cycles and to limit adult dispersal, and therefore, can affect population persistence. Borchardt (1993) found that high flows can dislodge macroinvertebrates, increasing drift lost downstream; this could increase dispersal but could also disrupt community structure. Brown et al. (2009) found increased frequency and magnitude of high flow events in u rbanized areas, primarily because of increased impervious surface associated with urbanization but suggest ed that this pattern is not universal and can be affected by st ormwater management practices. Walsh et al. (2009) propose d that effective imperviou sness, or hydrologic connectivity, of the catchment is more detrimental to stream ecosystems than simple total impervious surface area. Moerke et al. (2004) reported unexpectedly rapid recolonization of macroinvertebrates following channel disturbance caus ed by a restoration project on Juday Creek (South Bend, IN). Macroinvertebrate densities in the restored reaches recovered to densities comparable to the unrestored (undisturbed) upstream section within 14 days, and diversity recovered by 120 days post res toration (Moerke et al. 2004). Moerke et al. (2004) compare d these results to studies in other areas, in which macroinvertebrate communities required 70 150 days to recover in terms of density, and 250 days to 2 years to re establish prior diversity. S easo nal factors may have influenced the impressive recolonization of Juday Creek because the restoration took place in the early autumn, a time of active drift and oviposition by aerial adults (Moerke et al. 2004). However, urbanization may alter seasonal patterns of macroinvertebrate activity.
82 In a study of streams in western Georgia, Helms et al. (2009) found that the influence of urban iz ation on macroinvertebrate communities appears to be consistent throughout the year and reduces seasonal changes in asse mblages. The amount of alteration in invertebrate assemblages, as compared to reference conditions, has been shown to vary with the extent of urbanization. Moore and Palmer (2005) studied 29 headwater streams in Maryland and found that macroinvertebrate t axa richness was related negatively and linearly (no statistical threshold) to the amount of impervious surface cover in the catchment. Brown et al. (2009) also found a negative linear correlation between macroinvertebrate metrics and urbanization metri cs, in a combined standardized study of nine metropolitan areas across the U.S. This research also found no evidence for threshold effects, indicating that stream communities exhibited no init ial resistance to urbanization (Brown et al. 2009). This resul t means that there appeared to be no level of urbanization that had no discernible negative effect on fish, algae, or invertebrate assemblages (Brown et al. 2009). Prior land use was an important factor in both of these studies and is also discussed by M aloney et al. (2008) and Moerke and Lamberti (2006a ) Brown et al. (2009) suggest ed that pre urbanization agricultural impacts could reduce the resilience of macroinvert ebrate communities as compared to sites that were converted from forest to urban or su burban use. However, Moore and Palmer (2005) found that the agricultural sites in their study actually supported high macroinvertebrate diversity, in contrast to the general trend reported in the literature. After consulting with county soil conservation district officials, they attributed these findings to high local implementation of best management practices,
83 i ncluding grassed waterways, vegetative filter strips, contour farming, nutrient management, manure storage, rotational grazing, and no till farming (Moore and Palmer 2005). Moore and Palmer (2005) also reported that Urban streams with high amounts of intact riparian forest cover exhibited [macroinvertebrate] biodiversity levels more comparable to less urban areas despite high amounts of impervious cover in these catchments. This result supports the hypothesis that effective imperviousness has a greater effect on stream ecosystems than total imperviousness (Walsh et al. 2009). Vermonden et al. (2009) found that urban drainage systems in the Nether lands can provide important habitat for macroinvertebrates, including red list (uncommon) species, and that these areas can sustain a macroinvertebrate biodiversity comparable to that of drainage systems in rural areas (ditches and canals) and (semi)natur al watercourses These results suggest that, with certain patterns of land use and stormwater management, urban areas may be able to support diverse and functional macroinvertebrate assemblages. 1.4.10 Summary In summary, biomonitoring of freshwater streams thr ough benthic macroinvertebrate sampling is an emerging field that offers significant advantages over physiochemical monitoring alone. The data collection process can be adapted to varying levels of sophistication required to answer different types of water shed management and aquatic ecology questions, and many protocols are specifically designed to engage students and community volunteers, allowing residents to feel more connected to their watershed and hopefully more conscious of activities that impact eco system health.
84 Chapter 2: Project design 2.1.1 Overview This project was conducted in South Bend, Indiana from May 31 through August 7, 2010, as part of the Summer 2010 GLOBES REU program at the University of Notre Dame. GLOBES, or Global Linkages of Biology, the Enviro nment, and Society, is an interdisciplinary Ph.D. training program at Notre Dame which is funded by an Integrative Graduate Education, Research, and Traineeship (IGERT) grant from the National Science Foundation (NSF). Specifically, this research was condu cted with support from NSF IGERT training grant #0504495. I applied to this REU program and was accepted to work with GLOBES fellow Patrick Shirey, a graduate student in the Lamberti Stream Ecology Lab, on a project broadly titled, Ecological response t o environmental change: science and policy. My specific summer research goals were to investigate urban influence on stream ecosystem pr oduction in South Bend, IN and develop field and lab skills, including measurement of stream features, deployment of Hydrolab MiniSondes use of ArcGIS and data analysis (initial project descriptions by Anderson 2010) 2.1.2 Site history and descriptions This summer project was intended to complement the Lamberti Lab's ongoing research on Juday Creek by extending their lo ng term monitoring regime to the nearby Bowman Creek. Juday Creek flows through the Notre Dame campus and joins the St. Joseph River which eventually flows into Lake Michigan (Fig. 2.1) The Lamberti L ab has studied this stream for 3 0 years collecting ma croinvertebrate samples since 1981
85 (Lamberti and Berg 1995) Notre Dame researchers have also studied changes in fish and macroinvertebrate populations before and after a 1997 habitat improvement effort that restored a 1 km channelized section of Juday Cre ek to a more natural, meandering condition (Moerke et al. 2004) T he City of South Bend Public Works Department is currently planning to restore sections of Bowman Creek, one of the most ecologically impaired tributaries to the St. Joseph River (Gilot 2009; Deegan 2010) Bowman Creek flows through underground culverts fo r over half its urban, downstream length, and the city plans to restore some of these sections to a more natural aboveground condition.
86 b) a) Figure 2. 1: a) Juday Creek ( above ) and Bowman Creek ( below ) are tributaries in b) the St. Joseph River watershed Maps from Google Earth and USGS
87 Bowman Creek has been subject to extensive historical and ongoing ecological impairment. The stream's headwaters drain mainly agricultural and residential lands, followed by a stretch of wooded land which includes the AM General Test Track, where trails for testing Hummers and other large SUVs cross back and forth over the creek. (We asked AM General representatives for permission to sample on their property but were unsuccessful.) Bowman Creek then flows through an urban area contaminated from former defense manufacturing plants, paint factories, a vacant abattoir, a coal dismantlement facility, and residential dumping (Gilot 2009) The stream flows underground through pipes and culverts for over half its journey from the edge of South Bend to its juncture with the St. Joseph River. Portions of Bowman Creek have run dry during droughts creating barriers to fish movement. The stream also floods during rain events, bringing large amounts of trash and debris into the channel and onto the banks. Flooding is exacerbated by input from the city's combined sewer outflow (CSO ) system. This combined sewer system is the result of ad hoc additions to storm sewers originally constructed in the early 1900s (City of South Bend 2010) During dry weather, sewage is routed to treatment plants, but during wet weather, mixed stormwater and sewage flow over low barriers within the pipes and are discharged into streams and rivers (Fig. 2.2). Residents have also experienced sewer backups into baseme nts. E. coli concentrations are often a problem in Bowman Creek and the St. Joseph River after rain events (Gilot 2009; City of South Bend 2010)
88 Figure 2. 2: Combined sewer outflow, from City of South Bend (2010) The Public Works Department of the City of South Bend is currently planning to under take a restoration project involving several sections of Bowman Creek. Their stated goals are to ensure that water flows through the entire length of the creek to improve water quality, and minimize underground flow through pipes, minimize the amount of creek water that enters the city sewer system, enable the creek to sustain macroinvertebrate communities, increase the use of the creek as an educational tool, enhance the aesthetic appeal to surrounding neighborhoods, and minimize the risk of
89 flooding ( Gilot 2009) The process of excavating and restoring buried urban waterways is also known as daylighting the stream. The Bowman Creek restoration is part of a larger citywide effort to comply with federal Clean Water Act mandates. Goals of this broader project include: additional separations of storm and sanitary sewers, expanded capacity to retain storm water and a greater emphasis on green solutions, which address storm drainage through natural solutions on site (City of South Bend 2010) To achieve these goals, the city has contracted with the ecological firm JF New. As quoted in the June 2010 River Report from the City of South Bend, Mayor Steve Luecke said that JF New is training our engineers regarding creative, natural and cost effective solut ions, such as rain gardens, bioswales and other best management practices to capture storm water near the source rather than transport it. We anticipate this sustainable, low impact development approach will save money, enhance our environment and position South Bend as an ecological leader. T he citys overall focus on whole catchment distributed stormwater management solutions is consistent with recent research based restoration recommendations (Walsh et al. 2009; Palmer et al. 2010) However, certain asp ects of the project are projected to increase stormwater discharge to Bowman Creek. Specific proposed discharge locations include a section near Haney Ave. and Dubail Ave, where stor m water will be discharged naturally through Bowman Creek and another s ection near High St., where surface runoff from Riley High School could be discharged into Bowman Creek (City of South Bend 2010) These direct connections between stormwater pipes and the stream would increase the effective imperviousness of the catchme nt meaning that even small rainfall
90 events can produce sufficient surface runoff to cause frequent disturbance through regular delivery of water and pollutants (Walsh et al. 2005a ) Effective imperviousness has been found to be the variable most highly correlated with urb an stream ecosystem impairment (Walsh et al. 2009) If feasible at these sites, stormwater management options that promote retention and infiltration rather than direct discharge to the stream could be better options for improving ecosys tem function in Bowman Creek. 2.1.3 Project goals, questions, and hypotheses This project started with two main goals. First, the research would provide a pre restoration ecological assessment of Bowman Creek, describing a baseline condition that could then be compared to the post restoration condition. This comparison would assist in evaluating the specific effects of the city's restoration effort and help gauge its overall success in improving the ecological health and habitat value of the stream. Second, be cause Bowman Creek flows through so many long stretches of underground culvert, we chose to design the project in a way that could help determine the specific effects of these buried segments on stream metabolism and ecosystem integrity. Broadly, we plan ned to collect data from sites above and below culverts of different lengths, in order to determine: 1) whether the presence of culverts affected the stream ecosystem and if so, 2) whether the magnitude of the impact varied with the length of the culvert. In deciding what types of data to collect, we first asked: What does a culvert actually do to a stream? When a section of stream is rerouted through a culvert, the water is essentially enclosed by concrete on all sides. A concrete ceiling shuts out the su nlight, plunging the waterway into darkness. Concrete walls replace the natural banks,
91 riparian vegetation, and streamside soils with smooth, hard surfaces A concrete floor places a barrier between the stream and the underlying groundwater, interfering wi th hyporheic flow (Dahm et al. 2007) and nutrient processing dynamics (Groffman et al. 2005) Underground flow can affect stream ecosystems through two main mechanisms: lack of sunlight and channel alterations. We hypothesized that the lack of sunlight re aching the stream will prevent photosynthesis from occurring in the underground areas. Thus, sections of the stream flowing through dark culverts will support few to no autotrophs (algae and plants), and little to no primary production (carbon fixation by in stream autotrophs) will occur in these reaches. We decided to quantify these effects by following the protocol developed by Dr. Jennifer Tank's lab at Notre Dame (Griffiths et al. 2010, unpublished standard operating procedure based on Grace an d Imberge r 2006). This method uses deployable dataloggers to collect in stream data on water temperature and dissolved oxygen as well as an on shore meter to record light levels. These devices record data at ten minute intervals over the course of several days. Die l fluctuations, along with stream flow information are fed into a computer model. The model calculates gross primary production (GPP), ecosystem respiration (ER) and reaeration (k). D ataloggers w ere placed upstream and downstream of culvert s in order to c ompare any differences in metabolic output due to the underground flow. We also decided to collect several other types of data to refine and expand our understanding of ecosystem function at different sites along Bowman Creek and to determine whether the culverts had a measurable effect. The deployable dataloggers were
92 also able to record daily fluctuations in pH and specific conductivity during their time in the stream, in addition to the temperature and dissolved oxygen data required for the model. Samp les of small rocks were taken at each site so that the composition of the biofilm could be analyzed for chlorophyll a co ntent and ash free dry mass. These measurements would allow us to compare the relative heterotrophic and photosynthetic activity in the biofilm at each site. The abundance and community composition of benthic macroinvertebrates could also be used to give a general indication of secondary productivity (Benke and Huryn 2007) However, a rigorous quantification of secondary productivity would require lab processing outside the capacity of this project (measuring the length of each insect specimen and obtaining weight values by using published lengthmass regression tables) (Kohlhepp and Hellenthal 1992; Lamberti and Berg 1995; Entrekin et al. 2007) Water samples were a lso taken for chemical analysis to determine the amount of dissolved organic carbon as well as ammonium, nitrate and soluble reactive phosphorus These measurements could assist in determining limiting factors for stream ecosyste m function and help assess water quality. Channel alterations due to underground flow through culverts can also affect stream ecosystem function. These changes could include altered hydrography (water velocity, reaeration, etc.), altered water flow and nu trient exchange with the hyporheic zone/groundwater, changes in streambed composition and/or changes in the retention of organic m atter. Hall and Tank (2005) Groffman et al. (2005) Mulholland et al. (1997) and Fellows et al. (2001) provide more informat ion on hyporheic flow and its contributions to whole stream metabolism. Aldridge et al. (2009) Lancaster and Hildrew
93 (1993) and Borchardt (1993) discuss ed the importance of debris retention in maintaining refugia spatial habitat complexity, and ecosyste m metabolic function, as well as the detrimental scouring effects of flashy urban discharge Baker et al. (2004) provides additional information and metrics for studying flashy hydrography in Midwestern streams. While Ehrman and Lamberti (1992) and Cordova et al. (2008) have researched retention, large woody debris, and flow characteristics in Juday Creek and other stream systems, we decided to focus on the net effects of these habitat alterations on stream biota. (While quantifying these intermediate mecha nisms would be valuable research, these studies would be more useful after a significant biological effect had been detected, as a means of further investigation and explanation.) To do so, we planned to sample community compositions of indicator taxa: dia toms and benthic macroinvertebrates. Biological samples would be collected, identified, counted, and analyzed using published metrics to create an index value of stream ecosystem health at each site. 2.1.4 Timeline The first month at the University of Notre Dame was largely devoted to finalizing the study design and procedures, researching and contacting streamside property owners for permission to access Bowman Creek through their land, gathering supp lies, and calibrating equipment. Due to equipment issues and June thunderstorms, t he bulk of the fieldwork occurred in July and extended through the beginning of August. Final summer presentations were held on August 7.
