The Effects of Hydrologic Pulsing on Created Riparian

Home , Effluent

EFFECTS OF A PULSING HYDROPERIOD ON A CREATED RIPARIAN RIVER DIVERSION WETLAND

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of the Ohio State University

Daniel F. Fink, M.S.

* * * * *

The Ohio State University

2007

Dissertation Committee: Approved by

William J. Mitsch, adviser

M. Siobhan Fennessy ______Jay F. Martin Adviser

Franklin W. Schwartz Environmental Science Graduate Program

ABSTRACT

In this study, succession, hydroperiod, and water-quality dynamics are documented for a whole-ecosystem study involving a 3-ha created riparian wetland at the

Schiermeier Olentangy River Wetland Research Park at The Ohio State University in

Columbus, Ohio USA during 2003 through 2006. The effect of delivering influent water to the wetland as seasonal pulses or as a continuous steady flow on the hydroperiod, water chemistry, avian use, plant primary production, and plant community structure within the wetland was investigated. A simulation model was then developed using short-term measurements of river and wetland stage to predict long-term patterns of succession and plant development within the wetland. This wetland typically receives seven or eight natural flood pulses each year from the Olentangy River. Of 21 species planted in 1997, only Scirpus americanus and Juncus effusus remained as important macrophyte species during the study six to seven years after planting. Typha spp.

(angustifolia and latifolia), colonized naturally and was the dominant macrophyte in most of the wetland during this study.

Mean nutrient reductions per flood pulse for nitrate-nitrite, total Kjeldahl nitrogen

(TKN), soluble reactive phosphorus (SRP), and total phosphorus (TP) during 2003-2004 were 0.71 g-N m-2, 0.92 g-N m-2, 0.016 g-P m-2, and 0.08 g-P m-2 respectively. The

- annual reductions of N-NO3 , TN, P-SRP, and TP were 74%, 41%, 46%, and 31% by

ii - mass. A greater removal of NO3 and TP occurred in the emergent marsh section of the wetland than the open water section. Conversely TKN increased through the emergent marsh and decreased through the open water. Overall, the oxbow design appeared to be successful in ecological terms and similar diversion wetlands are recommended for other locations to examine their function under different climates and hydroperiods.

While hydrologic pulses affect wetland function and may enhance productivity and contribute stability to ecosystems, the overall effects of these pulses on biogeochemical processes in riparian river diversion wetland ecosystems have not been clearly demonstrated experimentally at an ecosystem scale. The natural hydroperiod of the riparian wetland was varied in 2005; its natural flood pulses were removed and replaced with an artificial steady-flow supplied by submersed pumps. The wetland received 8 natural river pulses from April 2004 – March 2005, then a steady flow of river water was pumped through the wetland from April 2005 – March 2006. In 2004 the wetland received a total of 27 m3 m-2 yr-1 (cubic meters of influent water per square meter of wetland area per year) of inflow followed by 20 m3 m-2 yr-1 of inflow in the steady flow year 2005. The removal rate was higher for nitrate-nitrogen, total nitrogen, and total phosphorus during the year with flood-pulsing than during the steady-flow year. The only nutrient that did not have a difference in removal rate in the two years was soluble reactive phosphorus. There were differences in spatial dynamics of most of the nutrients in flood-pulsing and non-pulsing years and between wet and dry seasons in all years of the study. In all cases, however, total nitrogen concentrations in the wetland increased through the emergent marsh and then decreased across the open water basin.

iii There was greater avian use of the created oxbow wetland during pulsing than in steady flow conditions. The guild of bird species most affected by the removal of a pulsing hydroperiod was shorebirds, which had a 75% reduction in the number of species observed during steady-flow. Peak observed shorebird use during the pulsing hydroperiod year corresponded to dry conditions in the wetland in the later part of the growing season during shorebird migration time.

Primary productivity of macrophytes was compared among the 2004 pulsing hydroperiod year, the 2005 stead-flow year, and in 2006 when natural pulsing was again restored (6 natural river pulses occurred from April 2006 – September 2006). The wetland was significantly more productive (α = 0.05) during pulsing compared to steady flow conditions. Wetland macrophyte productivity decreased from 7,560 kg yr-1`to 6,544 kg yr-1and total macrophyte areal coverage decreased from 32.3% to 25.7% of the wetland after the pulsing hydroperiod was replaced by the steady-flow hydroperiod.

When a flood pulse hydroperiod was restored in 2006, total net macrophyte productivity rebounded to 9,916 kg yr-1 and areal coverage returned to 32.1%. There was a positive correlation between the flashiness of the wetland hydroperiod and productivity (r = 0.75).

A pulsing hydroperiod should enable mitigation wetlands to reach target macrophyte coverage goals more easily.

A simulation model was developed use short-term hydrologic data, such as the data collected in this study, to predict long-term successional development in a created riparian wetland ecosystem. A hydrologic submodel predicted wetland water stage and depth based on precipitation, potential evapotranspiration, river stage, overland outflows, and groundwater exchange. Vegetation submodels calculated the growth of trees and

iv emergent macrophytes based on water depth and the duration and timing of flooding.

The model also predicted the state of the ecosystem annually as being open water, mudflat, marsh, or forested wetland. Data from 2004 were used to calibrate the models.

Simulation results were compared with actual data from 2003, 2005, and 2006 to validate the model. When the wetland inflow was pulsing, the model predicted water depth within 6.0 cm 75% of the time and when it was a steady flow the model predicted water depth within 2.5 cm 75% of the time.

Simulations of floral succession were evaluated at 8 (current wetland age) and

100 years. Simulations for current wetland age correctly predicted the state of the ecosystem. One hundred year simulations showed annual variations in macrophyte coverage over much of the wetland, but a consistent growth of trees on the wetland fringe. Simulations also predicted a larger area of mudflat if flashiness increased and a greater area of open water if flashiness decreased or if beavers dammed the outflow.

When a sediment accretion routine was added to the model, 100-yr simulations predicted that the wetland will eventually become completely covered with macrophytes and that the area of forested wetland will cover not only the outer fringe, but much of the internal portion of the wetland basin as well. Adding several model features including a dispersal model for propagules and a shorter time step for vegetation could improve model function.

Dedicated to my family and friends, who have cared so much for me.

vi ACKNOWLEDGEMENTS

I gratefully thank my advisor, William Mitsch for his wisdom and his patience while guiding me through this process. I also thank Siobhan Fennessy, Jay Martin, and

Frank Schwartz for serving on my dissertation committee and for their support and insight. The quality of this project is much better for their contributions. I am also grateful to my fellow “wetlanders” throughout my stay at Ohio State. Their help and friendship made my time here at The Ohio State University a joyful time in my life.

Funding for my education and research came from The Environmental Sciences Graduate

Program, the National Science Foundation’s GK-12 Fellowship program, the U.S.

Department of Agriculture NRI CSREES Award 2003-35102-13518 and a Payne grant from the Ohio Agricultural Research and Development Center of The Ohio State

University. Lastly I thank my friends and family for supporting me through my long journey in academia. I especially thank my wife Colleen for her love, support, and patience. I have never been happier then since you have been in my life.

vii VITA

12 December, 1975………………… Born – Elyria, Ohio

1998………………………………... B.A., Physics and Environmental Studies

Ohio Wesleyan University

Delaware, Ohio

1998-2001………………………….. Graduate Teaching Assistant,

The Ohio State University

2001………………………………... M.S. Environmental Science Graduate

Program, The Ohio State University

Columbus, Ohio

2002 – 2003………………………... Graduate Teaching and Research Assistant

The Ohio State University, Columbus Ohio

2003 – 2004………………………... GK-12 National Science Foundation

Teaching Fellow, Columbus, Ohio

2005 – 2006………………………... Graduate Teaching and Research Assistant

The Ohio State University, Columbus Ohio

viii PUBLICATIONS

Peer-reviewed journal articles

Fink, D.F. and W.J. Mitsch. 2004. Seasonal and storm event nutrient removal by a created wetland in an agricultural watershed. Ecological Engineering 23:313-325.

Fink, D.F. and W.J. Mitsch. in press. Hydrology and nutrient biogeochemistry in a created river diversion oxbow wetland. Ecological Engineering.

Zhang, L., W.J. Mitsch, and D.F. Fink. 2004. Hydrology, water quality, and restoration potential for the Upper Big Darby Creek, Central Ohio. Ohio Journal of Science 105:46-56.

Publications in the Olentangy River Wetland Research Park Annual Reports

Anderson, C.J., D.F. Fink, W.J. Mitsch. 2003. Vegetation establishment in the mitigation billabong at the Olentangy River Wetland Research park, 2000-2002. In: Mitsch, W.J., L. Zhang, C.J. Anderson (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2002, pp. 101-106

Gilbert, J.M., D.F. Fink, and M. Greene. 1999. Soil properties of three newly created wetlands. In: Mitsch, W.J., V. Bouchard (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 1998, pp. 113-118.

Fink, D.F. and W.J. Mitsch. 2000. Effectiveness of a newly constructed wetland on agricultural run-off. In: Mitsch, W.J., L. Zhang (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2001, pp. 191-196.

Fink, D.F. and W.J. Mitsch. 2001. Wetlands for controlling nonpoint source pollution from agriculture: Indian Lake Wetland Demonstration Project, Logan County, OH. In: Mitsch, W.J., L. Zhang (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2000, pp. 161-178.

Fink, D.F. and W.J. Mitsch. 2005. Fish and amphibian abundance in created riparian marshes with pulsing hydrology. In: Mitsch, W.J., L. Zhang, and A.E. Altor (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2004, pp. 125-128.

Fink. D.F. and W.J. Mitsch. 2005. Hydrology, biogeochemistry, and plant community development in a created river diversion oxbow wetland in the Ohio River Basin, USA. In: Mitsch, W.J., L. Zhang, A.E. Altor (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2004, pp. 137-148.

ix Fink, D.F. 2006. Analysis of Phragmites australis haplotypes at the ORWRP based on morphological characteristics. In: Mitsch, W.J., L. Zhang, and C. Tuttle, K. Jones (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2005, pp. 127-132.

Fink, D.F. and W.J. Mitsch. 2006. The effect of the removal of hydrologic pulsing on a river-diversion riparian wetland. In: Mitsch, W.J., L. Zhang, and C. Tuttle, K. Jones (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2005, pp. 133-148.

Mitsch, W.J., D.F. Fink, L. Zhang, 2002. Net primary productivity of macrophyte communities after eight growing seasons in experimental planted and unplanted marshes. In: Mitsch, W.J., L. Zhang (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2001, pp. 43-48.

Mitsch, W.J., L. Zhang, N. Dillon, and D.F. Fink. 2004. Biogeochemical patterns of created riparian wetlands: Tenth year results (2003). In: Mitsch, W.J., L. Zhang, and C. Tuttle (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2003, pp. 59-68.

Zhang, L., W.J. Mitsch, D.F. Fink. 2004. Hydrology, water quality, and restoration potential for the Upper Big Darby Creek, Central Ohio. In: Mitsch, W.J., L. Zhang, C. Tuttle (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2003, pp. 207-218.

Zhang, L., W.J. Mitsch, C.L. Tuttle, and D.F. Fink. 2006. Water budgets of the two Olentangy River experimental wetlands in 2005. In: Mitsch, W.J., L. Zhang, and C. Tuttle, K. Jones (eds.), Olentangy River Wetland Research Park at the Ohio State University: Annual Report 2005, pp. 25-36.

Technical Reports

Mitsch, W.J., and D.F. Fink. 2001. Wetlands for Controlling Nonpoint Source Pollution from Agriculture. Indian Lake Wetland Demonstration Project Logan County, OH. Final Report. 18 pp.. Indian Lake Watershed Project, Bellfontaine, OH.