94 2.1.5 Study sites From upstream to downstream, the study sites chosen along Bowman Creek are abbrev iated as LOCST (Locust Rd .) URB (Urbanskis property) MAIN (Main St.) GOLF (Studebaker Golf Course) DAYTON (Dayton St.) RAV (Ravina Park) and LWE (Lincolnway East) Biofilm samples w ere taken at an additional site, RRCULV, a railroad crossing culvert in the middle of the LWE reach. Dataloggers were not deployed at the RAV site due to insufficient stream depth. The DAYTON site is 290 m upstream of the start of the RAV reach, with no intervening long culverts Data from the DAYTON culvert outflow can ass ist in characterizing upstream inputs to the RAV site. GPS coordi nates are given at the site of Mini Sonde deployment. Fig. 2.1 shows the location of sites along the stream. Table 2. 1 summarizes stream channel characteristics between sites Table 2. 2 provides a summary of other relevant site information. S ites are described in greater detail in Appendix I Table 2. 1: Stream channel length between sites Stre am channel length between sites (in m) Site Open channel Culvert Total %Culvert LOCST (no long culverts in headwaters upstream) URB 2050 55 2105 3 MAIN 2180 230 2410 10 GOLF 355 600 955 63 DAYTON 210 340 550 62 RAV 255 35 290 12 LWE 290 795 1085 73
95 Table 2. 2: Site information Site Location Nearest road crossing GPS Datalogger deployment Samples taken LOCST Locust Rd. Locust Rd. and Jackson Rd. 4137'17.09"N 8616'56.12"W July 30 Aug. 2 Aug. 2, 2 010 URB Urbanski property Keria Tr. and Serbian Ln. 4138'6.01"N 8616'11.10"W July 28 30 and July 30 Aug. 2 July 22, 2010 MAIN Main St. S. Main St and W. Fairview Ave. 4138'53.99"N 8615'5.25"W July 14 16 and July 28 30 July 21, 2010 GOLF Studebaker Golf Course Accessible via E. Ewing Ave or High St. 4139'9.71"N 8614'33.45"W July 30 Aug 2 July 21, 2010 DAYTON Dayton St. E. Dayton St. and High St. 4139'21.67"N 8614'18.95"W July 14 16 n/a RAV site is downstream RAV Ravina Park E Broadway St. and Dale Ave. 4139'32.96"N, 8614'15.92"W n/a DAYTON site is upstream July 23, 2010 LWE Lincolnway East Main St. and Lincolnway East 4139'48.44"N 8614'12.37"W July 14 16 and July 28 30 July 23, 2010
96 Chapter 3: Materials and methods 3.1 Field protocols Note: A more detailed list of field sampling materials and methods is included on the CD attached to the back cover of the library copy of this thesis. Physical data Two different sets of equipment were used to collect physical data on the strea m study sites. Hach Hydrolab MiniSonde 5 deployable programmable dataloggers ( Figure 3. 1) were used to monitor in stream dissolved oxygen, pH, temperature and specific conductivity at ten minute intervals over a period of sever al days at each site. Figure 3. 1: Hach Hydrolab equipment and accessories. From top to bottom and left to right, the pieces of equipment shown in Figure 3. 1 include: protectiv e PVC tube, with one end cap removed (note holes drilled in case and cap to allow water flow); red cable lock; Hach Hydrolab MiniSonde 5 (tube with blue MS 5 sensor end is on the left); clear plastic calibration cap; weighted metal
97 deployment cap; 1 L Nalgene bottle for LDO calibration; serial to serial cable for connecting Surveyor to computer; charger cord for Surveyor; Sonde to serial cable; and Hach Hydrolab Surveyor 4a. In order to model whole stream metabolism, data from the MiniSondes was to be combined with stream discharge measurements and light level readings from Odyssey photosynthetic irradiance loggers (Dataflow Systems Ltd.) ( Figure 3. 2) These programmable, deployable devices measure photosynthetically active radiation (PAR) in microEinsteins per second per square meter. An E instein is defin ed as one mole of photons. The s e data describes the amount of usable light energy ( in the range of 400 to 700 nanometers ) that is available for photosynthesis, or the power per square meter (Hauer and Will 2007) Figure 3. 2: Odyssey photosynthetic irradience loggers and mounting setup. Water samples for chemical testing At each site, three water samples were taken for nutrient testing and three samples were taken for dissolved organic carbon (DOC) testing Stream water samples were filtered through glass fiber filters (grade F) to remove debris prior to filling sample bottles. Samples were transported from the field in a cooler with freezer packs and either
98 refrigerated (DOC) or frozen (nutrients) upon ret urning to the lab. Biofilm: c hlorophyll a samples and ash free dry mass Rock samples were taken for biofil m analysis. Four to six small (approx. 5 cm x 5 cm x 2 .5 cm) rocks were collected from each site, put into Whirl Pak bags, and placed in a cooler wit h freezer packs to be transported to the lab. Benthic macroinvertebrates Three benthic macroinvertebrate samples were collected from representative riffles at each site using a Surber sampler with a 30 cm x 30 cm frame and 1 mm mesh ( Figure 3. 3) Sampling was time constrained to one minute of scrubbing rocks and agitating substrate in the Surber frame, allowing the water flow to sweep dislodged invertebrates into the mesh net. This mesh was then inverted and the sample was rinsed into a white plastic pan. Clinging invertebrates were removed from the net with forceps. S ample s were elutriated, poured through a metal sieve to drain remaining stream water, and rinsed into a Whirl Pak bag with 95% ethanol Samples were transported to t he lab and stored for later identification and analysis.
99 Figure 3. 3: Collecting macroinvertebrates with a Surber sampler. 3.2 Lab materials and protocols Note: A more detailed list of laboratory materials and methods is included on the CD attached to the back cover of the library copy of this thesis Water samples for chemical testing Water samples collected for nutrient testing were processed and analyzed for ammonium, nitrate, and soluble reactive phosphor us concentrations A Lachat QuikChem 8500 Flow Injection Analysis System ( Figure 3. 4) was used to perform these highly sensitive tests. W ater samples collected for dissolved organic carbon testing were filtered and acid stabilize d with 2N HCl. Samples were analyzed for DOC concentrations using a Shimadzu TOC 5000A combustion oxidation analyzer
100 Figure 3. 4: The Lachat QuikChem 8500, in all its great and terrible glory Chlorophyll a samples Biofilm samples were scrubbed from the collected rocks and vacuum filtered onto glass fiber filters ( Figure 3.5) See Steinman et al. (2007) for details Chlorophyll a was ext racted from the filters usi ng methanol ( Figure 3.6) and analyzed with a flu o rometer ( Figure 3. 7). S amples were dried to a constant weight and ashed in a 500C kiln The ashfree dry mass (AFDM) is the mass of the co mbustible organic matter in a sample. AFDM values were obtained by subtracting the ashed mass of the biofilm sample s from the pre ashing dry mass Chlorophyll a concentrations and AFDM results were adjusted for the upper surface area of the rock samples an d used to calculate the Aut ot rophic Index or the ratio of AFDM to chl a. Higher values g enerally indicate more heterotrophic conditions,
101 while lower values indicate more autotrophic conditions. According to Steinman et al. (2007) normal AI values rang e from 50 to 200. See Bergey and Getty (2006) for a discussion of various surface area estimation techniques. Figure 3. 5: Vacuum manifold with filter cups
102 Figure 3. 6: Chlorophyll a extracted from b iofilm samples (shown after analysis) Figure 3. 7: Chlorophyll a analysis using flurometer
103 Benthic macroinvertebrates Equipment for processing invertebrate sam ples is shown in Figure 3. 8. Each benthic macroinvertebrate sample was poured from the Whirl Pak bag into a rectangular plastic container. The bag was rinsed with ethanol to ensure that all specimens were removed. Large pieces of debris (twigs, leaves and rocks) were examined for invertebrates rinsed, and discarded. Readily noticeable invertebrates were picked from the sample usin g fine tipped forceps. Each taxon was sorted into the appropriate section of a divided holding tray ( Figure 3. 9) The holding tray contained sufficient ethanol to cover the specimens. A portion of the ethanol and debris from the sample was placed into an empty petri dish and examined under a dissecting microscope. The debris was s earched systematically, from one side of the dish to the other. Invertebrates were removed and placed into the holding tray. The process was repeated until all of the debris from each sample had been examined. Empty caddisfly cases were not included in counts. Hydroptilid cases were recorded due to high abundance but were not used in calculations. Fragments of invertebrates were not counted, unless the head and half or more of the body was present, or more than half of a headless body. Small oligochaetes were difficult to count due to the possibility of breakage. Halves of small oligochaetes were recorded as half counts. Invertebrates were identified primarily through the use of McCafferty (1983) and Merritt et al. (2008). Specimens were identified to th e lowest practicable level (Lenat and Resh 2001). Most organisms were identified to family or genus. Several beetles were identified to the species level, while leeches were identified only to the subclass
104 Hirudinea. After identification, invertebrates w ere counted recorded, and transferred to a glass sample vial containing 70% ethanol and a paper label. Figure 3. 8: Processing invertebrate samples Figure 3. 9: Invertebrate sample in sorting tray
105 Chapter 4: Results 4.1 Physical data 4.1.1 Equipment issues The Hydrolab MiniSondes presented rem arkable and unique calibration challenges. These deployable dataloggers are very impressive pieces of technology, but their outstanding a ccuracy and capabilities come at the price of a discouraging tendency to develop mysterious errors. Our first test run on Juday Creek provide d an excellent example of this type of issue From July 2 July 5, we deployed one MiniSonde 5 in a culvert where Indiana State Road 933 crosses Juday Creek and placed another MiniSonde approximately 300 meters downstream past two other culverts (Interstate 90 and Cleveland Rd). The results from the first test run were very promising as shown in Figure 4. 1. Both MiniSondes recorded excellent smooth diel respiration curves (changes in dissolved oxygen concentration over the course of the day), and the results showed an obvious and interesting difference between the upstream and downstream site s However, we were concerned that the downstream MiniSonde had lost battery power halfway through the deployment We cleaned the terminals and replaced the internal and external batteries. Calibration problems and intermittent power losses persisted dur ing testing in the lab. Three working MiniSondes were borrowed from another department to conduct another field test in Juday Creek. The downstream MiniSonde from the first test was redeployed side by side with a functional Minisonde. Two additional datal oggers were deployed, one at the SR 933 road crossing above the culverts and the other at a wooded site 2.4 km upstream
106 T he results were clear: as shown in Figure 4. 2, a ll three reliable MiniSondes showed very tight correlation of their diel respiration curves, and the unreliable MiniSonde displayed a very similar pattern from its first deployment in total disagreement with the data that had been recorded by the functioning MiniSonde just inches away. The unreliable MiniSonde a lso lost power midway through the deployment We were quite concerned by these results, for two reasons. First, all three of the (presumably) working MiniSondes had recorded dissolved oxygen curves that were nearly the same the differences would be biolo gically negligible despite being deployed in sites above and below culverts. If our hypotheses about the effects of culverts on stream ecosystems were correct, we would have expected to see some difference between these sites. However, the culverts on Ju day Creek are fairly short. W e decided to send the unreliable MiniSondes in for repair and use the three working MiniSondes to determine whether Bowman Creeks significantly longer culverts would show any detectable effects on stream ecosystem function
107 Figure 4. 1: Promising results from the first test LDO = Luminescent Dissolved Oxygen
108 Figure 4. 2: Further testing revealed that the promising results we re an anomaly caused by equipment malfunction
109 4.1.2 Sonde data by site Figure 4. 2 shows an example of a set of fairly smooth diel respiration curves. Data recorded by the three reliable MiniSondes could potentially be plugged into the whole stream metabolism modeling program and used to determine gross primary productivity (GPP), ecosystem respiration (ER) and reaeration (k), if combined with appropriate data on sunlight and stream velocity. However, t he model will not function prope rly if the curves are t oo jagged and cannot be integrated ( Peter Levi, per s. comm 25 June 2010). D iel respiration curves recorded from Bowman Creek were more variable than those from Juday Creek Data collected at the URB, MAIN and LWE sites from July 28 30 showed jagged fluctuations at the downstream sites and were not suitable for use in the metabolism modeling program or other data analysis The two site open channel method also requires relatively constant flow and a section of stream that is as u niform as possible (Grace and Imberger 2006). Data collected at the MAIN and DAYTON sites from July 14 16 ( Figure 4. 3) were unusable due to an unexpectedly heavy rain event on July 15. Water depth and site accessibility limited our options for datalogger deployment, and c onditions between the monitoring stations on Bowman Creek were heterogeneous. Additionally, t he choice of deployment locations may not have satisfied the requirements for this method of measuring metabolism. The probe separation (the distance between the dataloggers) is a key factor in achieving good results for the two station [open channel] method (Grace and Imberger 2006). In order to calculate the
1 10 minimum, maximum and optimal distances between sites, the me an stream velocity and reaeration coefficient must be measured first. Our first attempt to measure stream velocity using rhodamine WT dye (Bott 2007) was unsuccessful due to a sensor malfunction We had planned to calculate the reaeration constant from die l curves with the modeling software W ithout these parameters we wer e unable to determine whether the separation between sites on Bowman Creek was suitable for the two station method There are other methods of measuring reaeration (Grace and Imberger 2 006). T hese tracer inject ion methods are fairly involved, and were outside the scope of a short term, wide ranging pilot study with many other components. W e decided to focus on recording usable diel respiration curves from Bowman Creek before fine tuning the metabolism model. MAIN and DAYTON July 14 July 16: The MAIN and DAYTON sites were separated by 1.5 km The curves shown in Figure 4. 3 are not as smooth as those from Juday Creek. Lower stream discharge may have contribute d to this difference. The sample period was interrupted by an unexpected heavy rain event on July 15. The effects of this downpour and flash flooding on debris accumulation at the MAIN site culvert were show n in Figure 1. 8 on p. 19. Several interesting observations can be made about these data, although they are not usable for metabolism modeling. The MAIN site is on the upstream edge of town and receives water flow from some of the less urbanized are as of the stream. The DAYTON site is i n the outflow of a long culvert and is downstream of a long series of culverts. It is interesting to note that the water flowing out of the culvert was consistently cooler ( Figure 4. 4) and bet ter oxygenated ( Figure 4. 3) than its upstream cou nterpart Large fish
111 were observed in the DAYTON culvert. Urban streams are typically warmer than their nonurban counterparts (Pouyat et al. 2007) so the culvert may serve as a fi sh refuge during warm, low oxygen summer conditions. Before the rain event, water temperatures at the upstream site were 1 2 C higher than the downstream site ( Figure 4. 4). During the rain event temperatures readings converg ed and remained similar for the next 12 hours B oth sites received precipitation inputs of the same temperature. Increased flow velocity may also have homogenized conditions somewhat throughout the reach. Figure 4. 3: MAIN and DAYTON LDO, July 14 July 16
112 Figure 4. 4: MAIN and DAYTON temperature, July 14 16 LOCST, URB and GOLF, July 30 Aug 2 : The LOCST and URB sites were separated by 2.1 km and the URB and golf sites were separated by 3.2 km. This set of data shows some interesting trends. The water flowing out of the culvert under the Riley High School parking lot and track field (the GOLF site) has higher dissolved oxygen content ( Figure 4. 5) despite having the highest temperature ( Figure 4. 6) and shows only modest variation in diel respiration. The URB site shows even less variation in dissolved oxygen througho ut the course of the day and had sl ightly lower average DO than the tunnel outflow. The LOCST ditch shows strong variation, with levels that decrease to impaired DO levels at night (especially on Aug. 2).