FIELDS OF STUDY

Major Field: Environmental Sciences

Specializations: Wetland ecology and restoration Systems ecology and ecosystem modeling

TABLE OF CONTENTS

Abstract………………………………………………………………………………….. ii

Dedication………………………………………………………………………………..vi

Acknowledgements……………………………………………………………………...vii

Vita……………………………………………………………………………………...viii

List of Figures…………………………………………………………………………...xv

List of Tables………………………………………………………………………….... xx

1. INTRODUCTION…………………………………………………………………… 1

1.1 Rationale and significance…………..……………………………………….. ...2 1.2 Creation, restoration, and mitigation……………………………………….. ….2 1.3 Ecosystem modeling…………………………………………………………… 4 1.4 A whole-ecosystem approach………………………………………………….. 4 1.5 Goals and objectives………………...…………………………………...…….. 5 1.6 Literature cited……………………………………………………………….…8

2. NUTRIENT BIOGEOCHEMISTRYAND MACROPHYTE ASEMBLAGES IN A CREATED WETLAND……………………………………………….……… ……12

2.1 Abstract……………………………………………………………………..…12 2.2 Introduction……………………………………………………………..……..13 2.3 Methods…………………………………………………………………….… 16 2.3.1 Site description……………………………………………………….… 16 2.3.2 Hydroperiod.……………………………………………………………. 17 2.3.3 Water quality………………………………………………………..….. 18 2.3.4 Statistical methods……………………………………………………… 19 2.4 Results……………………………………………………………………… ...20 2.4.1 Hydrologic loading and hydroperiod ………….………………………..20 2.4.2 Vegetation…………………………………………………………….…21 2.4.3 Nutrients and sediments………………………………………………... 22 2.5 Discussion……………………………………………………………….…….27 2.5.1 Water quality dynamics…………………………………………..…….. 27 2.5.2 Spatial patterns…………………………………………………..……... 28

xi 2.5.3 Nutrient loading and retention rates……………………………………. 29 2.5.4 Vegetation dynamics in a diversion wetland…………………………… 30 2.6 Conclusions………………………………………………………………...… 30 2.7 Acknowledgements…………………………………………………..………..31 2.8 Literature cited…………………………………………………….….……….32

3. THE EFFECT OF REMOVING HYDROLOGIC PULSING ON A RIVER DIVERSION RIPARIAN WETLAND…………………………….…….... 53

3.1 Abstract………………………………………………………………………..53 3.2 Introduction………………………………………………………………..…..54 3.2.1 Goals and objectives……………………………………………………. 56 3.3 Methods…………………………………………...………………………….. 56 3.3.1 Site description…………………………………….…………………… 56 3.3.2 Hydrologic conditions.…………………………………...…………….. 57 3.3.3 Water quality……………………………………………………..…….. 58 3.3.4 Chemical analysis………………………………………………………. 59 3.3.5 Avian use……………………….………………………………………. 60 3.3.6 Statistical methods……………………………………………………… 61 3.4 Results……………………………………………………..…………………….62 3.4.1 Hydrologic loading and hydroperiod…………………….………….…..62 3.4.2 Nutrient loading and retention rates……………………………….…… 63 3.4.3 Seasonal and spatial nutrient dynamics………………………………… 64 3.4.4 Avian use………………………….……………………………………. 68 3.5 Discussion……………………………………...………………………………..69 3.5.1 The effect of pulsing……………………….…………………………… 69 3.5.2 Comparison to other river diversions………….……………………….. 70 3.5.3 Spatial patterns of nitrogen removal.……………..………………….….71 3.5.4 Management implications…………………………………….……...… 72 3.5.5 Avian effects………………………………………………………….. ...73 3.5.6 Conclusions………………………………………………………...… ...74 3.6 Acknowledgements……………………………………………………...…… 75 3.7 Literature cited…………………………………………………………….…..76

4. THE EFFECT OF REMOVINGAND REINTRODUCING HYDROLOGIC PULSES ON PRODUCTIVITY OF A RIVER-DIVERSION WETLAND……… 100

4.1 Abstract…………………………………….………………………………...100 4.2 Introduction…………………………………….…………………………….101 4.2.1 Goals and objectives……………………...…………………………… 102 4.3 Methods……………………………………………...…………………...… …103 4.3.1 Site description………………………………...……………………. ...103 4.3.2 Hydroperiod…………………………………………..…………….…. 104 4.3.3 Macrophyte productivity and diversity……………….…………….. ...105 4.3.4 Statistical methods……………………………………...……………. ..106 4.4 Results……………………………………………………………….……….. .106

xii 4.4.1 Hydrologic loading and hydroperiod…………………………………..106 4.4.2 Macrophyte communities…………………………………….……...... 107 4.4.3 Macrophyte productivity………………………...…………….….… ...107 4.5 Discussion……………………………………………………………………109 4.5.1 The effect of pulsing…………………………………………….…….. 109 4.5.2 Causes for variation in macrophyte productivity……………...……….111 4.5.3 Nutrients and wetland vegetation patterns………...………..………… 115 4.5.4 Management implications…………………………...….…………….. 115 4.6 Acknowledgements…………………………………………………………. 116 4.7 Literature cited…………………………………………………………...…..117

5. USING SHORT-TERM HYDROLOGY DATA TO PREDICT LONG- TERM SUCCESSIONAL DEVELOPMENT IN A RIPARIAN WETLAND..… .141

5.1 Abstract………………………………………………………………………141 5.2 Introduction…………………………………………………………..……... 142 5.2.1 Goals and objectives………………………………………...………… 145 5.2.2 Site description………………………………………………………... 146 5.3 Modelling Methods……………………...……………………………………. 146 5.3.1 Calibration……………………………………………………….……. 148 5.3.2 Sensitivity analysis…………………………………………………..... 148 5.3.3 Integration and modeling techniques…………………………………..149 5.4 Model structure…………………………………………………………… ...... 150 5.4.1 Hydrologic submodel……………………………………………..…....150 5.4.2 Vegetation submodels………………………………………………… 151 5.4.2.1 Sediment accretion subroutine…………………………………. …152 5.4.2.2 Tree growth submodel…………………………………………..… 152 5.4.2.3 Impact of trees on hydroperiod…………………………………… 154 5.4.2.4 Emergent macrophyte growth submodel……………………..…… 155 5.4.3 Ecosystem state determination…………………………………...…… 157 5.5 Results and Discussion………………………………………..…………….. 157 5.5.1 Simulations……………………………………….…………………… 157 5.5.1.1 Simulation 1 – baseline, 2004 hydroperiod…………….…………. 158 5.5.1.2 Simulation 2 – dramatic flood pulsing………………………..… ...158 5.5.1.3 Simulation 3 – pulsing removed, 2005 hydrologic data...……..…. .159 5.5.1.4 Simulation 4 – sediment accretion………………………………..,.160 5.5.1.5 Simulation 5 – beaver ponding…………………...……….……… 161 5.5.1.6 Simulation 6 – 10 years of Olentangy River data…………..……...161 5.5.1.6.1 Ecosystem succession…………………………………… ...161 5.5.1.6.2 Total productivity………………………………………… .163 5.5.2 Wetland succession……………………………………………...……. 164 5.5.3 Sensitivity analysis…………………………………..…….….………. 166 5.5.4 Model accuracy…………………………………...……….…..……… 167 5.5.5 Model limitations…………………………….………………...………168 5.6 Conclusions……………………………………..…………………...……….169 5.7 Acknowledgements………………………………………………………… .170

xiii 5.8 Literature cited……………………………...………………………………..171

References………………………………………………………………...………… ...206 Appendix A – STELLA Code…………………………………………...…………. …222

xiv

LIST OF FIGURES

Figure Page

2.1. Site map of the Wilma H. Schiermeier Olentangy River Wetland Research Park on The Ohio State University campus. The 2.8 ha created oxbow wetland is between the experimental wetlands and the bottomland hardwood forest. Grab-sample locations are marked with a white circle and the seven biomass sampling transects are marked with white lines.... ……..37

2.2. Annual hydrograph for a created diversion oxbow wetland and Olentangy River in central Ohio, USA for (a) 2003 and (b) 2004…….…………..…. ….....39

2.3. Dominant vegetation communities in the created oxbow wetland in a) 2003 and b) 2004. The area in the 2004 map marked with the dotted lines shows the extent of the spread Xanthium strumarium following prolonged drawdown after vegetation and productivity surveys in this study. The location of biomass sampling transects are marked on Fig. 2.1.……..…... …….41

2.4. Mean total phosphorus (TP) reduction during flood pulses in 2003 (15 of 17 total pulses) and 2004 (7 of 7 total pulses). A positive percentage indicates a net reduction in TP whereas a negative percentage indicates a net export in TP. Error bars show standard error. The number of TP samples measured during each pulse is indicated below each mark…….… …...43

- - 2.5. Kriging diagrams of the a) nitrate+nitrate (NO3 + NO2 ), b) total Kjeldahl nitrogen (TKN), c) total phosphorus (TP), and d) total suspended solids (turbidity-NTU) in a created diversion oxbow wetland in central Ohio. Nutrient sampling locations are shown in Fig. 2.1……………………… ……...45

2.6. Nitrate loading and export during the second major flood event of 2004. Net retention is the difference between the two. The flood event occurred February 3-15, 2004.………………………………………………...…… …….47

xv 3.1 Site map of the Wilma H. Schiermeier Olentangy River Wetland Research Park on The Ohio State University campus. The 2.8 ha created oxbow wetland is between the experimental wetlands and the bottomland hardwood forest. Grab-sample locations are marked with a white circle. The area to the north of the “created oxbow” label is predominantly emergent marsh and the area to the south of the label is predominantly open water.………………………………………………………….……… ..…82

3.2 Inflow (top) and outflow (bottom) control structures in the created oxbow wetland. The Red Field TideflexTM check valve, top, opens via water pressure when the river elevation is higher than the wetland and closes when the river elevation is lower than the wetland water level, and water pressure is removed. Water then flows back to the Olentangy River though an outflow control weir, bottom.…………………………………. ….…84

3.3 Water inflow rate for the created oxbow wetland between April 2004 and March 2006. April 2004 through March 2005 was a pulsing hydroperiod year while April 2005 through March 2006 was steady-flow hydroperiod year when flood-pulses were removed from the wetland……………….. …...…86

- 3.4 Kriging diagrams of the nitrate + nitrate (NO3 -N), total nitrogen (TN), soluble reactive phosphorus (SPR), total phosphorus (TP), and total suspended solids (TSS) in the created oxbow wetland showing annual mean concentrations for 2004 and 2005. Inflow of river water is on the left of each diagram; outflow is on right. Sampling locations are shown in Fig. 3.1.………………………………………………………………..…… …...88

- 3.5 Kriging diagrams of the nitrate + nitrate (NO3 -N), total nitrogen (TN), soluble reactive phosphorus (SPR), total phosphorus (TP), and total suspended solids (TSS) in the created oxbow wetland during the wet season in 2004 and 2005. Inflow of river water is on the left of each diagram; outflow is on right. Sampling locations are shown in Fig. 3.1..... …....90

- 3.6 Kriging diagrams of the nitrate + nitrate (NO3 -N), total nitrogen (TN), soluble reactive phosphorus (SPR), total phosphorus (TP), and total suspended solids (TSS) in the created oxbow wetland during the dry season in 2004 and 2005. Inflow of river water is on the left of each diagram; outflow is on right. Sampling locations are shown on Fig. 3.1…. …...92

3.7 Number of new avian species-observed per sampling effort in the created oxbow wetland with a pulsing hydroperiod (April 2004 – March 2005) and steady-flow hydroperiod (April 2005 – March 2006)…………………...... 94

xvi 4.1 Site map of the Wilma H. Schiermeier Olentangy River Wetland Research Park on The Ohio State University campus. The 2.8 ha created oxbow wetland is between the experimental wetlands and the bottomland hardwood forest. The locations of the seven biomass sampling transects are marked with white lines………………………………..……………… …..122

4.2 Hydrograph for the created oxbow wetland and the Olentangy River from April 2004 to September 2006.……………………….…………………. ……124

4.3 Macrophyte communities in the created oxbow wetland during pulsing (2004 and 2006) and steady flow (2005) conditions during peak biomass. The area in the 2004 map marked with dotted lines shows the spread of Xanthium strumarium following a prolonged drawdown that occurred after the biomass surveys were completed.………...………………...………. ….…126

4.4 Comparisons of: (a) the community diversity index and percent areal coverage of emergent macrophytes (r2 = 0.97), (b) the community diversity index and the Richardson-Baker flashiness index (r2 = 0.98), and (c) the total macrophyte production of the wetland and the Richardson- Baker flashiness index (r2 = 0.75)……………………………...………….. ….128

4.5 Total net aboveground primary productivity in the created oxbow wetland during pulsing (2004 and 2006) and steady flow (2005) during peak biomass. Labels a and b indicate statistical differences in the mean biomass between years (p = 0.36 between 2004 and 2005; p = 0.05 between 2004 and 2006; p = 0.02 between 2005 and 2006). Error bars indicate standard error……………………….…………………...………. …...130

4.6 Total production (kg yr-1) of the 5 dominant macrophyte communities in the created oxbow wetland in 2004 (pulsing hydrology), 2005 (steady flow hydrology), and 2006 (restored pulsing hydrology). The “mixed macrophyte” community is comprised predominately of: Eleocharis sp., Scirpus americanus, Juncus effusus, Leersia oryzoides, Sagittaria spp., Verbesnia alterniflora, and Cyperus strigosis..………………….……….. …...132

4.7 Effects of pulsing on wetland ecosystem services. Arrows indicate the direction of increased service for each process. A bi-directional arrow indicates uncertainty as to the effect of pulsing with the relative size of the arrowhead suggesting the general impact………………………..……….……134