113 Figure 4. 5: LOCST, URB and GOLF LDO, July 30 Aug 2 Figure 4. 6: LOCST, URB and GOLF Temperature, July 30 Aug 2
114 As shown in Figure 4.7, pH values also displayed diel variations. These patterns of d aily fluctuation were smaller than the between site differences in pH. When respiration exceeds primary productivity, CO2 concentrations increase and pH d ecreases (Maberly 1996) Figure 4. 7: LOCST, URB an d GOLF pH, July 30 Aug 2 Light meter data Because we were not able to use the Sonde data for the whole stream metabolism modeling program, light meter data were not used in any analys e s See Figure 4. 8 for a representative fig ure showing the data recorded by an Odyssey PAR logger at the shady, wooded URB site over a period of several days from July 29 to August 6. Note that the amount of available light peaked sharply wh en the sun is directly overhead. A more open, unshade d sit e would show broader peak s because more sunlight could strike the stream
115 surface at an angle. The flat line on August 7th in Figure 4. 8 is due to the fact that the logger was still recording when it was stored in a dark bag in the lab T hese devices require such little energy that they are set to record until the data is downloaded to a compute r or the batteries are removed. They have no external on off switch. Figure 4. 8: Photosynt hetically Active Radiation, in microEinsteins per second per square meter 4.2 Chemical data The following figures display the chemical analysis results for water samples collected at different sites, arranged from upstream to downstream. S amples were collect ed on differ ent dates, and this information is listed in Table 2. 1, p. 95. F igures 4.9 4.14 show mean nutrient concentration values (of the three samples analyzed at each
116 site), along with error bar s displaying the standard error of the mean. 4.2.1 Nutrient data. Water samples were analyzed using a Lachat QuikChem 8500 Flow Injection Analysis System. S amples from this pr oject were analyzed in the same session as Lamberti Lab samples from oligotrophic sites in Alaska, so the equipment was calibrated for low nutrient levels. The highest calibration standards were 250 gN/L (for ammonium and nitrate) and 40 gP/L (for soluble reactive phosphorus). Although many of the Bowman Creek samples exceeded these r eference concentrations, the calibration curve is typically linear, even to higher concentrations (Peter Levi, pers. comm. 9 August 2010) The lab also tests agricultural water samples, which often contain much higher concentrations of nutrients than tho se found in Bowman Creek (Mike Brueseke, pers. comm.). Ammonium The ammonium data ( Figure 4. 9) display one especially striking result: an enormously high concentration at the LOCST site. The mean recorded value of the three LOCST samples (which were in close agreement) was 2095 gN/L, more than thirty times that of the next highest reading downstream (68 gN/L, at the URB site). In comparison Moerke and Lamberti (2006b) found concentrations of 3.8 to 63.6 gN/L ammonium (mean = 2 7.8 gN/L) in a study of 22 streams in southwestern Michigan. The LOCST site is the fa rthest upstream that was sampled in this project, and may be receiving agricultural runoff or possibly septic tank inputs from reaches that were not wellcharacterized by our study. The total (NH3NH4 +) N concentration at the LOCST site was 2.1 mg/L on
117 Aug. 2, exceeding the 2009 Draft EPA chronic criterion of 1.8 mg/L of ammonia (at pH 8 and 25 C) for freshwater sites without mussels present (U.S. EPA 2009) However, the equilibrium between toxic free ammonia (NH3) and less harmful ionized ammonium (NH4 +) is highly dependent on pH and temperature. During the deployment period, the pH at the LOCST site was between 7.2 and 7.3 ( Figure 4. 7 ) and midd ay high temperatures were ~24 25 C ( Figure 4. 6) At this pH, the formation of NH4 + is favored and the proportion o f free ammonia is small. More information is needed to assess pH variation at the LOCST site, because at higher lev els ammonia concentrations could present a hazard for aquatic life at this location. For comparison, pH at the GOLF site downstream varied from 8.0 to 8.2 during the same period ( Figure 4. 7) The error bars in this graph are very small and may be difficult to see. Figure 4. 9: Ammonium in Bowman Creek
118 Nitrate Figure 4. 10 shows nitrate concentrations in Bowman Creek water samples. In comparison Moerke an d Lamberti (2006b) reported nitrate concentrations from undetectable levels to 7020 gN/L (mean = 1110 gN/L) in a study of 22 streams in southwestern Michigan. Figure 4. 10: Nitrate in Bowman Creek Soluble Reactive Phosphorus Peter Levi reported that certain soluble reactive phosphorus (SRP) samples had a white precipitate that neither he nor Mi ke Brueseke could explain. The anomalous samples tested significantly lower for SRP than the other samples from t he same site, which were generally in very close agreem ent with one another. Figure 4. 11 shows the mean values and error bars with all samples included. Figure 4. 12 shows the difference in SRP values betw een samples with and without precipitates. Figure 4. 13 shows the mean
119 values and error bars with the precipitated samples removed. In comparison Moerke and Lamberti (2006b) reported SRP concentrations from undetectable levels to 95.2 gP/L ammonium (mean = 16.2 gP/L) in a study of 22 streams in southwestern Michigan. Figure 4. 11 : Soluble Reactive Phosphorus in Bowman Creek
120 Figure 4. 12: Soluble Reactive Phosphorus in Bowman Creek, showing outliers Figure 4. 13: Soluble Reactive Phosphorus in Bowman Creek, outliers removed
121 4.2.2 Dissolved Organic Carbon data F igure 4. 14 shows dissolved organic carbon (DOC) concentrations at sites along Bowman Creek. A Shimadzu TOC 5000A combustion oxidation analyzer was used to measure DOC in water samples. The third sample bottle for the GOLF sit e was cracked and slowly leake d, resulting in a larger error bar for that site. S amples from other sites were very consistent. Though the differences in DOC concentration between some sites ma y be statistically significant, especially the higher value at the upstream LOCST site, these small differences in total DOC are unlikely to be biologically significant. Carbon dynamics have important regulatory effects on nutrient processing, but t he lability and quality (availability and energy density) of instream carbon sources are often more important than the total amount ( Allan and Castillo 2007; Tank et al. 2010). Figure 4. 14: Dissolved Organic Carbon in Bowman Creek
122 4.3 Biological data 4.3.1 Biofilm: chlorophyll a density and ashfree dry mass The following figures display the biofilm analysis results for samples collected at Bowman Creek sites, from upstream to downstream. T hese samples were collected on different da tes, as described in Table 2. 1 on p. 95. Chlorophyll a concentrations are shown in Figure 4. 15, ash free dry mass (AFDM) density is shown in Figure 4. 16, and the ratio of AFDM to chl a (the autotrophic index) is shown in Figure 4. 17. The RRCULV sample was taken from inside the downstream railroad crossing culvert at the LWE site on 7/28/2010 (not included in Table 2. 1) This was the only sample taken from inside a culvert and the RRCULV culvert is actually fairly short compared to some of the others along Bowman Creek. We hypothesized that dark conditions inside culverts, especially long culverts, would preclude photosynthesis and that algae would not be able to grow in these sit es. It would have been interesting to see whether biofilm samples taken from the middle of longer culverts would have been consistent with this hypothesis. Unfortunately, time constraints and safety concerns prevented these samples from being collected Bi ofilm samples from the culvert and the wooded URB site were low in chl a ( Figure 4. 15) most likely due to limited light. Somewhat counterintuitively, lower values of the Autotrophic Index (AI) indicate higher photosynthetic activi ty (Steinman et al. 2007). Figure 4. 17 shows that biofilm samples from unshaded sites had much lower AI values (higher autotrophic activity) than samples from shaded sites (URB and RRCULV).
123 Figure 4. 15: Chlorophyll a density in Bowman Creek Figure 4. 16: Ash Free Dry Mass in Bowman Creek biofilm
124 Figure 4. 17: Autotrophic Index in Bo wman Creek Biofilm (normal range is from Steinman et al. 2007) Chlorophyll a measurements from the MAIN, GOLF, RAV and LWE sites are comparable to values that have been reported from Juday Creek. On artificial substrates, chlorophyll a concentrations wer e low during much of the year when the canopy was welldeveloped (monthly mean chlorophyll a from April to October = 35.2 mg/m2). In contrast, periphyton was substantially higher when the canopy was open following autumn leaf fall (monthly mean chlorophyl l a from November to March = 106.2 mg/m2) (Schwenneker 1985 phide Lamberti and Berg 1995). Chlorophyll a concentrations recorded from the unshaded, high nutrient LOCST site are comparable to the autumn values reported from Juday Creek.
125 4.3.2 Benthic macroi nvertebrate community analysis Eighteen benthic macroinvertebrate samples were collected from Bowman Creek, containing 3741 identified specimens Figure 4. 18 shows the abundance of invertebrates at each site, reported as the mean number of individuals collected in a 0.09 m2 sample area. Three samples were taken from each site. Lower macroinvertebrate abundance was measured in the downstream RAV and LWE sites, despite the presence of large amounts of organic debris and leaf litter in the RAV and LWE 3 samples. Amphipods ( Gammarus sp.) were the most abundant taxa, with 925 individuals. Other abundant taxa included elmid beetles (825 specimens ; examples in Figure 4. 22b,c,d and e ), hydropsychid caddisflies (78 4; Figure 4. 20), chironomid midge larvae (641) baetid mayfly larvae (136) and black fly larvae (67; Figure 4. 21a) Figure 4. 18: Invertebrate abundanc e
126 Thirty five unique taxa were identified overall, representing 28 different families from 15 orders. Specimen counts for each taxon and sample are listed in Tables 7.1 and 7.2, in Appendix II Figure 4. 19 shows the cumulative invertebrate taxa richness for each site. The total number of unique taxa is reported rather than mean richness because between site differences were not meaningful (the standard error bars were large and overlapping). These results undoubtedly underestimate the actual number of species present, but the sample effort was insufficient to plot species accumulation curves or calculate expected richness. Figure 4. 19: Total invertebrate taxa richness per site
127 Four taxa from the Surber samples were identified to species: Macronychus glabratus Say 1825 ( Figure 4. 22c) Optioservus fastiditus (LeConte) 1850 Limnius ( Figure 4. 22b) Stenelmis crenata (Say) 1824 Elimis ( Figure 4. 22e) Peltodytes lengi Roberts 1913 ( Figure 4. 21b) Additional species of aquatic beetles were observed at the URB site, and a supplemental sample of submerged wood debris was collected. Specimens from this sample were not included in calculations. Two species were identified from this sample: Helichus striatus LeConte 1852 ( Figure 4. 22a) Sperchopsis tesselatus (Ziegler) 1844 Sperche us ( Fi gure 4. 23) These species were not collected at any other sites during the summer research. The presence of Sperchopsis tesselatus was an unexpected result. S. tesselatus is a fairly uncommon beetle (Spangler 1961), and it is listed as a species of State S pecial Concern in Wisconsin (WDNR 2009). The URB site was consistent with published descriptions of S. tesselatus habitat preferences which are unusual among hydrophilid beetles : Undercut gravelly and sandy stream banks with overhanging roots seem to be especially suitable (Spangler 1961).
128 All specimens on pp. 128 129 were photographed by the author. Scale bars display mm. Figure 4 20 : a) Hydropsyche and b) Cheumatopsyche (Trichoptera: Hydropsy chidae) Figure 4 21 : a) Black fly larva (Diptera: Simuliidae) b) Peltodytes lengi (Coleoptera: Haliplidae) a) b) a) b)
129 Figure 4 22 : Dryopidae: a) Helichus striatus ; Elmidae: b) Optioservus fastiditus c) Macronychus glabratus d) Dubiraphia sp. e) Stenelmis crenata Figure 4 23 : Sperchopsis tesselatus (Coleoptera: Hydrophilidae) a) b) c) d) e)
130 Functional feeding group (FFG) metric s were calculated using methods described by Merritt and Cummins (2007) (s ee Table 1. 1 on p. 74 for a mo re detailed list of FFG metrics). Inver tebrate taxa with unknown FFG designations were excluded from these analyses. Fewer than 20 specimens were excluded from any sample, with the exception of LWE 3. The 61 small annelids from this sample were difficult to identify after preservation in ethano l. Comparatively low invertebrate abundance at the RAV and LWE sites reduced the number of specimens available for analysis. Sample sizes were further reduced when specimens were partitioned into functional feeding groups. This limitation may have created artifacts in the results of the FFG calculations. Figure 4. 24 compares the proportions of different invertebrate guilds c ollected from each site. Results are displayed in pie charts t o facilitate comparison with Figure 1. 14. As predicted by the river continuum concept (Vannote et al. 1980), upstream communities were dominated by shredder invertebrates. Shredders break down leaf litter and other coarse particulate organic matter (CPOM) into fine particu late organic matter (FPOM) as they feed on the biofilms that cover these surfaces (Merritt and Cummins 2007). The dominance of collector invertebrates, which feed on FPOM generally increased from LOCST to RAV. The largely unshaded, nutrient enriched LWE site supported an increased percentage of scraper invertebrates These organisms graze on photosynthetic biofilms (Merritt and Cummins 2007).
131 Figure 4. 24: Comparisons of the relative abundance of d ifferent functional feeding groups
132 Functional feeding group abundance was also used to calculate several FFG metrics as described by Merritt and Cummins 2007. Results for FPOM transport were within normal ranges. Results for channel substrate stabilit y showed all sites below stable levels. The top down predator control metric was lower than typical values for all sites, but between site differences were unclear due to small sample sizes. Figure 4. 25 shows the ratio of phot osynthesis to respiration (P/R) as estimated by FFG analysis. This metric represents the abundance of Scrapers/(Shredders+Total Collectors) Calculated P/R values were < 0.75 for all sites, indicating heterotrophic conditions and a dependence on outside ( allochthonous) carbon fixation (Merritt and Cummins 2007). The P/R ratio is discussed in greater detail in Section 1.3.3. Figure 4. 26 shows an estimate of the ratio of CPOM (leaf litter) to FPOM (detri tus) in the stream. This metric represents the abundance of Shredders/(Total Collectors). Upstream sites had higher values. Figure 4. 25: P/R Ratio, based on functional feeding group analysis
133 Figure 4. 26: CPOM/ FPOM Ratio, based on functional feeding group analysis Invertebrate abundance data were also used to provide an estimate of water quality at each site. Abundance counts for each taxa were multiplied by a pollution tolerance factor. The higher the number, the greater the tolerance to impaired water quality. For example, leeches (Hirudinea) were assigned the highest tolerance factor (10), while hydropsychid caddisflies were assigned a factor of 4 (Carte r et al. 2007) Stoneflies (Plecoptera) are considered pollution sensitive organisms (factors of 0,1, or 2). No stoneflies were found in Bowman Creek. The sum of these weighted counts was divided by the total abundance to create a weighted index value for each sample. The Family Biotic Index, also known as the Hilsenhoff Index (Hilsenhoff 1988) was used for these calculations. Table 4. 1 shows the relationship between the Family Biotic Index and wat er quality (Carter et al. 2007). Table 4. 2 shows the index scores and water quality values for
134 sites on Bowman Creek. Figure 4. 27 shows the mean water quality value at each site. These results agree with the data obtained through nutrient and biofilm analysis, supporting our hypothesis that the water quality in Bowman Creek became increasingly impaired along an urban gradient from the URB site to the LWE site. Further research is needed to determine the factors influencing water quality at the headwaters of Bowman Creek. The LOCST site is unlikely to represent an appropriate reference condition. However, characterization of this section of the stream is important for understanding downstream water quality and the overall spatial heterogeneit y of Bowman Creek. Table 4. 1: Water quality values based on Family Biotic Index. Values from Hilsenhoff (1988); method reported in Carter et al. (2007) Water Quality Values based on Family Biotic Index 0.003.75 Excellent 3.764.25 Very good 4.265.00 Good 5.015.75 Fair 5.766.50 Fairly poor 6.517.25 Poor 7.2610.00 Very poor
135 Table 4. 2: Water quality values based on mean Family Biotic Index (FBI). SEM = standard error of the mean SITE FBI SEM Water Quality LOCST 4.67 0.30 Good URB 4.15 0.05 Very good MAIN 4.36 0.09 Good GOLF 5.03 0.26 Fair RAV 5.13 0.09 Fair LWE 6.26 0.64 Fairly poor Figure 4. 27: Family Biotic Index
136 Chapter 5: Conclusions and discussion 5.1 Discussion of results and p otential sources of error The Merriam Webster dictionary provides several definitions for the word problem, including an intricate unsettled question and a source of perplexity, distress, or vexation. By both definitions, I encountered many problems in the course of this investigation. However, I believe I have learned a great deal from the endeavor, and also from the frank acknowledgment of the pro jects shortcomings and strengths. As Richard Feynman said in Cargo Cult Science his 1974 commencement address at Caltech (reprinted in (Feynman and Leighton 1997) : It's a kind of scientific integrity, a principle of scientific thought that correspond s to a kind of utter honest y a kind of leaning over backwards. For example, if you're doing an experiment, you should report everything that you think might make it invalid not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you've eliminated by some other experiment, and how they worked to make sure the other fellow can tell they have been eliminated.  The first principle is that you must not fool yourself and you are the easiest person to fool. So you have to be very careful about that. After you've not fooled yourself, it's easy not to fool other scientists. You just have to be honest in a conventional way after that. The most serious issues with my project stem from the mismatch between the complex questions I was interested in asking and the very short amount of time available to answer them. I f I had begun collecting data at the beginning of the summer REU session, I might have been able to gather sufficient data to address a simpler, more straightforward
137 research question. However, I should have realized that the two main goals of my research investigating the effect of culverts on stream ecosystems and collecting baseline ecological data on Bowman Creek prior to a planned restoration project could not be completed in a single summer. Instead, this project should be considered a pilot study. The results show that further investigation is warranted, and the data from this project can be used to direct future research effort s and monitoring programs. Data collected in this study were not sufficient to draw broadly applicable conclusions about the effect of culverts on stream ecosystem function. Between stream replicate samples are needed to adequately characterize this syst em of interactions We initially planned to sample sites on another urban stream (McCoys Creek, just over the state line in Buchanan, Michigan) but time constraints prevented this research from being carried out. Because all samples in this project were taken from sites on the same stream, it is not possible to separate the specific effects of culverts from the background changes that one would expect to find along an upstream to downstream grad ient. The location s of sample sites along the creek were not ideal for isolating the effects of culverts on ecosystem function Site selection was necessarily limited by accessibility (landowner permission) and suitability for MiniSonde deployment (sufficient water depth). The LOCST and URB sites were intended to serve as reference conditions, without culverts, but these sites were so different from one another that they were of little use in establishing a natural or undisturbed baseline for the creek. The LOCST site was essentially a roadside ditch receiving agr icultural and construction runoff, while the URB site was heavily wooded and shaded The MAIN site was at the edge of the city, receiving flow from wooded upstream areas but surrounded by urban
138 infrastructure. The downstream sites (GOLF, DAYTON, RAV, LWE and RRCULV) were all heavily affected by culvert flow and urbanization, but these impacts differed from site to site. A more rigorous study design would incorporate between stream replicates for comparison of upstream reference sites and downstream culvert affected sites. Because the sample sites in this project could not be separated into meaningful, distinct categories or placed along a onedimensional gradient of urbanization, only descriptive statistics were calculated for the data collected. The sampling regime followed by this summer project did establish that the ecological conditions on Bowman Creek var ied from upstream to downstream. This spatial heterogeneity should be taken into consideration when designing future studies and restoration efforts. Between site variations in diel respiration curves were much higher for Bowman Creek than for Juday Creek. D istances between sample sites were comparable. The f arthest upstream and farthest downstream stations on Juday Creek were separated by 2.4 km of str eam channel Site separation varied from 0.9 km to 3.2 km during Bowman Creek datalogger deployments. Juday Creek is a larger stream (3rd order) than Bowman Creek (1st order) and has a higher baseline discharge. Juday Creeks higher flow volume and velocit y could have contributed to the tighter correlation of readings among the Juday Creek sites. Additionally, there is a small dam on Bowman Creek between the URB and MAIN sites, on the AM General property. This dam could have created a discontinuity ( sensu W ard 1989) between the upstream sites (LOCST and URB) and the downstream sites (MAIN to LWE). Site separation distances and spatial heterogeneity also contributed to the technical difficulties we encountered in measuring whole stream metabolism as
139 discus sed in Section 4.1.1 ( p. 105) T he few diel respiration curves recorded from Bowman Creek were not consistent enough to fit the models parameters and the sections between sites were not uniform. Further research is needed to quantify the stream discharge and reaeration coefficient at sites along Bowman Creek. These measurements could then be used to investigate the longitudinal lag and persistence of culvert impacts. If the culverts did have an ef fect on stream ecosystems, how long into the culvert would normal, upstream water conditions persist? How long downstream would the effects of the culvert persist? Do culverts exert a greater influence on water quality or hydrology and habitat quality? Bo th of the topics addressed in this research (culvert effects and baseline data) exhibit ed temporal and spatial variation. T he results of this pilot study are not necessarily generalizable to streams other than Bowman Creek, or to season s other than mid sum mer. The abundance of macroinvertebrate species often varies seasonally. In Juday Creek, nymphs of the stonefly Taeniopteryx nivalis Fitch 1847 diapause deep in the substrate from the emergence of the adults in February until September (Kohlhepp and Hell enthal 1992) I f this species were present in Bowman Creek it would not have been collected in our summer samples. However, Helms et al. (2009) found that the effect of urbanization on stream macroinvertebrates appears to be consistent throughout the yea r and reduces seasonal changes in assemblages. Large fish were observed inhabiting the outflow of the DAYTON culvert. More research is needed to determine whether these sites can provide cool water oxygenated refugia during warm, low DO summer condition s. Additional sampling is required to establish whether fish inhabit these structures during other seasons and what effect
140 culverts have on water temperature during the extreme cold of an Upper Midwest winter. Several additional sampling and research des ign issues could have affected the results of this study The Surber sampler used in this project had mesh with 1 mm openings which are larger than the 0.5 mm standard for benthic macroinvertebrate sampling (Hauer and Resh 2007). The looser weave could ha ve allowed more chironomid midge larvae and other small invertebrates to pass through the net, skewing the resulting community composition data. Despite the slightly coarser mesh, abundant chironomids and early instar beetle larvae were recorded from the S urber samples. Another potential source of error is the variation in culvert design. The culverts have different shapes (ci rcular vs. rectangular ), cross sectional areas and substrate types. For example, the short railroad crossing culverts exhibit a fa irly natural distribution of rocks and sediment, while the downstream end of the NIPSCO culvert was nearly devoid of sediment. Any single aspect of the culvert s could be isolated and tested, given the proper study sites and experimental designs. This study did not establish clear differences between nonculvert and culvert sites. O ther confounding variables (urban gradient, culvert length and culvert design) also reduced the explanatory power of these findings. The problems discussed in this section were not un ique to this project. Instead, these challenge s are often encountered in urban stream ecology. Walsh et al. (2005 b ) explain : Our understanding of stressor mechanisms and their interactions is limited by a lack of experimental (i.e., causal) evidence. The unraveling of the interactions between smallscale stressors may ultimately prove experimentally intractable. The overall design of this project could be described a spacefor time substitution, using a spatial gradient of urbanization as a stand in for temporal change in urbanization at a single site
141 (Carter et al. 2009). This study could also be understood as the first half of a before after control impact monitoring project (Carter et al. 2009), but no conclusions can be drawn from this aspect o f the project until follow up data are collected after the citys restoration efforts. Space for time substitutions often include only a small number of streams in a given urban area, so these studies typically have very few degrees of freedom with which to separate the effects of different stressors (Carter et al. 2009). Short sampling periods and temporal fluctuations further decrease the explanatory power of these studies and given this uncertainty, SFT may be more appropriate in generating hypotheses about functional parameters than it is in testing them (Pickett 1989) Combinations of experimental and monitoring approaches, as described by Carter et al. 2009 and Culp and Baird (2007) may be more effective for rigorous testing of hypotheses generated by spacefor time substitutions such as this project. This project could have benefited from greater engagement with the surrounding community although I did have some interesting and excellent conversations with streamside residents Better communicat ion with the Public Works Department of the City of South Bend would facilitate efforts to coordinate sampling and to ensure that the data collected would be useful to the city. A volunteer monitoring program, involving collaboration between community memb ers and university researchers, could provide useful b aseline data on Bowman Creek This type of involvement w ould result in more experiential interaction with the urban ecosystem, hopefully leading to greater understanding, appreciation and conservation o f the freshwater resources on which we depend.