4.8 Pattern of Typha net above ground primary productivity per unit area (NAPP, g dry weight m-2 yr-1) compared to the nitrate + nitrite concentration in the surface water from the inflow, through the emergent marsh, open water, and outflow of the created oxbow wetland.…...……. ……136

xvii

5.1 Hydrarch succession in wetlands with reversals of successional stage caused by allogenic disturbance ………………..…………….………….... ….179

5.2 Odum diagram of interactions between hydraulic forcing functions and potential ecosystem states (open water, mudflat, marsh, and forested wetland). The forcing functions are precipitation, solar inputs, river stage or pumped inflow, and the seed bank in the wetland. The ecosystem state selected as a result of water depth is indicated by the order in which they are located in the Odum diagram.…………………..……………………..…...181

TM 5.3 STELLA diagram of the hydrologic submodel. Inputs are river (Rstage), rainfall (Pt), and temperature (T). Output is the surface elevation of the water in the wetland (MSL) in both daily and in seasonal time steps. The hydrologic input to the model from the river can be switched from natural river flood pulsing (Rstage and Pipeinflow subroutine) to a predetermined artificial pumping rate (Pump)……………………………………..……….….183

5.4 STELLATM diagram of the woody vegetation submodel. Inputs are the surface elevation of the wetland water (MSLW) and the surface elevation of wetland sediments (MSLL). Output is the diameter (TD) and height (TH) of tree production and the hydraulic lift (H) due to the trees………. ..…..185

5.5 STELLATM diagram of the macrophyte vegetation submodel. Inputs are surface elevation of the wetland water (MSLW), the surface elevation of the wetland sediments (MSLL), the diameter of and trees (TD), and the hydraulic lift (HL). Output is the biomass production of the emergent macrophytes (MacB).………………………………………………………. ....187

5.6 Site map indicating the land surface elevation contours (MSL ft) at which succession simulations were conducted.…………………………………... .…189

5.7 Ecosystem state at the end of five different100 year simulations. Simulation results are a) 2004 hydroperiod, b) 2004 hydroperiod with the variance of the pulse peaks increased by 50% , c) 2005 hydroperiod (steady flow), d) 2004 hydroperiod with sediment accretion, and e) 2004 hydroperiod with the outflow weir height raised by 1 ft (0.3048 m)………. ....191

5.8 Time series of the simulated macrophyte and woody biomass in the created wetland according to a 100-year extrapolation of the stage of the Olentangy River from April 2004 – March 2005 with a sediment deposition submodel. Each graph represents a different elevation contour (Fig. 5.6) within the wetland basin, with a) corresponding to the 723.00 ft MSL contour, b) the 723.43 ft MSL contour, c) the 724.10 ft MSL contour, d) the 724.25 ft MSL contour, e) the 725.00 ft MSL contour, f) the 726.00 ft MSL contour, and g) the 727.00 ft MSL contour.………….……193

xviii

5.9 Time series of simulated macrophyte and woody biomass in the created wetland according to a 100-year extrapolation of the last 10 years of Olentangy River Stage (1996-2005). Each graph represents a different elevation contour (Fig. 5.6) within the wetland basin, with a) corresponding to the 723.00 ft MSL contour, b) the 723.43 ft MSL contour, c) the 724.10 ft MSL contour, d) the 724.25 ft MSL contour, e) the 725.00 ft MSL contour, f) the 726.00 ft MSL contour, and g) the 727.00 ft MSL contour.……………………………………………..……... ….195

5.10 Comparison of simulated and actual elevation of surface waters from April 2003 – March 2006. The model was calibrated for April 2004 – March 2005 and validated for April 2003 – March 2004 and for April 2005 – March 2006………………………………………………………………. ……197

xix

LIST of TABLES

Table Page

2.1 Frequency and duration of flood pulses in a created oxbow wetland on the Olentangy River in central Ohio, USA …………………………………….…...48

2.2 Estimated net aboveground primary productivity, based on peak biomass, in the created oxbow in 2004 for the 11 most dominant emergent macrophytes. The “mixed wetland vegetation” in this table refers to grouped FACW species that individually made up less than 0.1% of the total macrophyte biomass in the wetland….………………………...…….. …...49

2.3 Mean nutrient concentrations and turbidity [SRP = soluble reactive - - phosphorus, TP = total phosphorus, NO3 + NO2 = nitrate and nitrite, TN = total nitrogen, and TSS = total suspended solids] in the oxbow wetland created at the Olentangy River Wetland Research Park, 2003-2004, when - flooded river water is flowing through the wetland. TN = TKN + NO3 + - NO2 = (as N)………………………………………………………….. ..………50

- 2.4 Annual mean loading and removal of nitrate-nitrite (NO3 ), total nitrogen (TN), soluble reactive phosphorus (SRP), and total phosphorus (TP) during 2004 in the river diversion oxbow (ave ± std. error). Rates are in g- N m-2 yr-1 or g-P m-2 yr-1 as is appropriate. The mean loading, export, and the retention of the nutrients during the eight flood pulses in the mitigation oxbow wetland in 2004 are also shown. Rates and the mass retention are calculated according to the number of actual days of flow; note that the duration of the inflow and outflow are different…………….…………….. …...51

2.5 Annual and per pulse removal of nitrate (as N) and total phosphorus in selected created river diversion and riparian wetlands.…………………… ……52

3.1 Inflow and outflow of the oxbow wetland in April 2004 – March 2006. Hydraulic retention time (HRT) was calculated only for days in which there was no inflow or outflow. The Richardson-Baker flashiness index (R-B index) for the created oxbow wetland was calculated on a monthly basis for pulsing and steady-flow hydroperiods. Wet season is defined as November through June; dry season is defined as July through October… ……95

xx - 3.2 Annual mean loading and retention of nitrate-nitrite (NO3 -N), total nitrogen (TN), soluble reactive phosphorus (SRP), and total phosphorus (TP) during pulsing hydroperiod (2004) and steady-flow (2005) years in the river diversion oxbow (ave ± std. error). Rates are in g-N m-2 yr-1 or g- P m-2 yr-1 as is appropriate. Rates and the mass retention during the pulsing year are calculated according to the number of actual days of flow; note that the duration of the inflow and outflow are different.……..……. …….96

3.3 Nutrient and suspended sediment concentrations (ave ± std error (# samples) in the created oxbow wetland at the Olentangy River Wetland Research Park during pulsing (April 2003 – March 2005) and non-pulsing (April 2005 – March 2006) conditions when flooded river water is flowing through the wetland (SRP = soluble reactive phosphorus, TP = total - - phosphorus, NO3 + NO2 = nitrate and nitrite, TN = total nitrogen, and TSS = total suspended solids)……...……………………………………. ……..97

3.4 Bird species observed utilizing the created oxbow wetland with a pulsing hydroperiod, April 2004 – March 2005, and with a steady-flow hydroperiod, April 2005 – March 2006………...………………………….. …...98

3.5 Annual and per pulse removal of nitrate (as N) and total phosphorus in selected created river wetlands.………………………………...………… ...…..99

4.1 Inflow, outflow, mean hydraulic retention time (HRT; ave ± std error), and water depth (ave ± std dev) of the mitigation oxbow wetland in April 2004 – Sept 2006. Units of inflow and outflow are cubic meters of water flux per square meter of wetland area (2.8 ha). HRT was only calculated for days in which there was inflow or outflow. The Richardson-Baker (R-B index) flashiness index for the created oxbow wetland was calculated on a monthly basis for the pulsing and steady-flow hydrologic conditions. The Wet Season is defined as November through June; the Dry Season is defined as July through October. *The second pulsing year (2006) only encompasses a partial year (170 days, April 2006 to September 2006)..….. ….137

4.2 Percent areal coverage and the Community Diversity Index (CDI) of the 5 dominant macrophyte communities in the created oxbow wetland in 2004 (pulsing hydrology), 2005 (steady flow hydrology), and 2006 (restored pulsing hydrology). The “Mixed macrophyte” community is comprised predominately of: Scirpus americanus, Eleocharis sp., Juncus effusus, Leersia oryzoides, Sagittaria spp., Verbesnia alterniflora, and Cyperus strigosis. NAPP is calculated for the area where macrophytes actually grew and not for the area of the entire oxbow wetland. CDI = Σ (Ciln(Ci)). Where Ci is the present cover of community “i” and N is the number of macrophyte communities (Mitsch et al., 2005a).…………………………. ..…138

xxi 4.3 Percent of the total biomass comprised by each of the 5 dominant macrophyte communities in the created oxbow wetland in 2004 (pulsing hydrology), 2005 (steady flow hydrology), and 2006 (restored pulsing hydrology). The “mixed. macrophyte” community is comprised predominately of: Eleocharis sp., Scirpus americanus, Juncus effusus, Leersia oryzoides, Sagittaria spp., Verbesnia alterniflora, and Cyperus strigosis....…………………………………………………………………..….139

4.4 Net aboveground primary productivity per unit area of vegetation (NOT total production) for the 5 dominant macrophyte communities in the created oxbow wetland in 2004 (pulsing hydroperiod), 2005 (steady-flow hydroperiod), and 2006 (restored pulsing hydroperiod). Superscripts denote significance between years for each macrophyte community (p < 0.05). The “mixed macrophyte” community is comprised predominately of: Eleocharis sp., Scirpus americanus, Juncus effusus, Leersia oryzoides, Sagittaria spp., Verbesnia alterniflora, and Cyperus strigosis..…………… …140

5.1 Four potential ecosystem successional stages predicted by the simulation model.………………………………………………………………………. …198

5.2 Model parameters, definitions, values, and sources for the created oxbow wetland hydrologic submodel and sediment accretion sub-routine………... …199

5.3 Model parameters, definitions, values, and sources for the created oxbow wetland tree submodel.……………………………………………………. …..200

5.4 Tree species growth coefficients that can be used in the tree growth model. Coefficients were calculated according to Botkin (1993) unless otherwise noted. GT, β0, β1, b2, b3 and DegD for a given species were calculated using age, height, and diameter parameters, which were taken from Petrides (1988), Sargent (1933), and Jenkins et al. (2003). Values of TOD were taken from Pearlstine et al. (1985) or calculated using from Pretides (1988) and Sargent (1933) as marked. The species used in the model presented here is Populus deltoides.……………………………...………... ….201

5.5 Model parameters, definitions, values, and sources for the created oxbow wetland macrophyte submodel…………………………………………….. ….202

5.6 Growth characteristics and parameters for functional emergent macrophyte groups used in the macrophyte growth sub-model (from Ellison and Bedford, 1993; model calibration). The functional group used in the model presented here is a combination of groups 1 and 3 from Boutin and Keddy (1993).………………………………………………..… ....203

xxii 5.7 Predicted basal area and net above-ground primary productivity (NAPP) per square meter of each elevation zone at the end of 8 (present year) and 100 years given a 100 year extrapolation of the past 10 years of Olentangy River flow. “Total production” refers to the entire wetland basin (multiplying the results from the unit model by the area of the wetland within the given contour band)…………………………………………….. ….204

5.8 Comparison of the state variable sensitivities (Sx) to different model parameters…..……………………………………………………………..…...205

xxiii

CHAPTER 1

INTRODUCTION

The most productive ecosystems are generally those that receive pulses of energy from external sources (e.g., tides, river floods, pulses of runoff, upwelling) in addition to solar energy (Nixon, 1988; W.E. Odum et al., 1995, Odum, 2000). At the same time, they make the system more “open” to exchange of elements with adjacent systems, which not only may enhance productivity but may also contribute to the stability of the ecosystem. This concept of stability of ecosystems in the face of continued pulses was termed pulse-stability by E.P. Odum (1983, 1995). A flood-pulse hydroperiod (the periodic flooding from an adjacent river or lake) that was once common to riverine wetlands in the Midwest United States (Baker et al., 2004), has been shown to increase nutrient uptake and plant productivity in riparian wetlands around the world (Mitsch and

Ewel 1979, Junk et al., 1989; Tockner et al., 2000). There may be the situation in Ohio as well because many of the river waters have high nutrient loads that can have a

“fertilizer effect” on wetland vegetation (Spink et al., 1998; Mitsch and Gosselink, 2000;

Gusewell et al., 2003; Gathumbi et al., 2005). A flood-pulse hydroperiod , however, can be a stress as well as a subsidy. When inundation is prolonged, flooding may impede plant growth and nutrient uptake (van der Valk and Davis, 1978; Mitsch and Rust, 1984;

Kozlowski, 2002).

1.1 Rationale and Significance

As river diversions become an increasingly popular design method for the

ecological engineering of watersheds and deltas (e.g., Reyes et al., 2000), it is important

to evaluate these ecosystems in enough detail to determine how well they function during

infrequent but important hydrologic events such as spring floods, rainstorms, and ice

melts. At these times the nutrient retaining functions, and storage capacity of a wetland

can be greatly diminished (Novitzki 1982; Raisen and Mitchel, 1995; Raisen et al., 1997).

Some studies have showed that certain designs of created wetlands can compromise

retention ability (Wong and Somes, 1995; Fink and Mitsch, 2004). In addition to the

water quality improvement functions of river diversion wetlands, increasing the

connectivity between rivers and floodplains can have a marked impact upon floral and

faunal communities (Brown, 1981; Waters and Shay, 1992; Toner and Keddy, 1997;

Newman et al., 1998; Giovanni and Da Motta Marques, 1998; Tanner et al., 1999;

Casanova and Brock, 2000; Henry et al., 2002; Kellogg et al., 2003) and directly affect chemical and physical processes governing nutrient and suspended solid dynamics within wetlands (Knight et al., 1987; Kadlec, 1994; Boustany et al., 1997; Lane et al., 1999,

2003; Blahnik and Day, 2000; Mitsch and Gosselink, 2000; Spieles and Mitsch, 2000).

1.2 Creation, Restoration, and Mitigation

When we lose wetlands, we lose their ability to provide valuable ecosystem services such as water quality improvement, sediment retention, flood prevention, and habitat for a variety of animals and plants (Mitsch and Gosselink, 2000). The wetland

2 mitigation policy is supposed prevent the loss of these ecosytesm services by requiring that lost wetlands be replaced by newly created or restored wetlands (US EPA, 1990;

Reinartz and Warne, 1993; Mitsch and Gosselink, 2000, Cole and Schafer, 2002).

However, many mitigation, restoration, and creation projects have not been very successful at replacing the functions lost due to the destruction of wetlands for development and agriculture(e.g. Maguire, 1985; Kentula et al., 1992; Mitsch and

Wilson, 1996; Zedler and Callaway, 2000; Cole and Schafer, 2002). One of the reasons for this rate of failure is that there is an inadequate understanding regarding how various ecological, biogeochemical, and hydrological functions of a wetland interact (Cole and

Schafer, 2002). For example, even though such indicators as macrophyte cover, species composition, and hydraulic loading in a wetland are frequently used to determine the legal and ecological success of these wetlands (Reinartz and Warne, 1993; Mitsch et al.,

1998; NRC, 2001; Cole and Schafer, 2002), it is unclear precisely how these features interact within the ecosystem as a whole (Mitsch and Wilson, 1996).

Research is needed to help wetland designers, builder s, and managers create mitigated and restored wetlands in a manner that replaces lost ecosystem functions and not just lost area (Cole and Schafer, 2002). This is important both ecologically and legally as the United State Environmental Protection Agency and the United State Army

Corps of Engineers have stated that mitigation must focus on functional replacement and not simply on area or structure (US EPA, 1990). Mitigation, created and restored wetlands will certainly not perform the intended functions if the landscape position is incorrect and the proper physical structure and characteristic hydrology are lacking (Cole and Schafer, 2002).

3 1.3 Ecosystem Modelling

One way to improve our understanding of how the hydroperiod and plant primary production and succession interact is to develop a simulation model to describe these processes. At this time there are not many, if any, such models that can be readily used by wetland designers to help them create wetlands. For a model to be a useful tool for wetland designers, builders, and managers, it is beneficial for it to be a relatively simple model (so that the cost, time, and effort to acquire the necessary input data is prohibitively high) yet still effectively predict both current ecosystem function and the functional development of a wetland.

Wetland designers may benefit from having simulation models that include both function and form in the same model. The model developed in this study will describe varying developmental scenarios in a created wetland, with a focus on a particular riparian oxbow design. This type of model will potentially suggest variables that can be manipulated to affect ecosystem development. In particular, the depth of water, the accretion of sediment, the role of macrophytes (type, productivity, and cover), and the basin morphology were key features in the modeling effort.

1.4 A whole-ecosystem approach

While there may be no optimal scale for ecosystem experimentation, studies that attempt to link wetland function with structure are often done at inappropriate spatial and temporal scales (Mitsch and Day, 2004). It is not always sound to apply conclusions from small-scale, short-term studies to real-world, full-scale conditions (Ahn and Mitsch,

2002; Mitsch et al., 2005a). Whole ecosystem studies are sometimes criticized because

4 the size, cost, and logistics do not allow for much, if any, replication. These weaknesses, however, are compensated for by the results of whole-ecosystem studies having a lower variability and by being more likely to reveal ecosystem properties that are not apparent in smaller scale experiments (Odum, 1990; Carpenter, 1998; Kemp et al., 2001). For example, even though biotic community structure is often taken as a surrogate for ecosystem function (Kentula et al., 1992), the presence of physical or biological structure may not indicate functional replacement (Campbell et al., 2002). Until our understanding of ecological indicators improves, it is necessary to do in depth whole ecosystem studies to determine what the true functionality of created ecosystems is. Such studies are not often done (Rienartz and Warne, 1993) and there is still much need for improvement in the building of wetlands (Mitsch and Wilson, 1996).

1.5 Goals and Objectives

The overall goal for this research was to understand the role of hydrologic pulsing in the functioning of a created riparian river diversion wetland in the Midwestern USA.

Heterogeneity in wetland morphology, hydroperiod, and maturity on wetland biogeochemistry, primary productivity, and successional development were evaluated.

The specific objectives of this dissertation were to:

1. determine the overall efficacy of the wetland at attenuating influent nutrient loads

(Chapter 2);investigate the importance of seasonal hydrologic pulses vs. steady

5 flow conditions on the wetland’s ability to attenuate nitrogen, phosphorus, and

suspended solids and on the use of created riparian wetland by avian species

(Chapter 3);

3. determine the effect of removing and subsequently restoring seasonal hydrologic

pulses on the primary productivity and floral community diversity in the wetland

(Chapter 4); and

4. develop a simulation model that could predict the long-term vegetation patterns in

this created riparian wetland (Chapter 5).

This project involved manipulating pulse and non-pulse fluxes of nutrient-laden waters into adjacent riparian wetlands on a fourth-order river. Nutrient retention was expected to be affected by the properties of the soil water interface, retention time, flood pulse frequency and duration, loading rates, and macrophytes. The macrophytes were also expected to change in response to changes in the wetland hydroperiod. The macrophytes may also prove to have a significant role in nutrient cycling by affecting residence time and by direct nutrient uptake and decay/deposition.

This research was carried out at a created river diversion wetland along the

Olentangy River at the Wilma H. Schiermeier Olentangy River Wetland Research Park at

The Ohio State University, Columbus, OH, USA (Mitsch et al., 1998b; 2005b). The first study (Chapter 2) focused specifically on water quality functions and the development of herbaceous plant communities in this riparian wetland. I hypothesized that the different zones of the wetland would support different biogeochemical functions. I also

6 hypothesized that the greatest reduction of nitrate-nitrogen would occur in the emergent marsh portion of the wetland and the greatest reduction of total suspended solids and total phosphorus would occur in the open water portion of the wetland. All nutrient and sediment dynamics were expected to vary with seasons. This study was intended to provide a baseline for subsequent studies on the effects of variations in hydroperiod on wetland ecosystem function.

A second study (Chapter 3) evaluated the effect of replacing the natural pulsing hydroperiod with a steady inflow of river water provided artificially by submerged pumps. This study evaluated the hypotheses that the wetland would have greater nutrient removal during pulsing conditions and that the wetland would support a greater number of avian species during pulsing conditions.

The third study (Chapter 4) examined the effect of removing and subsequently restoring a pulsing hydroperiod on macrophyte productivity and diversity. It was hypothesized that wetland productivity, areal coverage, and community diversity would decrease when pulsing was removed and increase again when pulsing was restored.

A fourth study (Chapter 5) developed a unit model that was designed to take short-term hydrologic data and make long-term predictions about the development of the created wetland. The primary goal for the model is to accurately predict the successional maturation of this wetland design so that wetland designers and builders will be able to know with relative certainty what functions a similarly designed wetland will support if built in a different location. Simulations investigated the effects of variations in water delivery (both natural and artificial), sediment accretion, and beaver activity.

7 1.6 Literature Cited

Ahn, C. and W.J. Mitsch. 2002. Scaling considerations of mesocosm wetlands in simulating large created freshwater marshes. Ecological Engineering 18:327-342.

Baker, D.B., R..P. Richards, T.T. Loftus, J.W. Kramer. 2004. A new flashiness: Characteristics and applications to Midwestern rivers and streams. Journal of the American Water Resources Association 40:503-522.

Blahnik, T. and J.W. Day. 2000. The effects of varied hydraulic and nutrient loading rates on water quality and hydrologic distributions in a natural forested treatment wetland. Wetlands 20:48-61.

Boustany, R.G., C.R. Crozier, J.M. Rybczyk, R.R. Twilley. 1997. Denitrification in a south Louisiana wetland forest receiving treated sewage effluent. Wetlands Ecology and Management 4:273-283.

Brown, S.L. 1981. A comparison of the structure, primary productivity, and transpiration of cypress ecosystems in Florida. Ecological Monographs 51:405-415.

Campbell, D.A., C.A. Cole, R.P. Brooks. 2002. A comparison of created and natural wetlands in Pennsylvania, USA. Wetlands Ecology and Management 10:41-49.

Carpenter, S.R. 1998. The need for large-scale experiments to assess and predict the response of ecosystems to perturbation. pp. 287-312. In: Pace, M.L. and P.M. Groffman (eds.), Successes, Limitations, and Frontiers of Ecosystem Science. Springer-Verlag, NY.

Casanova, M.T and M.A. Brock. 2000. How do depth, duration, and frequency of flooding influence the establishment of wetland plant communities. Plant Ecology 147:237-250.

Cole, C.A. and D. Schafer. 2002. Section 404 wetland mitigation and permit success criteria in Pennsylvania, USA, 1986-1999. Environmental Management 30:508-515.

Fink, D.F. and W.J. Mitsch. 2004. Seasonal and storm event nutrient removal by a created wetland in an agricultural watershed. Ecological Engineering 23:313-325.

Gathumbi, S.M., P.J. Bohlen, D.A. Graetz. 2005. Nutrient enrichment of wetland vegetation and sediments in subtropical pastures. Soil Science Society of America Journal 69:539-548.

Giovanni, S.GT. and D.M.L. Da Motta Marques. 1998. Establishment of three emergent macrophytes under different water regimes. Water Science and Technology 40:233- 240.

8 Gusewell, S., U. Bollens, P. Ryser, F. Klotzli. 2003. Contrasting effects of nitrogen, phosphorus and water regime on first- and second-year growth of 16 wetland plant species. Functional Ecology 17:754-765.

Henry, C.P., C. Amoros, N. Roset. 2002. Restoration ecology of riverine wetlands: A 5- year post-operation survey on the Rhône River, France. Ecological Engineering 18: 543-554.

Junk, W. J. 1999. The flood pulse concept of large rivers: Learning from the tropics. Archiv für Hydrobiologie 115: 261–280.

Kadlec, R. H. 1994. Detention and mixing in free water wetlands. Ecological Engineering 3:1–36.

Kellogg, C.H., S.D Bridgham, S.A. Leight. 2003. Effects of water level , shade, and time on germination and growth of freshwater marsh plants along a simulated successional gradient. Journal of Ecology 91:274-282.

Kemp, W.M., J.E. Peterson, R.H. Gardner. 2001. Scale-dependence and the problem of extrapolation: Implications for experimental and natural coastal ecosystems. pp. 3-57 In: Gardner, R.H., W.M. Kemp, V.S. Kennedy, and J. Peterson (eds), Scaling Relationships in Experimental Ecology. Columbia University Press, NY.

Kentula, M.E., Brooks, R.B., Gwin, S.E., Holland, C.C., Sherman, A.D. Sifneos, J.C. 1992. An Approach To Improving Decision Making in Wetland Restoration and Creation Island Press, Washington, D.C.

Knight, R. L., T. W. McKim, H. R. Kohl. 1987. Performance of a natural wetland treatment system for wastewater management. Journal of Water Pollution Control Federation 59:746–754.

Kozlowski, T.T. 2002. Physiological-ecological impacts of flooding on riparian forest ecosystems. Wetlands 22:550-561.

Lane, R.R., J.W. Day, B. Thibodeaux. 1999. Water quality analysis of a freshwater diversion at Caernarvon, Louisiana. Estuaries 22:327–336.

Maguire, C.E. 1985. Wetland replacement evaluation. Contract No. DACW-65-85-D- 0068. U.S. Army Corps of Engineers, Norfolk District, Virginia.

Mitsch, W.J. and K.C. Ewel. 1979. Comparative biomass and growth of cypress in Florida wetlands. American Midland Naturalist 101:417-426.

Mitsch, W.J. and W.G. Rust. 1984. Tree growth responses to flooding in a bottomland forest in northeastern Illinois. Forest Science 30:499-510.

9 Mitsch, W.J. and R.F. Wilson. 1996. Improving the success of wetland creation and restoration with know-how, time, and self-design. Ecological Applications 6:77-83.