142 5.2 F uture research directions This project was in many respects more of a pilot study than a solid, standalone research endeavor. I gained a much greater appreciation for the dynamic and complex nature of stream ecosystems, and the challenges associated with experimental design in studies of landscape level urban disturbance. There are many ways in which the results of this pilot study could be used to inform future investigations into this under studied system One goal of our study was to collect pre restoration e cological data on Bowman Creek. T he most valuable extension of this research would be to continue collecting data, to gain a greater understanding of seasonal and year to year variation. Placing the data in a temporal context would greatly improve its utility in establishing a baseline condition for the ecosystem function of the creek. Maintaining a long term monitoring regime after the restoration project would yield valuable insights on the streams response and provide data that could be used to refine subsequent restoration efforts. Another goal was to determine the effect s of culv erts on stream ecosystem health. E xpanding the project to include between stream replicates w ould be the single best choice to improve the broader relevance of the se results. There are several additional projects that would complement the data already collected F ish sampling would be an excellent way to expand our understanding of the Bowman Creek ecosystem. The Lambert i Lab has conducted prior electrofishing projects on Juday Creek ( Moerke and Lamberti 2003) E lectroshocking sections of culvert on Bowman Creek would be especially interesting. This research, if conducted in different seasons, could help determine whether the culverts function as a fish refuge during temperature extremes.
143 Also, if open channel metabolism modeling is not compatible with short between site distances chamber respirometers could provide another method of assessing differences in stream metabo lism between culvert and non culvert sites (Bott 2007) As an extension of the insect data, collected samples could be further analyzed to quantify secondary productivity (Kohlhepp and Hellenthal 1992; Lamberti and Berg 1995; Benke and Huryn 2007; Entrekin et al. 2009) If significant differences in ecosystem composition were observed between wellchosen sites with and without culverts, secondary studies would be useful to elucidate the intermediate mechanisms by which the culverts were affecting the biolo gical communities. For example, studies could be designed to address whether water flows faster through culverts, or whether the substrate composition differs between culvert and nonculvert stream sections (Moerke et al. 2004; Faustini and Kaufmann 2007) Informal observations show that culverts appear to have fewer stable attachment surfaces for aquatic organisms (such as rocks and large woody debris) and fewer accumulations of coarse particulate organic matter such as leafpacks. This suggests that culver ts could affect habitat and food sources for aquatic organisms by altering transport and retention rat es for rocks and organic debris. Q uantification of flow and retention rates ( using rhodamine dye, dowels, ginkgo leaves and other materials ) could prove v ery informative (Lamberti and Gregory 2007) This research could build on the retention studies conducted by Aldridge et al. ( 2009) Cordova et al. (2008) and Borchardt ( 1993 ) Retention in Bowman Creek culverts could also be compared to retention in Juday Creek (Ehrman and Lamberti 1992) Studies of nutrient uptake and spiraling lengths in long culverts, as compared to upstream and downstream reaches, could help determine the
144 effect of stream burial on stream nutrient dynamics (Stream Solute Workshop 1990) Investigations of the effects of culverts on water flow and nutrient exchange with the hyporheic zone and groundwater could also provide valuable information on the watershed scale hydrological effects of culverts (Mulholland et al. 1997; Groffman et al. 2005) 5.3 Implications for restoration The citys focus on whole catchment restoration strategies and low impact development show great promise for improving ecosystem function and water quality in the streams and rivers surrounding South Bend. However, cu rrent proposed projects would result in the direct discharge of stormwater to Bowman Creek at an additional two sites one near Haney Ave. and Dubail Ave. and another near Riley High School (City of South Bend 2010). This approach is strongly discouraged u nder the guidelines developed by Palmer et al. (2005; 2010) and Walsh et al. (2009), if the goal is to improve stream ecosystem function.
145 Chapter 6: Works Cited Adams LW, Dove LE, Leedy DL. 1984. Public Attitudes toward Urban Wetlands for Stormwater Control and W ildlife Enhancement. Wildlife Society Bulletin 12:299 303. Aldridge KT, Brookes JD, Ganf GG. 2009. Rehabilitation of Stream Ecosystem Functions through the Reintroduction of Coarse Particulate Organic Matter. Restoration Ecology 17:97106. Allan JD, Castil lo MM. 2007. Stream Ecology: Structure and function of running waters. 2nd ed. Dordrecht, The Netherlands: Springer Anderson KE, Paul AJ, McCauley E, Jackson LJ, Post JR, Nisbet RM. 2006. Instream Flow Needs in Streams and Rivers: The Importance of Underst anding Ecological Dynamics. Frontiers in Ecology and the Environment 4:309 318. Anderson V. 2010. REU Project List 2010. GLOBES, University of Notre Dame [Internet]. Available from: http://globes.nd.edu/reulist2010.shtml#policy Armitage PD, Cranston PS, Pi nder LCV. 1995. The Chironomidae: biology and ecology of non biting midges. London: Chapman and Hall Arscott DB, Jackson JK, Kratzer EB. 2006. Role of Rarity and Taxonomic Resolution in a Regional and Spatial Analysis of Stream Macroinvertebrates. Journal of the North American Benthological Society 25:977 997. Bailey RC, Norris RH, Reynoldson TB. 2001. Taxonomic Resolution of Benthic Macroinvertebrate Communities in Bioassessments. Journal of the North American Benthological Society 20:280 286. Baker DB, Ri chards RP, Loftus TT, Kramer JW. 2004. A New Flashiness Index: Characteristics and Applications to Midwestern Rivers and Streams. Journal of the American Water Resources Association 40:503 522. Ball P. 2009. Flow. New York: Oxford University Press Barbour MT, Gerritsen J, Griffith GE, Frydenborg R, McCarron E, White JS, Bastian ML. 1996. A Framework for Biological Criteria for Florida Streams Using Benthic Macroinvertebrates. Journal of the North American Benthological Society 15:185211. Barbour MT, Gerrit sen J, Snyder BD, Stribling JB. 1999. Rapid Bioassessment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic Macroinvertebrates, and Fish, Second Edition. Washington, D.C.: U.S. Environmental Protection Agency; Office of Water Available from: http://water.epa.gov/scitech/monitoring/rsl/bioassessment/index.cfm
146 Bejan A, Lorente S. 2008. Design with constructal theory. Hoboken N.J.: John Wiley & Sons Benfield EF. 2007. Decomposition of Leaf Material. In: Methods in Stream Ecology. 2nd ed. B urlington, MA: Academic Press. pp. 711 720. Benke AC, Huryn AD. 2007. Secondary Production of Macroinvertebrates. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 691710. Benyus JM. 2002. Biomimicry: Innovation Inspired by Nature New York: Harper Perennial Bergey E, Getty G. 2006. A Review of Methods for Measuring the Surface Area of Stream Substrates. Hydrobiologia 556:716. Berkowitz AR, Nilon CH, Hollweg KS. 2002. Understanding Urban Ecosystems. 1st ed. New York: Springer Be rnhardt ES, Band LE, Walsh CJ, Berke PE. 2008. Understanding, Managing, and Minimizing Urban Impacts on Surface Water Nitrogen Loading. Annals of the New York Academy of Sciences 1134:6196. Borchardt D, Statzner B. 1990. Ecological impact of urban stormwa ter runoff studied in experimental flumes: Population loss by drift and availability of refugial space. Aquatic Sciences Research Across Boundaries 52:299 314. Borchardt D. 1993. Effects of flow and refugia on drift loss of benthic macroinvertebrates: im plications for habitat restoration in lowland streams. Freshwater Biology 29:221227. Bournaud M, Cellot B, Richoux P, Berrahou A. 1996. Macroinvertebrate Community Structure and Environmental Characteristics along a Large River: Congruity of Patterns for Identification to Species or Family. Journal of the North American Benthological Society 15:232 253. Bott TL. 2007. Primary Productivity and Community Respiration. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 663690. Brown LR Cuffney TF, Coles JF, Fitzpatrick F, McMahon G, Steuer J, Bell AH, May JT. 2009. Urban streams across the USA: lessons learned from studies in 9 metropolitan areas. Journal of the North American Benthological Society 28:10511069. Cairns JJ, Pratt JR. 19 93. A history of biological monitoring using benthic macroinvertebrates. In: Freshwater Biomonitoring and Benthic Macroinvertebrates. New York: Chapman and Hall. pp. 10 27.
147 Cardinale BJ. 2011. Biodiversity improves water quality through niche partitioning. Nature 472:8689. Carpenter SR, Bennett EM. 2011. Reconsideration of the planetary boundary for phosphorus. Environ. Res. Lett. 6:014009. Carroll CR. 2008. Cities and the rivers that run through them. Basins and Coasts. Integrated Management of Coastal an d Freshwater Systems. 2:10 12. Carter JL, Resh VH, Hannaford MJ, Myers MJ. 2007. Macroinvertebrates as Biotic Indicators of Environmental Quality. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 805831. Carter T, Jackson CR, Ros emond A, Pringle C, Radcliffe D, Tollner W, Maerz J, Leigh D, Trice A. 2009. Beyond the urban gradient: barriers and opportunities for timely studies of urbanization effects on aquatic ecosystems. J. N. Am. Benthol. Soc. 28:10381050. Carter T, Jackson CR. 2007. Vegetated roofs for stormwater management at multiple spatial scales. Landscape and Urban Planning 80:8494. Caskey BJ, Frey JW, Selvaratnam S. 2010. Breakpoint Analysis and Assessment of Selected Stressor Variables on Benthic Macroinvertebrate and Fish Communities in Indiana Streams: Implications for Developing Nutrient Criteria. Reston, VA: U.S. Geological Survey Available from: http://pubs.usgs.gov/sir/2010/5026/pdf/sir20105026_web.pdf City of South Bend. 2010. River Report. Available from: http: //www.southbendin.gov/doc/BPW_CleanRiverRpt_0610.pdf Collins KA, Lawrence TJ, Stander EK, Jontos RJ, Kaushal SS, Newcomer TA, Grimm NB, Cole Ekberg ML. 2010. Opportunities and challenges for managing nitrogen in urban stormwater: A review and synthesis. Ec ological Engineering 36:15071519. Cordova JM, Rosi Marshall EJ, Tank JL, Lamberti GA. 2008. Coarse particulate organic matter transport in low gradient streams of the Upper Peninsula of Michigan. Journal of the North American Benthological Society 27:760 771. Courtemanch DL, Davies SP. 1987. A coefficient of community loss to assess detrimental change in aquatic communities. Water Research 21:217 222. Culp JM, Baird DJ. 2007. Establishing Cause Effect Relationships in MultiStressor Environments. In: Metho ds in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 835 854.