Mitsch, W.J., S. Johnson, M. Liptak. 1998. Planting and planting success of the new mitigation wetland at the Olentangy River Wetland Research Park in 1997. pp. 205- 210. In: Mitsch, W.J. and V. Bouchard (eds). Olentangy River Wetland Research Park Annual Report 1997, Ohio State University, Columbus, OH.

Mitsch, W.J and J.G. Gosselink. 2000. Wetlands, 3rd ed. John Wiley and Sons, NY.

Mitsch, W.J. and J.W. Day, Jr. 2004. Thinking big with whole ecosystem studies and ecosystem restoration—A legacy of H.T. Odum. Ecological Modelling 178:133-155.

Mitsch, W.J., J.W. Day, L. Zhang, R.R. Lane. 2005a. Nitrate-nitrogen retention in wetlands in the Mississippi River Basin. Ecological Engineering 24:267-278.

Mitsch, W.J., L. Zhang, C.J. Anderson, A. Altor, M. Hernandez. 2005b. Creating riverine wetlands: Ecological succession, nutrient retention, and pulsing effects. Ecological Engineering 25:510-527.

National Research Council. 1992. Restoration of Aquatic Ecosystems. National Academy Press, Washington, DC. 552 pp.

Newman, S., J. Schuette, J.B. Grace, K. Ritchey, T. Fontaine, K.R. Reddy, M. Pietrucha. 1998. Factors influencing cattail abundance in the northern everglades. Aquatic Botany 60:265-280.

Nixon, S.W., 1988. Physical energy inputs and the comparative ecology of lake and marine ecosystems. Limnology and Oceanography 33:1005–1025.

Novitzki, R.P. 1982. Hydrology of Wisconsin Wetlands: Wisconsin Geological and Natural History Survey Information Circular 40, Madison, WI, 22 p.

Odum, H.T. 1983. Systems Ecology – An Introduction. John Wiley and Sons, New York. Pp. 644.

Odum, E.P., 1990. Field experimental tests of ecosystem-level hypotheses. Trends in Ecological Evolution 5:204–205.

Odum W.E., E.P. Odum. H.T. Odum. 1995. Nature’s pulsing paradigm. Estuaries 18:547–555.

Odum, E. P. 2000. Tidal marshes as outwelling/pulsing systems. In M. P. Weinstein and D. A. Kreeger, eds. International Symposium: Concepts and Controversies in Tidal Marsh Ecology. Kluwer Academic Publishers, Dordrecht.

Raisen, G.W. and D.S. Mitchel. 1995. The use of wetlands for the control of non-point source pollution. Water Science and Technology 32:177-186.

Raisen, G.W., D.S. Mitchel, R.L. Croome. 1997. The effectiveness of a small constructed wetland in meliorating diffuse nutrient loadings from an Australian rural catchment. Ecological Engineering 9:19-36.

Reinartz, J.A. and E.L. Warne. 1993. Development of vegetation in small created wetlands in southeastern Wisconsin. Wetlands 13:153-164.

Reyes E, M.L. White, J.F. Martin, G.P. Kemp, J.W. Day, A. Aravamuthan. 2000. Landscape modeling of coastal habitat change in the Mississippi Delta. Ecology 81, 2331–2349.

Spieles, D.J. and W.J. Mitsch. 2000. The effects of season and hydrologic and chemical loading on nitrate retention in constructed wetlands: a comparison of low- and high- nutrient riverine systems. Ecological Engineering 14:77-91.

Spink, A., R.E. Sparks, M. van Oorschot, T.W. Verhoenven. 1998. Nutrient dynamics of large river floodplains. Regulated rivers: Research and Management 14:203-216.

Tanner, C.C., J. D’Eugenio, G.B. McBride, J.P.S. Sukias, K. Thompson. 1999. Effect of water level fluctuation on nitrogen removal from constructed wetland mesocosms. Ecological Engineering 12:67-92.

Tockner , K., F. Malard, J.V. Ward. 2000. An extension of the flood pulse concept. Hydrologic Processes 14:2861-2883.

Toner, M. and P. Keddy. 1997. River hydrology and riparian wetlands: A predictive model for ecological assembly. Ecological Applications 7:236-246.

U.S. Environmental Protection Agency. 1983. Methods for Chemical Analysis of Water and Wastes. 600/4-79-020, U.S. Environmental Protection Agency, Cincinnati, OH.

van der Valk, A.G. and C.B. Davis. 1978. The role of seed banks in the vegetation dynamics of prairie glacial marshes. Ecology. 59:322-335.

Waters, I. and J.M Shay. 1992. Effect of water depth on population parameters of Typha glauca stand. Canadian Journal of Botany 70:349-351.

Wong, T.H.F., and N.L.G. Somes. 1995. A stochastic approach to designing wetlands for stormwater pollution control. Water Science and Technology 32:145-151.

Zedler, J.B. and J.C. Callaway. 2000. Evaluating the progress of engineered tidal wetlands. Ecological Engineering 15:211-225.

CHAPTER 2

NUTRIENT BIOGEOCHEMISTRYAND MACROPHYTE ASEMBLAGES IN A

CREATED WETLAND

2.1 Abstract

A better understanding of the function of riparian wetlands is needed. In this study, hydrological, successional, and water-quality dynamics are documented for a whole-ecosystem study involving a 3-ha created riparian wetland at the Schiermeier

Olentangy River Wetland Research Park at The Ohio State University in Columbus, Ohio

USA during 2003 and 2004. This wetland typically receives seven or eight natural weeklong flood pulses each year from the Olentangy River. Of 21 species planted in

1997, only Scirpus americanus and Juncus effusus remained as important macrophyte species during the study seven years after planting. The dominant macrophyte in most of the wetland was Typha spp. (angustifolia and latifolia). Mean removal per flood pulse for nitrate-nitrite, total Kjeldahl nitrogen (TKN), soluble reactive phosphorus (SRP), and total phosphorus (TP) was 0.71 g-N m-2, 0.92 g-N m-2, 0.016 g-P m-2, and 0.08 g-P m-2

- respectively. The annual reductions of N-NO3 , TN, P-SRP, and TP were 74%, 41%,

- 46%, and 31% by mass. A greater attenuation of NO3 and TP occurred in the emergent marsh section of the wetland than the open water section. TKN concentration increased

12 through the emergent marsh and subsequently decreased through the open water section.

The created oxbow successfully removed nitrate and phosphorus and also had an acceptable assemblage and coverage of macrophytes. It is recommended that similar diversion wetlands be created in other locations to examine their function under different climatic and hydrological conditions.

Keywords: Created wetlands; Ohio River Basin; Olentangy River Wetland Research

Park; pulsing; river diversion; Typha spp.; wetland function; wetland restoration.

2.2 Introduction

The National Research Council (1992) called for the creation and restoration of 4 million ha of wetlands in the United States by 2010. Mitsch et al. (2001, 2005a) and

Mitsch and Day (2006) have suggested that the restoration of 2 million ha of wetlands is necessary in the Mississippi River basin alone to mitigate the hypoxic zone in downstream coastal waters of the Gulf of Mexico. One of the ideal locations for many of these created and restored wetlands is near rivers where river waters can be diverted mechanically or naturally into a wetland. Riparian wetlands are connected to systems that routinely provide river water and propagules. In turn wetlands provide habitat that serves as a nursery for fish and other aquatic life, improvement of water quality, and flood storage capacity.

River diversion wetlands are wetlands fed primarily by flooding streams, which bring sediments and chemicals into the wetlands along with seasonal floodwaters and for the water to then flow back into the stream (Mitsch and Day, 2006). Because there are

13 both artificial and natural levees along major sections of streams, it is often possible to

create such wetlands with minimal construction work, by removing portions of levees to

allow floodwater to enter the wetlands. Such wetlands mimic natural oxbows that capture

floodwater and sediments and slowly release the water back to the river after the flood

passes (Mitsch and Day, 2006).

As river diversions become an increasingly popular method for the ecological

engineering of wetlands in downstream portions of watersheds (e.g., Reyes et al., 2000), it is important to evaluate if this wetland type is also appropriate in the upper portion of watersheds in large catchments. Furthermore, if management nutrient pollution throughout the entire Mississippi watershed is to be attempted, we must have data that detail how specific wetland designs function (Reyes et al., 2000). A few studies have

compared natural and created wetlands in upper and lower portions of large watersheds,

but more study is needed regarding the function of specific river diversion designs in

various climates (e.g., Mitsch et al. 2005a).

There has been a great deal of research on how riparian wetlands function (Mitsch

and Gosselink, 2000; Kozlowski, 2002; Kao et al., 2003) and on the relative success of specific wetland restoration and creation projects (Mitsch and Wilson, 1996; Zedler and

Callaway, 2000; Cole and Schafer, 2002). However, insufficient research has been conducted with regards to the specific design features that lead to optimal functioning of created and restored river diversion wetlands. Also, while many studies have determined the overall efficacy of wetlands, few have analyzed these ecosystems in enough detail to determine if the various zones of the wetland are functioning similarly. In addition, most

14 studies do not explicitly take into account rare but important hydrologic events such as

spring floods, rainstorms, and ice melts. At these times the nutrient retaining functions,

and storage capacity of a wetland can be greatly diminished (Novitzki 1982; Raisen and

Mitchel, 1995; Raisen et al., 1997). Certain designs of created wetlands can compromise

retention ability (Wong and Somes, 1995; Fink and Mitsch, 2004).

The importance of river flood events is increasingly recognized in the field of

restoration ecology and attempts have been made to reconnect rivers with their natural

floodplains (Day et al., 1995; Galat et al., 1998; Hensel et al., 1998; Molles et al., 1998;

Toth et al., 1998; Henry et al., 2002; Mitsch and Day, 2006). In addition to the water

quality merits of river diversion wetlands, increasing the connectivity between rivers and

floodplains can have a marked impact upon floral and faunal communities. One potential

benefit of river diversion wetlands with fluctuating water levels is to create/restore a

mixed habitat of wooded and herbaceous wetland (Toner and Keddy, 1997).

A wetland’s hydroperiod directly affects the nutrient and suspended solid

dynamics within wetlands (Mitsch and Gosselink, 2000). The rate at which a wetland’s

water quality changes is generally acknowledged to be dependant on nutrient

concentrations in the inflow, the chemical form of the nutrient, and water flux (Knight et

al., 1987). For example, maximum efficiency of nitrogen removal occurs at loading rates

below 10 g-N m-2 yr-1 (Lane et al., 2003; Spieles and Mitsch, 2000), when diverted water

is spread out over the largest possible wetland area (Lane et al., 1999; Blahnik and Day,

- 2000), and when the NH3:NO3 ratio is less than 1.0 (Boustany et al., 1997).

Furthermore, because water is unevenly distributed in flow-through wetlands due to

15 different degrees of channelization, microtopography, animal activity, and patterns of vegetative growth, different parts of a diversion wetland will function differently

(Fennessy et al., 1994; Kadlec, 1994).

The objective of this project was to observe the function of a created river diversion wetland in the upper Ohio River basin. This study focused specifically on water quality functions and the development of herbaceous plant communities in this riparian wetland. I hypothesize that the different zones of the wetland will have different biogeochemical functions. Specifically that the greatest attenuation of nitrogen species will occur in the emergent marsh portion of the wetland and that attenuation of total suspended solids and total phosphorus in the open water portion of the wetland will vary with seasons.

2.3 Methods

2.3.1 Site Description

The 3-ha created riparian wetland (referred to here as a created oxbow) examined in this study is located on the floodplain of the Olentangy River in Central Ohio at the

Schiermeier Olentangy River Wetland Research Park at The Ohio State University

(Columbus, Ohio, USA; Fig. 2.1). Water enters the oxbow through a Red Field

TideflexTM check valve when the river elevation is higher than the wetland, and flows back to the Olentangy River though an outflow control weir by gravity. The wetland has two significant vegetation zones. The northern half (closest to the inflow) is an emergent marsh, and the southern half (closest to the outflow) is an open water basin. Lack of

16 vegetation in the southern half is likely due to high water conditions during spring, which

may prevent germination of emergent aquatic plants outside of the littoral zone.

2.3.2 Hydroperiod

The wetland’s hydroperiod was determined by measuring the water level of the

created oxbow and the river with staff gauges, and by measuring the flow of water into

and out of the wetland using a Swofer 2100 current meter and an ISCO 730 bubbler

module. A simple mathematical model was developed to describe the inflow based on

the elevation of the river relative to the oxbow water surface and on the shape of the inflow pipe and weir. Inflow into the wetland equaled:

Qin = (2.09)*PipeA*√ (Rstage-MSLW)*43560 (2.1)

Where Qin is the inflow, PipeA is the inundated area of the inflow pipe, Rstage is the

surface water elevation of the river, and MSLW is the surface water elevation of the

wetland. Outflow from the wetland equaled:

1.436 Qout = 10.16*(( MSLW -WL) )*43560 (2.2)

Where Qout is the outflow, and WL is the elevation of the top of the outflow weir (See

Chapter 5 for more details). Daily hydrologic budgets enabled calculation of loading and retention rates (by mass) of the various nutrients.

2.3.3 Water Quality

Dawn and dusk inflow and outflow grab samples were taken between October

2002 and October 2004 during these flood events. When the Olentangy River is at a high stage (above 220.9 m MSL), overflow occurs into the diversion wetland from the river. More detailed automatic sampling was conducted during some storm events using

ISCO 6874 autosamplers at the inflow and outflow points of the created oxbow.

Samplers were set to hourly frequencies during high flow periods. This more detailed sampling strategy allowed a comparison of the wetland’s functioning during high and low flow periods. Grab samples (n = 514 over the two water years) were analyzed for nitrate- nitrite, total Kjeldahl nitrogen, soluble reactive phosphorus, total phosphorus, and turbidity.

Manual and auto-sampler grab-samples were split into filtered (0.45 µm) and unfiltered subsamples, and analyzed for nitrogen and phosphorus concentrations.

Nitrate+nitrite, soluble reactive phosphorous (SRP), total Kjeldahl nitrogen (TKN), and total phosphorus (TP) concentrations were measured using Lachet (2000) modifications of standard methods (USEPA 1983, APHA 1998) with a Lachet QuickChem flow- injection analyzer (FIA+ 8000). TP and SRP were analyzed with an ascorbic acid and molybdate color reagent method. TP samples were digested in a block digester using 0.5 ml of a digestion solution made up of H2SO4, K2SO4 and mercuric sulfate and heated for

1 hour at 160oC and 1.5 hours at 380oC. TKN was determined using the salicylate and hypochlorite method. Nitrate+nitrite (hearafter referred to as nitrate or nitrate-N) was

18 determined by the sulfanilamide method after reduction in a cadmium column (Lachet,

2000). Total nitrogen (TN) was estimated to be the sum of TKN and nitrate+nitrite-N.

Total suspended solids were determined using an empirical correlation with turbidity developed for our study site (Harter and Mitsch, 2003). Turbidity was measured in the laboratory using a Hach 2100 N turbidimeter.

Quality control in the laboratory was maintained through the use of matrix duplicates, check standards, matrix spikes, and blanks which were all utilized each time the Lachet was run to verify data precision and accuracy. Matrix duplicates are environmental samples that are divided into two samples in the laboratory and analyzed separately and compared for precision; approximately 5% of the samples were checked with a matrix duplicate. Check standards and blanks are samples of known concentration that are tested and compared to the known value to test for accuracy. If a correlation analysis of the results indicated the test samples were within ± 5% accuracy, then the results were considered acceptable. A matrix spike is an environmental sample to which a known concentration of analyte has been added. Matrix spikes were used to ensure accuracy with 2% of the samples. The spike was taken through the entire analytical procedure and the recovery of the analyte was calculated. Results were expressed as percent recovery of the known amount spiked. The matrix spike was considered acceptable when >95% recovery was achieved.

2.3.4 Statistical Methods

Microsoft Excel and SPSS 11.0 were used to calculate the mean, standard deviation, and correlations between nutrient and other environmental values. When

19 comparing multiple variables, an analysis of variance (ANOVA) was done using

Bonferroni’s method for multiple comparisons. Direct comparisons of single variables

between the first and second water years of the study were done using a homoscedastic,

two-tailed Student’s T-Test for equal means. All tests were conducted at a 95%

confidence interval (α = 0.05).

Interpolated maps of nutrient concentration isobars within the wetland basin were

created with point (punctual) kriging methods using a geostatistical software package

(GS+ Geostatistics for the Environmental Sciences v. 7.0). Prior to interpolation, semi-

variance analysis was done to produce Gaussian (hyperbolic) variogram models of the

autocorrelation present in the data. The active lag distance was set to 40 m (50 % of the

maximum transect length). The lag class distance interval (step size) was selected to

provide the best autocorrelation fit for each individual data set and varied between 2 and

6 m. Anisotropic methods were used as the variation in nutrient concentration was

expected to vary differently between the primary axis of water flow (north-south) and the

axis of lateral water flow (east-west). To account for this variation, an offset angle of 90o with an offset tolerance of 22.5o was used for all geostatistical analyses. Interpolation

values for a specific location were weighed by distance and the degree of autocorrelation

present for that distance. The interpolated isobar maps were created using a uniform x-y

grid that generated 188 points at a space interval of 0.6 m and a search neighborhood of

32 nearest points with no limit on geographical distance.

2.4 Results

2.4.1 Hydrologic loading and hydroperiod

20 From its creation in 1996 through 2004, the 3-ha created oxbow received an

average of seven to eight natural flood pulses each year from the Olentangy River (Table

2.1). Inflow from 1998-2004 averaged 20 ± 4 m yr-1 (or m3 m-2 yr-1; cubic meters of

inflow per square meter of wetland area per year). The Olentangy River typically

provides frequent, short (5-6 days of inflow) flood pulses into the created oxbow. These

pulses typically result in 9-12 days of outflow from the wetland.

During the wet season (November-June) and the dry season (July-October) there

were 165 and 7 days of inflow respectively (Fig. 2.2a). In 2003, the wet season was

atypically dry and the dry season was atypically wet, with 53 days of flow in the wet

season and 41 days of flow in the dry season (Fig. 2.2b). In 2003 and 2004 respectively,

the oxbow received 21 m yr–1 and 27 m yr–1 of water through 17 and 8 independent flood pulses per yr, respectively (Table 2.1). The year 2004 was wet compared to most of the previous years of the wetland’s existence (see previous data in Mitsch and Day, 2006).

Despite differences between the frequency and magnitude of the flood pulses (Fig. 2.2 a,b), the mean duration of flood pulse inflows and outflows were similar. In both years, inflow pulses were 5-6 days in duration and outflow pulses lasted 9-10 days (Table 2.1).

2.3.2 Vegetation

In May 1997, the created oxbow was planted with 6900 rootstocks representing

21 species (Acorus calamus, Alisma plantago-aquatica, Asclepias incarnata,

Cephalanthus occidentalis, Equisetum sp., Iris versicolor, Juncus effusus, Lobelia

cardinalis, Polygonum spp., Pontederia cordata, Potomogeton pectinatus, Sagittaria

latifolia, Saururus cernuus, Schoenoplectus tabernaemontani, Scirpus americanus,

21 Scirpus cyperinus, Scirpus fluviatilis, Sparganium eurycarpum, Spartina pectinata, and

Zizania aquatica (Mitsch et al., 1998). Seventy-five percent of these species are still found growing in the wetland. However, only two of the original planted species,

Scirpus americanus and Juncus effusus, make up a significant portion of the 2003-04 wetland primary productivity (Table 2.2).

The macrophyte species contributing most to productivity are Typha sp.,

Eleocharis sp. and Scirpus americanus (Table 2.2), which together account for 68% of the macrophyte net primary productivity in the wetland. The dominant vegetation communities in the created oxbow are Typha sp., a woody fringe of Salix spp. and

Populus deltoides, a mixed community of Eleocharis sp., Juncus effusus, and Scirpus americanus, and expanding patches of Pontederia cordata (Fig. 2.3). While Typha sp. contributes most to the macrophyte productivity, its proportion of the productivity decreases from 83% at the inflow to 0% at the outflow. Of 105 species of plants identified in the created oxbow in 2003 and 2004, 55 were wetland indicator species

(classified as FACW or OBL). Indicator status was determined using the National

Wetland Indicator List for Region I of USA (Northeast; Reed, 1998).

The changing water coverage in the southern basin of the created oxbow affected the macrophyte cover. As water receded in the late summer, Xanthium strumarium rapidly colonized the exposed mudflat, and became the dominant species on the mudflat until the onset of the first heavy frost.

2.4.3 Nutrients and sediments

Although the mean concentration of SRP in the water flowing into the created

oxbow in 2003 (60 ± 1 μg L-1) was almost double the mean concentration in 2004 (33 ± 1

μg L-1; Table 2.3), the mean reduction in SRP concentration was only 11.8% higher.

The pattern of SRP concentration change was also different between the 2003 and 2004.

In 2003 the SRP concentration decreased by the midpoint of the wetland to 13± 1 μg L-1 and then increased to 27± 1 μg L-1 at the outflow. In 2004, there was no decrease in SRP concentration between the inflow and the midpoint followed by a decrease in the second half of the wetland to 19 ± 1 μg L-1.

There was a significant difference in the inflow concentration of TP between

years (p < 0.01). In 2003, the mean inflow concentration as only 92 ± 6 μg L-1 compared

to 203 ± 2 μg L-1 in 2004 (Table 2.3). The percent removal of TP was the same in both years despite the difference in the initial inflow concentration. During early spring high flow periods, the wetland retained phosphorus, but it was a source of phosphorus during some late spring floods and during large thunderstorm events in the drier summer months

(Fig. 2.4). TP concentrations did not decrease informally through the oxbow in either year of the study (Fig. 2.5). TP concentrations increased significantly (p = 0.05) between the midpoint and the outflow of the wetland in 2003. In 2004, TP increased to 230 ± 15

μg L-1 along the northeast bank in the emergent marsh near the inflow of the wetland and

had a significant decrease from the midpoint (36 ± 8 μg L-1) to the outflow

(19 ± 1 μg L-1). There were also differences in the lateral distribution of TP between the

23 two years. In 2003 TP concentrations were fairly uniform east to west across the wetland, where-as in 2004 the east bank had a 20% higher TP concentration than the west bank of the wetland through the emergent marsh.

- - -1 The inflow concentration of NO3 -NO2 in 2003 (4.40 ± 0.04 mg L ) was more than twice as high as the concentration in 2004 (1.81 ± 0.01 mg L-1; Table 2.3). The decrease in nitrate concentration was 13.6% higher in 2004, even though there was a lower average inflow concentration that year. The spatial patterns of nitrate concentrations in the created oxbow were similar between the two years, with was no difference between the concentrations in the channel and along either shore (Fig. 2.5).

It is unclear if the elevated nitrate concentration seen along the west bank of the created oxbow near the outflow is an artifact of the sampling method or if nitrate was produced within the wetland. Because the sampling method was not continuous, it is possible that the grab sampling did not accurately capture each flood pulse as it traveled through the wetland. During most flood-pulses there was an increase in the nitrate concentration at the inflow as nutrient enriched river water flooded into the created oxbow (Fig. 2.6). As this water moved through a wetland, some amount of the nitrate could be taken up by plants or transformed by denitrifying bacteria in to N2 gas and emitted into the atmosphere (Mitsch and Gosselink, 2000). The nitrate that was not retained or removed flowed through the wetland as a nutrient pulse whose magnitude dropped as it moved through the wetland. If by chance grab sampling for nutrients occurred near the end of this pulse, it would be possible for the peak of the pulse to be near the outflow, making it appear as though nitrate was being produced by the wetland when in reality some amount was simply moving through untransformed.

24 Nitrate could also have been produced within the wetland. Soluble organic

nitrogen from decaying plants and algae can be mineralized under both aerobic and

+ anaerobic pathways through ammonification to form ammonium ions (NH4 ).

Ammonium ions can then be taken up directly by anaerobic organisms, converted to

ammonia (NH3) and lost to the atmosphere via volatilization, or it can be immobilized

through ion exchange with negatively charged soil particles. Ammonium ions can also

be oxidized in the thin oxidized layer that exists at the surface of many wetlands soils by

- Nitrosomonas sp. microorganisms into nitrate (NO2 ) and then by Nitrobacter sp.

- microorganisms into nitrate (NO3 ) in a process called nitrification (Mitsch and

Gosselink, 2000). Nitrification can also occur in the oxidized rhizosphere of plants

where adequate oxygen is often available to convert the ammonium nitrogen to nitrate

nitrogen (Reddy and Graetz, 1988)

The inflow concentration of TN in 2004 was 3.04 ± 0.05 mg L-1 (Table 2.3).

While TN decreased by 24% in 2004, its TKN fraction increased by 200% through the emergent marsh portion of the wetland (Fig 2.5), and then dropped across the open water portion to 0.60 ± 0.01 mg L-1 for an overall increase in TKN of 25.6% in 2004 (Table

2.3), the only year in which it was measured. Concentrations of TKN were higher along the east bank, especially near the end of the emergent marsh area (3.30 ± 0.15 mg L-1),

and lower along the west bank of the oxbow in the emergent zone and at the midpoint

(2.40 ± 0.20 mg L-1). In the open water area, there were no differences in TKN between

the edges and the channel (1.50 ± 0.05 mg L-1). For turbidity, the only lateral difference

was in the open water basin, where both edges were less turbid than the channel.