148 Dahm CN, Valett H. Maurice, Baxter CV, Woessner WW. 2007. Hyporheic Zones. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 119142. Dang CK, Harrison S, Sturt MM, Giller PS, Jansen MAK. 2009. Is the elemental composition of stream invertebrates a determinant of tolerance to organic pollution? Journal of the North American Benthological Society 28:778784. DeAngelis DL. 1980. Energy Flow, Nutrient Cycling, and Ecosystem Resilience. Ecology 61:764771. Deegan D. 2010. Aquatic Community Monitoring in Elkhart and St. Joseph Counties on the St. Joseph River and Selected Tributaries 2009. Available from: http://www.southbendin.gov/docs/WW_AquaticsProgram2009Repor t.pdf Diaz RJ, Rosenberg R. 2008. Spreading Dead Zones and Consequences for Marine Ecosystems. Science 321:926 929. Ehrman TP, Lamberti GA. 1992. Hydraulic and Particulate Matter Retention in a 3rd Order Indiana Stream. Journal of the North American Benthological Society 11:341349. Elmore AJ, Kaushal SS. 2008. Disappearing headwaters: patterns of stream burial due to urbanization. Frontiers in Ecology and the Environment 6:308 312. Ensign SH, Doyle MW. 2006. Nutrient spiraling in streams and river networks. J. Geophys. Res. 111:CiteID G04009. Entrekin SA, Rosi Marshall EJ, Tank JL, Hoellein TJ, Lamberti GA. 2007 Macroinvertebrate secondary production in 3 forested streams of the upper Midwest, USA. JNABS 23:472460. Entrekin SA, Tank JL, Rosi Marshall EJ, Hoellein TJ, Lamberti GA. 2009. Response of secondary production by macroinvertebrates to large wood addition in three Michigan streams. Freshwater Biology 54:17411758. Fahnenstiel G, Pothoven S, Vanderploeg H, Klarer D, Nalepa T, Scavia D. 2010. Recent changes in primary production and phytoplankton in the offshore region of southeastern Lake Michigan. Journal of Great Lakes Research 36:20 29. Faustini JM, Kaufmann PR. 2007. Adequacy of Visually Classified Particle Count Statistics From Regional Stream H abitat Surveys1. JAWRA Journal of the American Water Resources Association 43:1293 1315. Fellows CS, Valett H. Maurice, Dahm CN. 2001. WholeStream Metabolism in Two Montane Streams: Contribution of the Hyporheic Zone. Limnology and Oceanography 46:523531.
149 Feyen J. 2008. Water and Urban Development Paradigms: Towards an Integration of Engineering, Design and Management Approaches. CRC Press Feynman RP, Leighton R. 1997. Surely Youre Joking, Mr. Feynman! New York: W. W. Norton & Company Feynman RP. 1967. T he character of physical law. Cambridge: M.I.T. Press Filippelli GM. 2008. The Global Phosphorus Cycle: Past, Present, and Future. ELEMENTS 4:89 95. FL DEP. 2006. Methods of Hester Dendy Multiplate Sampler Deployment. FL DEP, Bureau of Laboratories, Biology Section. Available from: http://publicfiles.dep.state.fl.us/dear/sas/library/docs/hester_dendy.pdf Francis RA, Hoggart SPG. 2008. Waste Not, Want Not: The Need to Utilize Existing Artificial Structures for Habitat Improvement Along Urban Rivers. Restorat ion Ecology 16:373381. Frissell CA, Liss WJ, Warren CE, Hurley MD. 1986. A hierarchical framework for stream habitat classification: Viewing streams in a watershed context. Environmental Management 10:199214. Geisel TS. 1971. The Lorax. New York: Random House Books for Young Readers Gilot GA. 2009. Bowman Creek Restoration Project. Available from: http://bowmancreek.michianastem.org/Alt Home Gordon LJ, Peterson GD, Bennett EM. 2008. Agricultural modifications of hydrological flows create ecological surpri ses. Trends in Ecology & Evolution 23:211219. Gotelli NJ, Ellison AM. 2004. A Primer Of Ecological Statistics. Sunderland, MA: Sinauer Associates Grace MR, Imberger SJ. 2006. Stream Metabolism: Performing & Interpreting Measurements. Water Studies Centre Monash University, Murray Darling Basin Commission and New South Wales Department of Environment and Climate Change Available from: http://www.sci.monash.edu.au/wsc/docs/tech manual v3.pdf Groffman PM, Bain DJ, Band LE, Belt KT, Brush GS, Grove JM, Pouyat RV, Yesilonis IC, Zipperer WC. 2003. Down by the riverside: urban riparian ecology. Frontiers in Ecology and the Environment 1:315321. Groffman PM, Dorsey AM, Mayer PM. 2005. N processing within geomorphic structures in urban streams. Journal of the North American Benthological Society 24:613 625.
150 Gullan PJ, Cranston PS. 2010. The Insects: An Outline of Entomology. West Sussex, UK: John Wiley and Sons Hall RO Jr., Tank JL. 2005. Correcting whole stream estimates of metabolism for groundwater input. Limnolo gy and Oceanography: Methods 3:222 229. Hauer FR, Lamberti GA. 2007. Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press Hauer FR, Will WR. 2007. Temperature, Light, and Oxygen. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 103117. Heisenberg W. 1958. Physics and Philosophy: The Revolution in Modern Science. New York: Harper and Row Helms BS, Schoonover JE, Feminella JW. 2009. Seasonal variability of landuse impacts on macroinvertebrate assemblages in streams of w estern Georgia, USA. Journal of the North American Benthological Society 28:991 1006. Helms BS. 2008. Response of Aquatic Biota to Changing Land Use Pattern in Streams of West Georgia, USA. Herbst DB, Silldorff EL. 2006. Comparison of the Performance of Di fferent Bioassessment Methods: Similar Evaluations of Biotic Integrity from Separate Programs and Procedures. Journal of the North American Benthological Society 25:513530. Hilsenhoff WL. 1982. Using a biotic index to evaluate water quality in streams. Te chnical Bulletin 132. Madison, WI: Wisconsin Department of Natural Resources Available from: http://digicoll.library.wisc.edu/cgibin/EcoNatRes/EcoNatRes idx?id=EcoNatRes.DNRBull132 Hilsenhoff WL. 1988. Rapid Field Assessment of Organic Pollution with a Fa mily Level Biotic Index. Journal of the North American Benthological Society 7:65 68. Howard AD, Dietrich WE, Seidl MA. 1994. Modeling fluvial erosion on regional to continental scales. Journal of Geophysical Research 99:13,97113,986. Ingall E, Jahnke R. 1997. Influence of water column anoxia on the elemental fractionation of carbon and phosphorus during sediment diagenesis. Marine Geology 139:219229. Jowett IG. 1993. A method for objectively identifying pool, run, and riffle habitats from physical measur ements. New Zealand Journal of Marine and Freshwater Research 27:241248. Kstner J, Blchl PE. 2007. Ammonia Production at the FeMo Cofactor of Nitrogenase:
151 Results from Density Functional Theory. Journal of the American Chemical Society 129:29983006. K aushal SS, Groffman PM, Mayer PM, Striz E, Gold AJ. 2008. Effects of stream restoration on denitrification in an urbanizing watershed. Ecological Applications 18:789804. Kay JJ. 1991. A Non equilibrium Thermodynamic Framework for Discussing Ecosystem Inte grity. Environmental Management 15:483495. Kohlhepp GW, Hellenthal RA. 1992. The Effects of Sediment Deposition on Insect Populations and Production in a Northern Indiana Stream. In: Proceedings of the 1991 Midwest Pollution Control Biologists meeting: En vironmental Indicators: Measurement and Assessment Endpoints. Report No. EPA 905/R 92/003. Chicago, IL: U.S. EPA, Region V, Environmental Sciences Division. pp. 73 84. Lamberti GA, Berg MB. 1995. Invertebrates and Other Benthic Features as Indicators of En vironmental Change in Juday Creek, Indiana. Natural Areas Journal 15:249258. Lamberti GA, Gregory SV. 2007. CPOM Transport, Retention, and Measurement. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 273 289. Lancaster J, Hildre w AG. 1993. Flow Refugia and the Microdistribution of Lotic Macroinvertebrates. Journal of the North American Benthological Society 12:385 393. Lehmkuhl DM. 1979. How to Know the Aquatic Insects. Dubque, Iowa: Wm. C. Brown Co. Lenat DR, Barbour MT. 1994. U sing benthic macroinvertebrate community structure for rapid, cost effective, water quality monitoring: Rapid bioassessment. In: Biological Monitoring of Aquatic Systems. Boca Raton, FL: CRC Press. pp. 187 215. Lenat DR, Resh VH. 2001. Taxonomy and Stream Ecology: The Benefits of Genus and Species Level Identifications. Journal of the North American Benthological Society 20:287298. Louv R. 2008. Last Child in the Woods: Saving Our Children From Nature Deficit Disorder. Revised and expanded. Chapel Hill, N C: Algonquin Books Maberly SC. 1996. Diel, episodic and seasonal changes in pH and concentrations of inorganic carbon in a productivelake. Freshwater Biology 35:579 598.
152 Maloney KO, Feminella JW, Mitchell RM, Miller SA, Mulholland PJ, Houser JN. 2008. Lan duse legacies and small streams: identifying relationships between historical land use and contemporary stream conditions. Journal of the North American Benthological Society 27:280 294. Maloney KO, Weller DE, Russell MJ, Hothorn T. 2009. Classifying the biological condition of small streams: an example using benthic macroinvertebrates. Journal of the North American Benthological Society 28:869 884. Marchal J. 2005. Research Article: An evaluation of the accuracy of order level biotic indices for southern A ppalachian streams. BIOS 76:6167. Marzolf ER, Mulholland PJ, Steinman AD. 1994. Improvements to the Diurnal Upstream Downstream Dissolved Oxygen Change Technique for Determining Whole Stream Metabolism in Small Streams. Can. J. Fish. Aquat. Sci. 51:15911599. McCafferty WP. 1983. Aquatic entomology: the fishermens and ecologists illustrated guide to insects and their relatives. Sudbury, MA: Jones and Bartlett Publishers McDonough W, Braungart M. 2002. Cradle to Cradle: Remaking the Way We Make Things. 1s t ed. New York: North Point Press Merritt RW, Cummins KW, Berg MB, Novak JA, Higgins MJ, Wessell KJ, Lessard JL. 2002. Development and Application of a Macroinvertebrate Functional Group Approach in the Bioassessment of Remnant River Oxbows in Southwest Fl orida. Journal of the North American Benthological Society 21:290 310. Merritt RW, Cummins KW, Berg MB. 2008. An Introduction to the Aquatic Insects of North America. 4th ed. Dubuque, IA: Kendall Hunt Publishing Merritt RW, Cummins KW. 2007. Trophic Relati onships of Macroinvertebrates. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 585 609. Meyer Judy L., Paul MJ, Taulbee WK. 2005. Stream ecosystem function in urbanizing landscapes. Journal of the North American Benthological Soc iety 24:602612. Meyer KM, Kump LR. 2008. Oceanic Euxinia in Earth History: Causes and Consequences. Annu. Rev. Earth Planet. Sci. 36:251288. Miao S, Carstenn S, Nungesser MK eds. 2009. Real world ecology: large scale and long term case studies and method s. New York: Springer
153 Mida JL, Scavia Donald, Fahnenstiel GL, Pothoven SA, Vanderploeg HA, Dolan DM. 2010. Longterm and recent changes in southern Lake Michigan water quality with implications for present trophic status. Journal of Great Lakes Research 36:4249. Minshall GW, Petersen RC, Cummins KW, Bott TL, Sedell JR, Cushing CE, Vannote RL. 1983. Interbiome Comparison of Stream Ecosystem Dynamics. Ecological Monographs 53:225. Moerke AH, Gerard KJ, Latimore JA, H ellenthal RA, Lamberti GA. 2004. Restor ation of an Indiana, USA, Stream: Bridging the Gap between Basic and Applied Lotic Ecology. Journal of the North American Benthological Society 23:647660. Moerke AH, Lamberti GA. 2003. Responses in Fish Community Structure to Restoration of Two Indiana St reams. North Amer. J. of Fisheries Mgmt 23:748 759. Moerke AH, Lamberti GA. 2006a. Relationships between Land Use and Stream Ecosystems: A Multistream Assessment in Southwestern Michigan. In: American Fisheries Society Symposium. Vol. 48. pp. 323338. Moer ke AH, Lamberti GA. 2006b. Scale dependent influences on water quality, habitat, and fish communities in streams of the Kalamazoo River Basin, Michigan (USA). Aquat. Sci. 68:193205. Moore AA, Palmer MA. 2005. Invertebrate Biodiversity in Agricultural and Urban Headwater Streams: Implications for Conservation and Management. Ecological Applications 15:11691177. Mulholland PJ, Marzolf ER, Hendricks SP, Wilkerson RV, Baybayan AK. 1995. Longitudinal Patterns of Nutrient Cycling and Periphyton Characteristics in Streams: A Test of Upstream Downstream Linkage. Journal of the North American Benthological Society 14:357 370. Mulholland PJ, Marzolf ER, Webster JR, Hart DR, Hendricks SP. 1997. Evidence That Hyporheic Zones Increase Heterotrophic Metabolism and Phosp horus Uptake in Forest Streams. Limnology and Oceanography 42:443451. Nerbonne JF, Ward B, Ollila A, Williams M, Vondracek B. 2008. Effect of sampling protocol and volunteer bias when sampling for macroinvertebrates. J. N. Am. benthol. Soc. 27:640 646. Ni emczynowicz J. 1999. Urban hydrology and water management present and future challenges. Urban Water 1:1 14.
154 Novotny V, Clark D, Griffin RJ, Booth D. 2001. Risk based urban watershed management under conflicting objectives. Water Sci. Technol 43:6978. Novotn V, Ahern J, Brown P. 2010. Water Centric Sustainable Communities Retrofitting, and Building the Next Urban Environment. Hoboken N.J.: Wiley Novotn V, Brown PR. 2007. Cities of the future: towards integrated sustainable water and landscape management rnational workshop held July 1214, 2006 in Wingspread Conference Center, Racine, WI. IWA Publishing Odum HT. 1956. Primary Production in Flowing Waters. Limnology and Oceanography 1:102117. Oksanen L. 2001. Logic of Experiments in Ecology: Is Pseudorepli cation a Pseudoissue? Oikos 94:2738. Ozaki K, Tajima S, Tajika E. 2011. Conditions required for oceanic anoxia/euxinia: Constraints from a one dimensional ocean biogeochemical cycle model. Earth and Planetary Science Letters 304:270 279. zgner H, Kendle AD. 2006. Public attitudes towards naturalistic versus designed landscapes in the city of Sheffield (UK). Landscape and Urban Planning 74:139157. Palmer MA, Ambrose RF, Poff NL. 1997. Ecological theory and community restoration ecology. Restoration Ecology 5:291300. Palmer MA, Bernhardt ES, Allan JD, Lake PS, Alexander G, Brooks S, Carr J, Clayton S, Dahm CN, Follstad Shah J, et al. 2005. Standards for ecologically successful river restoration. Journal of Applied Ecology 42:208 217. Palmer MA, Menninger HL, Bernhardt E. 2010. River restoration, habitat heterogeneity and biodiversity: a failure of theory or practice? Freshwater Biology 55:205 222. Palmer MA. 2008. Reforming Watershed Restoration: Science in Need of Application and Applications in Need of S cience. Estuaries and Coasts 32:1 17. Paul MJ, Meyer Judy L. 2001. Streams in the Urban Landscape. Annual Review of Ecology and Systematics 32:333365. Pickett STA. 1989. Space for time substitution as an alternative to long term studies. In: Long term stu dies in ecology: approaches and alternatives. New York: Springer. pp. 110135.
155 Pigay H, Gregory KJ, Bondarev V, Chin A, Dahlstrom N, Elosegi A, Gregory SV, Joshi V, Mutz M, Rinaldi M, et al. 2005. Public Perception as a Barrier to Introducing Wood in Ri vers for Restoration Purposes. Environmental Management 36:665674. Pinkham R. 2000. Daylighting: New Life for Buried Streams. Old Snowmass, CO: Rocky Mountain Institute Plotnikoff R, Wiseman C. 2001. Benthic Macroinvertebrate Biological Monitoring Protoco ls for Rivers and Streams: 2001 Revision. Olympia, Washington: Washington State Department of Ecology Available from: http://www.ecy.wa.gov/biblio/0103028.html Poff NL, Allan JD, Bain MB, Karr JR, Prestegaard KL, Richter BD, Sparks RE, Stromberg JC. 1997. The Natural Flow Regime. BioScience 47:769 784. Poff NL, Olden JD, Vieira NKM, Finn DS, Simmons MP, Kondratieff BC. 2006. Functional Trait Niches of North American Lotic Insects: Traits Based Ecological Applications in Light of Phylogenetic Relationships. Journal of the North American Benthological Society 25:730 755. Pouyat RV, Pataki DE, Belt KT, Groffman PM, Hom J, Band LE. 2007. Effects of urban land use change on biogeochemical cycles. In: Terrestrial ecosystems in a changing world. Berlin: Springer Ve rlag. pp. 4558. Available from: http://www.nrs.fs.fed.us/pubs/4129 Pringle CM, Triska FJ. 2007. Effects of Nutrient Enrichment on Periphyton. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 743759. Purcell AH, Corbin JD, Hans K E. 2007. Urban riparian restoration: an outdoor classroom for college and high school students collaborating in conservation. Madroo 54:258267. Ramrez A, De Jess Crespo R, Martin Cardona DM, MartnezRivera N, Burgos Caraballo S. 2009. Urban streams i n Puerto Rico: what can we learn from the tropics? Journal of the North American Benthological Society 28:10701079. Resh VH, Brown AV, Covich AP, Gurtz ME, Li HW, Minshall GW, Reice SR, Sheldon AL, Wallace JB, Wissmar RC. 1988. The Role of Disturbance in Stream Ecology. Journal of the North American Benthological Society 7:433 455. Reynoldson TB, Norris RH, Resh VH, Day KE, Rosenberg DM. 1997. The Reference Condition: A Comparison of Multimetric and Multivariate Approaches to Assess Water Quality Impairmen t Using Benthic Macroinvertebrates. Journal of the North American Benthological Society 16:833 852.
156 Rhoads BL, Wilson D, Urban M, Herricks EE. 1999. Interaction Between Scientists and Nonscientists in Community Based Watershed Management: Emergence of the Concept of Stream Naturalization. Environmental Management 24:297308. Robinson J. Stream ecology basics: A sampling training guide. University of Tennessee Knoxville and Kanugalihi Biological Consulting, for the Stream Monitoring Information Exchange a nd Discover Life in America. [Internet]. Available from: http://www.discoverlifeinamerica.org/dlia/education/activities_stream_ecology.pdf Rockstrm J, Steffen W, Noone K, Persson A, Chapin FSI, Lambin E, Lenton TM, Scheffer M, Folke C, Schellnhuber HJ, et al. 2009. Planetary boundaries: exploring the safe operating space for humanity. Ecology and Society 14:32. Rosenberg DM, Resh VH eds. 1993. Freshwater biomonitoring and benthic macroinvertebrates. New York: Chapman and Hall Roy AH, Dybas AL, Fritz KM, Lu bbers HR. 2009a Urbanization affects the extent and hydrologic permanence of headwater streams in a midwestern US metropolitan area. Journal of the North American Benthological Society 28:911 928. Roy AH, Purcell AH, Walsh CJ, Wenger SJ. 2009b Urbanizati on and stream ecology: five years later. Journal of the North American Benthological Society 28:908 910. Roy AH, Rosemond AD, Leigh DS, Paul MJ, Wallace JB. 2003. Habitat Specific Responses of Stream Insects to Land Cover Disturbance: Biological Consequenc es and Monitoring Implications. Journal of the North American Benthological Society 22:292 307. Royer TV, Tank JL, David MB. 2004. Transport and Fate of Nitrate in Headwater Agricultural Streams in Illinois. Journal of Environment Quality 33:1296. Rushton BT. 2001. Low Impact Parking Lot Design Reduces Runoff and Pollutant Loads. J. Water Resour. Plng. and Mgmt. 127:172. Schelske CL, Stoermer EF, Conley DJ, Robbins JA, Glover RM. 1983. Early Eutrophication in the Lower Great Lakes: New Evidence from Biogeni c Silica in Sediments. Science 222:320322. Schrader CC. 1995. Rural greenway planning: the role of streamland perception in landowner acceptance of land management strategies. Landscape and Urban Planning 33:375390. Schwarzenbach RP, Escher BI, Fenner K, Hofstetter TB, Johnson CA, von Gunten U, Wehrli B. 2006. The Challenge of Micropollutants in Aquatic Systems. Science 313:1072 1077.
157 Schwenneker BW 1985. The contribution of allochthonous and autochthonous organic material to aquatic insect secondary pr oduction rates in a north temperate stream. Ph.D. dissertation, University of Notre Dame, Notre Dame, Indiana. 273 pp. Sedley DN. 2003. Platos Cratylus. Cambridge, UK: Cambridge University Press Smil V. 2004. Enriching the Earth: Fritz Haber, Carl Bosch, and the Transformation of World Food Production. Cambridge, MA: MIT Press Smith RF, Alexander LC, Lamp WO. 2009. Dispersal by terrestrial stages of stream insects in urban watersheds: a synthesis of current knowledge. Journal of the North American Benthol ogical Society 28:10221037. Spangler PJ. 1961. Notes on the Biology and Distribution of Sperchopsis tessellatus (Ziegler) (Coleoptera: Hydrophilidae). The Coleopterists Bulletin 15:105 112. Stanford JA. 2007. Landscapes and Riverscapes. In: Methods in Str eam Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 321. Steinman AD, Lamberti GA, Leavitt PR. 2007j. Biomass and Pigments of Benthic Algae. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 357379. Steinman AD, Mulholland P J. 2007. Phosphorus Limitation, Uptake, and Turnover in Benthic Stream Algae. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 187 212. Strauss EA, Lamberti GA. 2000. Regulation of Nitrification in Aquatic Sediments by Organic Car bon. Limnology and Oceanography 45:18541859. Stream Solute Workshop. 1990. Concepts and Methods for Assessing Solute Dynamics in Stream Ecosystems. Journal of the North American Benthological Society 9:95 119. Suren AM, McMurtrie S. 2005. Assessing the ef fectiveness of enhancement activities in urban streams: II. Responses of invertebrate communities. River Res. Applic. 21:439453. Tank JL, Bernot MJ, Rosi Marshall EJ. 2007. Nitrogen Limitation and Uptake. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 213 238. Tank JL, Rosi Marshall EJ, Griffiths NA, Entrekin SA, Stephen ML. 2010. A review of allochthonous organic matter dynamics and metabolism in streams. J. N. Am. Benthol. Soc. 29:118146.
158 Tennis P. 2004. Pervious concrete pa vements. Second printing rev. Skokie Illinois: Portland Cement Association U.S. EPA. 1995. The Great Lakes Today: Concerns. In: The Great Lakes: An Environmental Atlas and Resource Book. 3rd ed. Chicago, IL: U.S. EPA Great Lakes National Program Office. Available from: http://www.epa.gov/glnpo/atlas/glat ch4.html U.S. EPA. 2009. Draft 2009 Update Aquatic Life Ambient Water Quality Criteria For Ammonia Freshwater. Washington, D.C.: U.S. Environmental Protection Agency; Office of Water Available from: htt p://water.epa.gov/scitech/swguidance/standards/criteria/aqlife/pollutants/ammo nia/upload/2009_12_23_criteria_ammonia_2009update.pdf U.S. EPA. 2010. Causal Analysis/Diagnosis Decision Information System (CADDIS). Office of Research and Development, Washingt on, DC [Internet]. Available from: http://www.epa.gov/caddis/ Vannote RL, Minshall GW, Cummins KW, Sedell JR, Cushing CE. 1980. The River Continuum Concept. Can. J. Fish. Aquat. Sci. 37:130137. Vermonden K, Leuven RSEW, van der Velde G, van Katwijk MM, Ro elofs JGM, Jan Hendriks A. 2009. Urban drainage systems: An undervalued habitat for aquatic macroinvertebrates. Biological Conservation 142:1105 1115. Vieira NKM, Poff NL, Carlisle DM, Moulton II SR, Koski ML, Kondratieff BC. 2006. A Database of Lotic Inve rtebrate Traits for North America. Reston, VA: U.S. Geological Survey Available from: http://pubs.usgs.gov/ds/ds187/pdf/ds187.pdf Walker B, Meyers JA. 2004. Thresholds in ecological and social ecological systems: a developing database. Ecology and Society [Internet] 9. Available from: http://www.ecologyandsociety.org/vol9/iss2/art3/ Walsh CJ, Fletcher TD, Ladson AR. 2005 a Stream Restoration in Urban Catchments through Redesigning Stormwater Systems: Looking to the Catchment to Save the Stream. Journal of t he North American Benthological Society 24:690 705. Walsh CJ, Fletcher TD, Ladson AR. 2009. Retention Capacity: A Metric to Link Stream Ecology and Storm Water Management. J. Hydrologic Engrg. 14:399. Walsh CJ, Roy AH, Feminella JW, Cottingham PD, Groffman PM, Morgan RP. 2005b The urban stream syndrome: current knowledge and the search for a cure. Journal of the North American Benthological Society 24:706 723. Walsh CJ. 2004. Protection of in stream biota from urban impacts: minimise catchment imperviousne ss or improve drainage design? Mar. Freshwater Res. 55:317326.
159 Ward JV. 1989. The Four Dimensional Nature of Lotic Ecosystems. Journal of the North American Benthological Society 8:2 8. Webster JR, Gurtz ME, Hains JJ, Meyer Judy L., Swank WT, Waide JB, Wa llace JB. 1983. Stability of Stream Ecosystems. In: Stream Ecology. Plenum Publishing. Webster JR, Patten BC. 1979. Effects of Watershed Perturbation on Stream Potassium and Calcium Dynamics. Ecological Monographs 49:51 72. Webster JR, Valett H. M. 2007. S olute Dynamics. In: Methods in Stream Ecology. 2nd ed. Burlington, MA: Academic Press. pp. 169 185. Wenger SJ, Roy AH, Jackson CR, Bernhardt ES, Carter TL, Filoso S, Gibson CA, Hession WC, Kaushal SS, Mart E, et al. 2009. Twenty six key research questions in urban stream ecology: an assessment of the state of the science. Journal of the North American Benthological Society 28:1080 1098. Whiting ER, Clifford HF. 1983. Invertebrates and urban runoff in a small northern stream, Edmonton, Alberta, Canada. Hydr obiologia 102:7380. Woese CR. 2004. A New Biology for a New Century. Microbiol. Mol. Biol. Rev. 68:173186. Wolfram S. 2002. A New Kind of Science. 1st ed. Champaign, IL: Wolfram Media WDNR. 2009. A Water Scavenger Beetle (Sperchopsis tessellatus). Endan gered Resources Program Species Information [Internet]. Available from: http://dnr.wi.gov/ORG/LAND/ER/biodiversity/index.asp?mode=info&Grp=6&Sp ecCode=IICOL71010 WSBT. 2010. South Bend unveils plan to revitalize south side neighborhood through Bowman Creek. WSBT News, Weather, Sports [Internet]. Available from: http://www.wsbt.com/news/local/92473294.html
160 Chapter 7: Appendices 7.1 Appendix I: Detailed study site information Figure 7. 1: a) LOCST site, a ditch next to a construction area. b) Map from Google Earth LOCST site named for Locust Rd. crossing GPS: 4137'17.09"N 8616'56.12"W Nearest road crossing: Jackson Rd. and Locust Rd. This site (Fig. 6.1) was the farthest upstream sampled. It was also the last site sampled. From July 30 through Aug. 2, MiniSonde 5 # UM40 was deployed in the roadcrossing a) b)
161 culvert at this site, concurrently with other MS5s at URB and GOLF. Chemical and biological samples were taken on Aug. 2. Subsites in this reach were downstream of the road crossing culvert visible in the photo above. Photo is facing south and upstream. Substrate was very silty, most likely due to sediment erosion from the subdivision being constructed in a large area to the southeast of this road crossing. Downstrea m portion of reach passing near houses with mowed lawns up to stream; somewhat shaded by trees. Many ranid frogs observed (likely Rana clamitans melanota [Rafinesque, 1820]), along with mud chimneys most likely constructed by burrowing crayfish.