25 Total suspended solids at the wetland inflow were significantly greater (p = 0.5)

in 2004 (19.5 ± 0.9 NTU) than in 2003 (14.9 ± 1.1 NTU; Table 2.3). By the midpoint of the wetland, there was no difference between years (14.6 ± 2.9 NTU). There was,

however, a significant difference (p = 0.05) between years at the outflow. In 2003 the

total suspended solids dropped 38.9% to 9.1 ± 0.9 NTU. In 2004 there was no significant

drop from the midpoint to the outflow of the wetland (Table 2.3). The most obvious

difference in total suspended solids between the two years is the significant increase (p =

0.01) and subsequent decrease (103 ± 15 NTU) in the middle of the open water portion of

the wetland. Loading and retention rates of the various nutrients were variable throughout

2004 (Table 2.4). The total annual loading rate in 2004 for nitrate-nitrogen, total

nitrogen, soluble reactive phosphorus, and total phosphorus were 32.2 g-N m-2 yr-1, 64.5 g-N m-2 yr-1, 0.48 g-P m-2 yr-1, and 6.1 g-P m-2 yr-1 respectively. Retention rates were 15.4

g-N m-2 yr-1, 32.3 g-N m-2 yr-1, 0.05 g-P m-2 yr-1, and 4.48 g-P m-2 yr-1 respectively for the

same nutrients.

During the eight discrete flood pulses, mean loading for nitrate-nitrogen, total

nitrogen, soluble reactive phosphorus, and total phosphorus was 0.97 g-N m-2 per pulse,

2.54 g-N m-2 per pulse, 0.036 g-P m-2 per pulse, and 0.27 g-P m-2 per pulse. Removal

during the eight flood pulses was 0.71 g-N m-2 per pulse, 0.92 g-N m-2 per pulse, 0.016 g-

P m-2 per pulse, and 0.08 g-P m-2 per pulse respectively (Table 2.4). Overall, there was a

- 74%, 41%, 46%, and 31% reduction in the mass of N-NO3 , TN, P-SRP, and TP

respectively throughout the entire year during eight pulses. Sometimes the oxbow would

receive a flood from the river that was not of sufficient magnitude to create outflow, or

outflow would occur solely as a result of precipitation. These two “incomplete pulse”

26 scenarios are not included in the “during pulse” calculations but they are incorporated in the total annual loading calculation. Also, the “during pulse” calculations are based only on the actual days of flow, whereas the annual calculations are not. Retention rates for

TN, SRP, and TP were all higher early in the year, before the start of the growing season.

Nutrient loading was variable not only between flood pulses, but also within flood pulse as shown in flood data from February 2004 (Fig. 2.4). Peaks in nutrient loading do not necessarily match with the moments of greatest nutrient concentration in the influent or the moments of greatest hydraulic flow. There was no significant correlation for either situation (r2 = 0.25 and r2 = 0.36, respectively).

2.5 Discussion

2.5.1 Water Quality Dynamics

The created oxbow in this study was an effective nutrient sink, especially during the initial mid-winter snowmelt season that is common in this part of Ohio. Nitrate- nitrogen removal was dependent on several hydrologic factors including hydraulic loading rate, water depth, and hydraulic retention time. The created oxbow wetland had a lower removal rate for nitrate during infrequent periods of high flow in the dry season, compared to the wet season flows. The created oxbow’s reduced ability to retain phosphorus during “dry season” flood pulses suggests that perhaps the wetland either reaches its assimilative capacity early in the year, or that the lower water levels and increased carp population combine to produce conditions conducive to the export of suspended solids and phosphorus.

27 Despite large differences in nutrient concentrations from year to year, retention of nutrients (by concentration) was similar each year. For example, total phosphorus retention by concentration was 26% each year despite the fact that mean inflow concentrations in 2004 were more than double the concentrations in 2003. The difference in mean annual concentration of total phosphorus in the inflow is likely the result of a greater percentage of the total samples in 2003 coming from the dry season than in 2004.

More samples were taken in 2004 because there was more rain, and thus more summer flooding in 2003 than in 2004. The difference in the timing and amount of flooding does not necessarily explain the difference in influent nitrate-nitrite levels, which can fluctuate dramatically from year to year in the Olentangy River (Mitsch et al., 2005a).

The increase in turbidity in the open water portion of the oxbow in 2004 coincided with an increased number of common carp (Cyprinus carpio) in the wetland.

Carp swam up the outflow of the oxbow during a large flood pulse from the river in late spring, and took up residence in the wetland.

Much of the fluctuation in total nitrogen is explained by changes in the TKN concentration. Ammonium concentrations within the wetland were negligible

(< 0.05 mg L-1); therefore, since the increase and subsequent decrease in TKN matches vegetation patterns observed in the created oxbow (see below) it is likely that the emergent marsh area contributes organic nitrogen to the water column.

2.5.2 Spatial Patterns

Ninety-two percent of the nitrate-nitrite loss occurs in the emergent marsh in the upper third of the oxbow. This indicates that at this point either the concentration of

28 nitrate-nitrite is too low to be further reduced by wetland processes or that the lack of emergent vegetation in open water areas results in an environment that is not conducive to denitrifying bacteria. The environment may be less suitable because the open water area has fewer plant roots or other plant tissues, and subsequently less organic matter on the benthic substrate for denitrifying bacteria to use as a carbon source. This explanation may also account for the differences between the nitrate concentration in the central channel and the edges of the wetland, as the emergent biomass is significantly greater at the edges. The increase and decrease in TKN and nitrate-nitrite appear to be dependent on inverse conditions. The portions of the wetland that show the greatest decrease in

- NO3 are the areas with the greatest increase in TKN.

There was a greater concentration of TP and SRP along the east bank than along the west bank. Phosphorus can be released from dried soils rich with sorbed phosphorus when these soils are inundated (Olilia et al., 1997). It is likely that higher concentrations of phosphorus in the edge zones were a result of the wetland area expanding and contracting as the oxbow received and discharged water during and following pulses.

2.5.3 Nutrient Loading and Retention Rates

The rates of nitrate-nitrogen and total phosphorus retention are within the ranges

-2 -1 of normal nutrient retention rates in wetlands of 10 - 40 g-NO3-N m yr and 0.5 - 5 g-P m-2 yr-1 reported by Mitsch et al. (2000). The nitrate-nitrogen rate is at the low end of the range predicted by that paper while the phosphorus retention is at the high end. Long- term experience with the adjacent experimental wetlands at this same Ohio location (see

Mitsch et al., 2005b) suggest that nitrate-nitrogen retention has increased or remained

29 steady from year to year while total phosphorus retention has decreased over that time.

The total nitrogen retention rate is also similar to those reported by other studies of river- fed created wetlands (Table 2.5). The created oxbow in this study also retained a greater amount of nitrate-nitrogen and a lesser amount of phosphate-phosphorus when compared to other pulsing systems studied in Australia and South Carolina.

2.5.4 Vegetation dynamics in a diversion wetland

Macrophyte vegetation affects and is affected by water quality in riparian wetlands. In the seven years since the oxbow was created, the extent of the Typha- dominated community has not expanded beyond the upper third of the wetland and this is where most of the nitrate-nitrogen is removed. The vegetation restriction to this location is likely due to the significant reduction of available nutrients in the wetland. By the end of the emergent marsh, nitrate-nitrite concentrations have been reduced to low levels, allowing other plants to better compete with Typha (Koch and Reddy, 1992). Another likely factor is that water levels during spring vegetation emergence are typically quite deep and there is minimal drawdown along the edges of the lower two-thirds of the oxbow. This is not favorable for Typha spp., or any other macrophyte, growth or germination (Keddy and Reznicek. 1986, Squires and van der Valk, 1992). Lastly, there is a long unobstructed east-west fetch on the lower two-thirds of the wetland. It is possible that the Typha community stops where it does more as a result of wind and subsequent wave action than as a result of the nutrient concentrations or water depth.

This question will be better answered as trees on the southwest bank continue to mature and form a more significant windbreak.

30 2.6 Conclusions

The wetland was a net sink for nitrogen and phosphorus during snow-melt, most spring floods, and on an annual basis. The wetland did export nitrogen and phosphorus during some of the flood pulses, most notably during river pulses caused by dry season thunderstorms. Overall, this oxbow design has shown itself to be a success in ecological terms. It would be worth replicating this wetland design in other locations to examine how it functions under a variety of climatic and hydrological conditions.

The created oxbow wetland has developed a diverse and reasonable assemblage of emergent macrophytes although it is dominated by Typha sp. in its upper third closest to the inflow. After nutrients are depleted by this emergent marsh section, more plant diversity occurs in the lower two-thirds of the wetland towards the outlflow. Frequent fluctuations in the water level in the southern section during spring flood events, and the drying out of the southern section into a large mudflat during late summer likely accounts for many of the differences between the two parts of the wetland.

2.6 Acknowledgements

Support provided by U.S. Department of Agriculture NRI CSREES Award 2003-

35102-13518 and a Payne grant from the Ohio Agricultural Research and Development

Center of The Ohio State University. Olentangy River Wetland Research Park publication 07-002. We also thank Maria Hernandez for coordinating the laboratory analysis of nutrient concentrations and Colleen Fink for her help in biomass harvesting.

31 2.7 Literature Cited

American Public Health Association. 1998. Standard Methods for the Analysis of Wastewater, 20 ed., APHA, Washington, DC.

Blahnik, T. and J.W. Day. 2000. The effects of varied hydraulic and nutrient loading rates on water quality and hydrologic distributions in a natural forested treatment wetland. Wetlands 20:48-61.

Boustany, R.G., C.R. Crozier, J.M. Rybczyk, R.R. Twilley. 1997. Denitrification in a south Louisiana wetland forest receiving treated sewage effluent. Wetlands Ecology and Management 4:273-283.

Casey, R.E. and S.J. Klaine. 2001. Nutrient attenuation by a riparian wetland during natural and artificial runoff events. Journal of Environmental Quality 30:1720-1731.

Cole, C.A. and D. Schafer. 2002. Section 404 wetland mitigation and permit success criteria in Pennsylvania, USA, 1986-1999. Env Management 30:508-515.

Day J.W., C.J. Madden, R.R. Twilley, R.F. Shaw, B.A. McKee, M.J. Dagg. 1995. The influence of Atchafalaya River discharge on Fourleague Bay, Louisiana (USA). pp. 151–160. In: Dyer K.R. and R.J. Orth (eds), Changes in Fluxes in Estuaries. Olsen and Olsen, New York,

Fennessy, M.S., C. Brueske, W.J. Mitsch. 1994. Sediment deposition patterns in restored freshwater marshes using sediment traps. Ecological Engineering 3:409-428.

Fink, D.F. and W.J. Mitsch. 2004 Seasonal and storm event nutrient removal by a created wetland in an agricultural watershed. Ecological Engineering 23:313-325.

Galat, D.L., L.H. Frederickson, D.D. Humburg, K.J. Bataille, J.R. Bodie, J. Dohrenwend. 1998. Flooding to restore connectivity of regulated, large-river wetlands (Lower Missouri River). BioScience 48:721–733.

Harter, S.K. and W.J. Mitsch. 2003. Patterns of short-term sedimentation in a freshwater created marsh. Journal of Environmental Quality 32:325-334.

Henry, C.P., C. Amoros, N. Roset. 2002. Restoration ecology of riverine wetlands: A 5- year post-operation survey on the Rhône River, France. Ecological Engineering 18:543-554.

Hensel P.F., J.W. Day, D. Pont, J.N. Day. 1998. Short-term sedimentation dynamics in the Rhone River Delta, France: The importance of riverine pulsing. Estuaries 2:52–65.

32 Kadlec, R. H. 1994. Detention and mixing in free water wetlands. Ecological Engineering 3:1–36.

Kao, J.T., J.E. Titus, W. Zhu. 2003. Differential nitrogen and phosphorus retention by five wetland plant species. Wetlands 23:979-987.

Keddy, P. A., and A. A. Reznicek. 1986. Great Lakes vegetation dynamics: the role of fluctuating water levels and buried seeds. Journal of Great Lakes Research 12:25–36.

Knight, R. L., T. W. McKim, H. R. Kohl. 1987. Performance of a natural wetland treatment system for wastewater management. Journal of Water Pollution Control Federation 59:746–754.

Koch, M.S., and K.R. Reddy. 1992. Distribution of soil and plant nutrients along as a trophic gradient in the Florida Everglades. Journal of Soil Science Society of America 56:1492-1499.

Kozlowski, T.T. 2002. Physiological-ecological impacts of flooding on riparian forest ecosystems. Wetlands 22:550-561.

Lachet Instruments. 2000. Methods Manual. Lachet Instruments, Milwaukee, WI, USA.

Lane, R.R., J.W. Day, and B. Thibodeaux. 1999. Water quality analysis of a freshwater diversion at Caernarvon, Louisiana. Estuaries 22:327–336.