162 Figure 7. 2: URB site, in a wooded area. b) Map from Google Earth URB site named for the Urbanskis, a very friendly couple living near the stream GPS: 4138'6.01"N 8616'11.10"W Nearest road crossing: Keria Tr. and Serbian Ln. This site (Fig. 6.2) was downstream of LOCST and upstream of MAIN. It was the least disturbed site, as far as the surrounding ecosystem. In the URB reach, the stream flows through a heavily wooded area, with significant but not overwhelming large woody debris in the channel. Many sandbanks, riffles, and pools are present in this section, which meanders much more than channelized portions farther up and downstream. Some b) a)
163 bank undercutting and erosion were observed. Chemical and biological samples were collected on July 22. MiniSonde 5 # UM55 was deployed here from July 28 30 (concurrently with other MS5s at MAIN and LWE), and was reset and redeployed from July 30 Aug. 2 (concurrently with other MS5s at LOCST and GOLF). Mini Sonde was deplo yed in midchannel of deep run, anchored with concrete block. Wood frogs ( Rana sylvatica LeConte, 1825) were observed, as well as extensive chimney building crayfish activity, even extending into the floodplain far from the stream channel. Mr. Urbanski repo rted digging up a crayfish burrow down to 18 ft. depth and finding a blind white crayfish at the bottom. (See photo of Mr. Urbanski and the author discussing a sample of stream invertebrates on p. x i v.)
164 Figure 7. 3: a) MAIN site, above the first major cu lvert. b) Map from Google Earth MAIN site named for Main St. crossing GPS: 4138'53.99"N 8615'5.25"W Nearest road crossing: S. Main St and W. Fairview Ave. Photo (Fig. 6.3a) is facing northeast and downs tream from a position southwest (upstream) of the culvert running under Main St. That is the first major culvert on the stream, marking the edge of the urbanized section of South Bend (Fig. 6.3b) This site had an interesting substrate composition, with ma ny old broken bricks. Trash and debris were very visible. This site experienced very high flows during a rain event between July 14 and July 15 (see Fig. 1.1). Chemical and biological samples were collected on July 21. b) a)
165 MiniSonde 5 # UM40 was deployed here from July 14 16 (concurrently with other MS5s at DAYTN and LWE) and from July 28 July 30 (concurrently with other MS5s at MAIN and LWE). Large amounts of algae observed here. Figure 7. 4: a) GOLF sit e, a golf course between two culverts. b) Map from Google Earth. GOLF site named for Studebaker Golf Course GPS: 4139'9.71"N 8614'33.45"W Nearest road crossing: Accessible via E. Ewing Ave or High St. Photo (Fig. 6.4a) is facing downstream (northeast) toward a footbridge in the middle of b) a)
166 the reach, which flo ws through the Studebaker Golf C ourse. The reach (Fig. 6.4b) is bounded on the upstream side by a long culvert that flows under the parking lot and running track of Riley High School, and is bounded on the downstream side by another long culvert that flows under the schools baseball field. Chemical and biological samples were collected on July 21. MiniSonde 5 # UM56 was deployed here from July 30 Aug 2, concurrently with other MS5s at LOCST and U RB. The MiniSonde was anchored just inside the culvert at the downstream end of this reach. This site was almost entirely unshaded by trees except at the extreme ends near the culverts, and extensive algal growth was observed in sunny areas. The stream ban ks were mowed to golf course standards and had been stabilized with riprap (granite rocks). Leeches ~5 7cm in length were observed here. Figure 7. 5: a) DAYTON site, the outflow of a long culvert. b) M ap from Google Earth a) b)
167 DAYTON site named for Dayton St. crossing GPS: 4139'21.67"N 8614'18.95"W Nearest road crossing: E. Dayton St. and High St. Photo (Fig. 2.8b) is facing upstream (southwest) into the culvert that flows out from under the Riley High School baseball fields. This site is downstream from the GOLF sampling site, separated by the golf course reach and a long culvert (Fig. 2.8a) Due to time constraints, chemical and biological samples were not collected from this site, but the RAV site is directly downstream with no culverts in between. MiniSonde 5 # UM55 was deployed at this site from July 14 16, concurrently with other MS5s at MAIN and LWE. The MiniSonde was anchored in the culvert channel, far enough into the tunnel that it would not be visible to a casual observer without a light. Large fish (potentially up to 30cm) were observed taking refuge in the end of this culvert and fleeing upstream when disturbed. These were most likely creek chub ( Semotilus atromaculatus [Mitchill, 1818]). S chools of smaller creek chub were visible throughout the section of stream below the culvert. Evidence of residential and industrial trash dumping observed. We talked to Ellen, a resident who had lived in the same house near the Dubail St. stream crossing for 35 years. She reported many wildlife sightings in the creek area, including a great blue heron and even a salmon. (Salmon are a natural part of this regions waterways, but there are significant fish barriers between this section of Bowman Creek and it s juncture with the St. Joseph River downstream.)
168 Figure 7. 6: a) RAV site, in Ravina Park. b) Map from Google Earth RAV site named for Ravina Park GPS: 4139'32.96"N 8614'15.92"W Nearest road cros sing: E. Broadway St. and Dale Ave. This site (Fig.6.6) is directly downstream from the DAYTON culvert, with no intervening culverts. This photo is facing upstream (south) toward the Indiana Ave. crossing. This reach is bounded on the upstream side by the Indiana Ave. crossing and on the downstream side by the start of a long culvert that runs under a Pepsi bottling facility and the NIPSCO plant. Chemical and biological samples were collected from this site on July a) b)
169 23. This site was not used for MiniSonde deployment due to insufficient water depth in the downstream culvert. (Sonde data from the DAYTON site and sample data from this site can be used to help characterize the combined reach.) This site flowed through a municipal park, but the stream channel its elf was heavily clogged with large woody debris (tree trunks and branches) that were accumulating rafts of trash and blocking the flow of water.
170 Figure 7. 7: a) NIPSCO culvert outflow. b ) Map from Google Earth. c) LWE site LWE site named for Lincolnway East crossing GPS: 4139'48.44"N 8614'12.37"W Nearest road crossing: Main St. and Lincolnway East This reach (Fig. 6.7c) is bounded on the upstream side by the outflow of the NIPSCO (No rthern Indiana Public Service Company) culvert (Fig. 6.7a), which is downstream of the RAV site. The reach is bounded on the downstream side by the Lincolnway East a) b) c)
171 crossing just before Bowman Creek flows into the St. Joseph River. This photo (Fig. 6.7c) is facing downstream (northeast). It was taken downstream of the NIPSCO culvert outflow and a large culvert which supported elevated train tracks crossing the stream. Another set of elevated train tracks crosses the creek at a culvert downstream of this posi tion (not visible in photo). Chemical and biological samples were collected at this site on July 23. An additional chlorophyll a sample was collected in the middle of the second railroad crossing culvert at this site (referred to as RRCULV sample). MiniSon de 5 # UM56 was deployed in the middle of the second railroad crossing culvert from July 14 July 16 (concurrently with other MS5s at MAIN and DAYTON) and from July 28 Aug 2 (concurrently with other MS5s at URB and MAIN. This site was chosen due to in sufficient water depth at the NIPSCO culvert outlet at the upstream portion of the reach. Schools of creek chub were observed in this reach, and a single juvenile rainbow trout ( Onchorhynchus mykiss Walbaum, 1792) was seen in the downstream section. Extens ive coverage of hairlike brown algae/bacteria was observed in sunny areas of the stream channel on July 14, but much of this was swept away by a scouring flood on July 15. A wooded area near the access road to this site was occupied by what appeared to be a well established hobo camp, in close proximity to two sets of train tracks. Unfortunately, safety concerns and time constraints prevented us from talking to the residents about their observations of the nearby stream.
172 7.2 Appendix II: Invertebrate counts Table 7. 1: Raw macroinvertebrate data for sites LOCST, URB, and MAIN LOCST URB MAIN 1 2 3 1 2 3 1 2 3 Highest taxa Family Genus Species Ephemeroptera Unidentified 2 Baetidae 2 1 49 9 Baetis 3 2 4 Trichoptera Unidentified larva 8 3 Unidentified pupa 1 4 2 3 4 1 2 Hydropsychidae 9 94 41 57 44 30 47 Cheumatopsyche 13 9 3 21 7 Hydropsyche 16 18 44 27 27 36 Philopotamidae 2 Chimarra Hydroptilidae Hytroptila or Ochrotrichia? 2 1 Neotrichia? 4 (empty cases sandy)* 2 76 4 4 (empty cases Leucotrichia?)* Diptera Unidentified 5 Simuliidae 17 22 4 1 1 2 Tipulidae Antocha Chirnomidae 12 33 53 2 20 26 37 22 101 Ceratopogonidae Dixidae 2 Empididae 1 3 4 Coleoptera Elmidae Unknown larvae 27 4 11 27 18 20 101 67 240 Macronynchus glabratus 1 2 Optioservus fastiditus 11 11 10 10 11 1 11 Stenelmis crenata 17 6 1 4 6 14 6 5 1
173 *Not included in total counts **Potentially P. duodecimpunctatus Dubiraphia 1 Haliplidae Peltodytes lengi** Hemiptera Corixidae 1 Veliidae Macroveliidae 1 Collembo la 1 2 1 Acarina Amphipoda Gammaridae Gammarus 416 137 40 83 57 173 2 4 11 Isopoda Asellidae Caecidotea 1 1 Decapoda (crayfish) 1 Turbellaria 3 3 2 2 1 Nematoida? 1 Annelida 15 2 LWE worms? Oligochaeta 8 1 1 1 Hirudinea Bivalvia (Tiny clams) 2 2 3 Gastropoda (Left handed pulmonate snails) 2 3 5 7 Planorbidae (Ramshorn snails) 1 6 1 4 Ancylini 1 (Rig ht handed snails) 1 TOTAL ABUNDANCE 549 238 140 255 220 351 241 190 490 TOTAL PER SITE 927 826 921 SAMPLE MEAN 309.0 275.3 307.0
174 Table 7. 2: Raw macroinvertebrate dat a for sites GOLF, RAV, and LWE GOLF RAV LWE TOTAL 1 2 3 1 2 3 1 2 3 Highest taxa Family Genus Species Ephemeroptera Unidentified 2 Baetidae 4 4 25 5 5 1 3 1 109 Baetis 3 15 27 Trichoptera Unidentified lar va 2 5 19 Unidentified pupa 1 1 6 24 Hydropsychidae 38 5 27 23 1 416 Cheumatopsyche 4 2 4 9 2 3 1 78 Hydropsyche 16 29 19 11 3 37 5 2 290 Philopotamidae 2 Chimarra 1 1 Hydroptilidae Hytroptila or Ochrotrichia? 11 4 2 20 Neotrichia? 5 9 (empty cases sandy)* 6 6 19 1 1 4 149 (empty cases Leucotrichia?)* 12 Diptera Unidentified 1 1 1 8 Simuliidae 1 2 13 3 1 67 Tipuli dae Antocha 1 1 Chirnomidae 15 86 140 15 12 42 3 1 21 641 Ceratopogonidae 2 2 Dixidae 2 Empididae 4 1 1 14 Coleoptera Elmidae Unknown larvae 103 4 34 6 4 7 1 674 Macronynchus glabratus 3
175 *Not included in total counts **Potentially P. duodecimpunctatus Op tioservus fastiditus 1 66 Stenelmis crenata 13 1 3 2 1 1 81 Dubiraphia 1 Haliplidae Peltodytes lengi** 1 1 Hemiptera Corixidae 1 Veliidae 1 1 Macroveliidae 1 Collembola 4 Acarina 1 1 Amphipoda 1 1 Gammaridae Gammarus 1 1 925 Isopoda Asellidae Caecidotea 4 2 1 9 Decapoda (crayfish) 1 Turbellaria 1 1 13 Nematoida? 3 2 6 Annelida 1 18 L WE worms? 1 61 62 Oligochaeta 3 4 1 2 1 22 Hirudinea 1 1 3 3 8 Bivalvia (Tiny clams) 12 5 3 2 1 1 33 Gastropoda (Left handed pulmonate snails) 2 4 3 6 4 3 3 1 1 54 Planorbidae (Ramshorn snails) 1 1 16 Ancylini 1 1 1 4 (Right handed snails) 1 3 TOTAL ABUNDANCE 220 159 284 89 58 121 18 7 111 3741 TOTAL PER SITE 663 268 136 SAMPLE MEAN 221.0 89.3 45.3 207.8
176 (ta panta rhei) everything flows