Lane, R.R., H.S. Mashriqui, G.P. Kemp, J.W. Day, J.N. Day, A. Hamilton. 2003. Potential nitrate removal from a river diversion into a Mississippi delta forested wetland. Ecological Engineering 20:237-249.

Mitsch, W.J., J.K. Cronk, X. Wu, R.W. Nairn D.L. Hey. 1995. Phosphorus retention in constructed freshwater riparian marshes. Ecological Applications 5:830-845.

Mitsch, W.J. and R.F. Wilson. 1996. Improving the success of wetland creation and restoration with know-how, time, and self-design. Ecological Applications 6:77-83.

Mitsch, W.J and J.G. Gosselink. 2000. Wetlands, 3rd ed. John Wiley and Sons, New York.

33 Mitsch, W.J., A.J. Horne, R.W. Nairn. 2000. Nitrogen and phosphorus retention in wetlands —Ecological approaches to solving excess nutrient problems. Ecological Engineering 14:1-7.

Mitsch, W.J., J.W. Day, W. Gilliam, P.M. Groffman, D.L. Hey, G.W. Randall, N. Wang. 2001. Reducing nitrogen loading to the Gulf of Mexico from the Mississippi River Basin: strategies to counter a persistent ecological problem. BioScience 51:373-388.

Mitsch, W.J., J.W. Day, L. Zhang, R.R. Lane. 2005a. Nitrate-nitrogen retention in wetlands in the Mississippi River Basin. Ecological Engineering 24:267-278.

Mitsch, W.J., L. Zhang, C.J. Anderson, A. Altor, M. Hernandez. 2005b. Creating riverine wetlands: Ecological succession, nutrient retention, and pulsing effects. Ecological Engineering 25:510-527.

Mitsch, W.J. and J.W. Day, Jr. 2006. Restoration of wetlands in the Mississippi-Ohio- Missouri (MOM) River Basin: Experience and needed research. Ecological Engineering 26:55-69.

Molles M.C., C.S. Crawford, L.M. Ellis, H.M. Valett, C.N. Dahm. 1998. Managed flooding for riparian ecosystem restoration. BioScience 48:748–756.

Nairn, R.W. and W.J. Mitsch. 2000. Phosphorus removal in created wetland ponds receiving river overflow. Ecological Engineering 14:107-126.

National Research Council. 1992. Restoration of Aquatic Ecosystems. National Academy Press, Washington, DC. 552 pp.

Novitzki, R.P. 1982. Hydrology of Wisconsin Wetlands: Wisconsin Geological and Natural History Survey Information Circular 40, Madison, WI, 22 p.

Olila, O.G, K.R. Reddy, D.L. Stites. 1997. Influence of draining on soil phosphorus forms and distribution in a constructed wetland. Ecological Engineering 14:107-126.

Phipps RG and W.G. Crumpton. 1994. Factors affecting nitrogen loss in experimental wetlands with different hydrologic loads. Ecological Engineering 3:399–408.

Raisen, G.W. and D.S. Mitchel. 1995. The use of wetlands for the control of non-point source pollution. Water Science and Technology 32:177-186.

34 Reddy, K.R. and D.A. Graetz. 1988. Carbon and nitrogen dynamics in wetland soils. Pp. 307-318. In: D.D. Hook, W.H. McKee, Jr., H.K. Smith, J. Gregory, V.G. Burrel, M.R. DeVoe, R.E. Sojka, S. Gilbert, R.Banks, L.G. Stolzy, C. Brooks, T.D. Mathews, and T.H Shear (eds.) The Ecology and Management of Wetland, Vol. 1: The Ecology of Wetlands. Timber Press, Portland, OR.

Reed, P.B. Jr. 1998. National list of plant species that occur in wetlands: Northeast (Region I). U.S. Fish and Wildlife Service, Washington, DC, Biological Report 88 (26.1).

Reyes E, M.L. White, J.F. Martin, G.P. Kemp, J.W. Day, A. Aravamuthan. 2000. Landscape modeling of coastal habitat change in the Mississippi Delta. Ecology 81:2331–2349.

Squires, L. and A.G. van der Valk. 1992. Water-depth tolerances of the dominant emergent macrophytes of the Delta Marsh, Manitoba. Canadian Journal of Botany 70:1860-1867.

Toner, M. and P. Keddy. 1997. River hydrology and riparian wetlands: A predictive model for ecological assembly. Ecological Applications 7:236-246.

Toth L.A., S.L. Melvin, D.A. Arrington, J. Chamberlain. 1998. Hydrologic manipulations of the channelized Kissimmee River. BioScience 48:757–765.

U.S. Environmental Protection Agency. 1983. Methods for Chemical Analysis of Water and Wastes. 600/4-79-020, U.S. Environmental Protection Agency, Cincinnati, OH.

Wong, T.H.F., and N.L.G. Somes. 1995. A stochastic approach to designing wetlands for stormwater pollution control. Water Science and Technology 32:145-151.

Zedler, J.B. and J.C. Callaway. 2000. Evaluating the progress of engineered tidal wetlands. Ecological Engineering 15:211-225.

Figure 2.1 Site map of the Wilma H. Schiermeier Olentangy River Wetland Research

Park on The Ohio State University campus. The 2.8 ha created oxbow wetland is between the experimental wetlands and the bottomland hardwood forest. Grab-sample locations are marked with a white circle and the seven biomass sampling transects are marked with white lines.

36 37

Figure 2.2 Annual hydrograph for a created diversion oxbow wetland and Olentangy

River in central Ohio, USA for (a) 2003 and (b) 2004.

Figure 2.3 Dominant vegetation communities in the created oxbow wetland in a) 2003

and b) 2004. The area in the 2004 map marked with the dotted lines shows the extent of

the spread Xanthium strumarium following prolonged drawdown after vegetation and productivity surveys in this study. The location of biomass sampling transects are marked on Fig. 2.1.

a) 2003 b) 2004

-Open water -Typha spp -Woody fringe -Salix spp. -Populus deltoides -Mixed communities -Eleocharis sp -Scirpus americanus -Juncus effusus -Pontedaria cordata -Xantium strumarium

0 200 feet 0 60 meters

Outflow Outflow

Figure 2.4 Mean total phosphorus (TP) reduction during flood pulses in 2003 (15 of 17 total pulses) and 2004 (7 of 7 total pulses). A positive percentage indicates a net reduction in TP whereas a negative percentage indicates a net export in TP. Error bars show standard error. The number of TP samples measured during each pulse is indicated below each mark.

42 100% 2003 pulses 75% 2004 pulses

50%

25%

0% 43 -25%

Percent reduction . -50%

-75%

-100% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

- - Figure 2.5 Kriging diagrams of the a) nitrate+nitrate (NO3 + NO2 ), b) total Kjeldahl nitrogen (TKN), c) total phosphorus (TP), and d) total suspended solids (turbidity-NTU) in a created diversion oxbow wetland in central Ohio. Nutrient sampling locations are shown in Fig. 2.1.

- -1 -1 a) NO3 mg-N L b) TKN mg-N L

Inflow Inflow Inflow

0 200 feet 2.00 0 60 3.50 2003 2004 2004 meters

0.05 0.60

Outflow Outflow Outflow c) TP mg-P L-1 d) Turbidity - NTU

Inflow Inflow Inflow Inflow

0.23 155 2003 2004 2003 2004

0.12 15

Outflow Outflow Outflow Outflow

Figure 2.6 Nitrate loading and export during the second major flood event of 2004. Net retention is the difference between the two. The flood event occurred February 3-15,

2004.

250

200 Outflow

-1 Inflow yr

-2 150 -N m g 100 m

0 2 Feb 2004 6 Feb 2004 10 Feb 2004 16 Feb 2004 Date

Mean duration of pulses, days

Pulses per year Inflow Outflow 1996-2002 7-8 5-6 12

2003 17 5-6 10

2004 8 5-6 9

Table 2.1 Frequency and duration of flood pulses in a created oxbow wetland on the

Olentangy River in central Ohio, USA.

NPP % of

Species g m-2 yr-1 total

Typha sp. 276 28.0

Leersia oryzoides 66 6.7

Juncus effusus 140 14.2

Verbesina alterniflora 35 3.5

Pontederia cordata 2 0.2

Scirpus americanus 191 19.4

Eleocharis sp. 209 21.2

Phragmites australis 26 2.7

Mixed wetland vegetation 23 2.3

Sparganium eurycarpum 16 1.6

Juncus canadensis 2 0.2

______

Total 986 100

Table 2.2 Estimated net aboveground primary productivity, based on peak biomass, in the

created oxbow in 2004 for the 11 most dominant emergent macrophytes. The “mixed

wetland vegetation” in this table refers to grouped FACW species that individually made up less than 0.1% of the total macrophyte biomass in the wetland.

Parameter Inflow Mid-point Outflow % Removal

2003 2004 2003 2004 2003 2004 2003 2004

SRP (μg-P L-1) 60 ± 1 (79)b 33 ± 1 (62)b 13 ± 1 (16)bc 36 ± 8 (6)b 27 ± 1 (97)c 19.0 ± 0.3 (87)c 55.3 43.5

TP (μg-P L-1) 92 ± 6 (8)b 203 ± 2 (77)b 77 ± 5 (7)bc 144 ± 11 (7)bc 68 ± 5 (6)b 150 ± 1 (102)b 25.7 26.3

- -1 b b bc bc b b b b NO3 (mg-N L ) 4.40 ± 0.04 (58) 1.81 ± 0.01 (83) 2.32 ± 0.14 (14) 0.66 ± 0.07 (7) 2.65 ± 0.02 (79) 0.77 ± 0.01 (120) 39.9 57.3

TN (mg-N L-1) - 3.04 ± 0.05 (77) - 3.17 ± 0.35 (7)c - 2.31 ± 0.02 (102)c - 24.0

TSS (mg L-1) 14.9 ± 1.1 (79)b 19.5 ± 0.9 (87)b 14.2 ± 1.9 (16)b 14.6 ± 2.9 (7)bc 9.1 ± 0.4 (99)bc 16.4 ± 0.4 (128)b 38.9b 15.9b 50 Mean ± Standard Error (number of samples) bSignificant difference between years (p < 0.05) cSignificant difference from upstream location (p < 0.05)

- Table 2.3 Mean nutrient concentrations and turbidity [SRP = soluble reactive phosphorus, TP = total phosphorus, NO3

- + NO2 = nitrate and nitrite, TN = total nitrogen, and TSS = total suspended solids] in the oxbow wetland created at the

Olentangy River Wetland Research Park, 2003-2004, when flooded river water is flowing through the wetland. TN =

- - TKN + NO3 + NO2 = (as N).

50 - NO3 -N TN SRP TP

Yearly Mean

Loading Rateg-X m-2 yr-1 32.2 ± 0.2 64.5 ± 0.4 0.48 ± 0.00 6.10 ± 0.04

Export Rate g-X m-2 yr-1 16.8 ± 0.2 32.2 ± 0.3 0.43 ± 0.01 1.62 ± 0.01

Retention

Rate g-X m-2 yr-1 15.4 ± 0.2 32.3 ± 0.2 0.05 ± 0.01 4.48 ± 0.03

Percent (by mass) 48 ± 3 50 ± 4 10 ± 1.0 73 ± 8.0

During Eight Pulses

Loading g-X m-2 pulse-1 0.97 ± 0.11 2.54 ± 0.47 0.036 ± 0.006 0.27 ± 0.07

Export g-X m-2 pulse-1 0.25 ± 0.01 1.34 ± 0.03 0.019 ± 0.004 0.19 ± 0.06

Retention

Total g-X m-2 pulse-1 0.71 ± 0.10 0.92 ± 0.44 0.016 ± 0.003 0.08 ± 0.01

Percent (by mass) 73.9 ± 11.1 40.6 ±6.8 46.8 ± 24.9 31.1 ± 6.1

Unit ‘X’ is N or P, as appropriate

- Table 2.4 Annual mean loading and removal of nitrate-nitrite (NO3 ), total nitrogen (TN), soluble reactive phosphorus (SRP), and total phosphorus (TP) during 2004 in the river diversion oxbow (ave ± std. error). Rates are in g-N m-2 yr-1 or g-P m-2 yr-1 as is

appropriate. The mean loading, export, and the retention of the nutrients during the eight

flood pulses in the mitigation oxbow wetland in 2004 are also shown. Rates and the mass

retention are calculated according to the number of actual days of flow; note that the

duration of the inflow and outflow are different.

Annual removal Removal per pulse - -2 -2 - -2 -2 Study Site g NO3 -N m g P m g NO3 -N m g-P m Reference Sustainable retention rates 10 – 40 0.5 – 5 Mitsch et al., 2000 for non-treatment wetlands Riparian oxbow, Columbus, OH 15 4.5 0.71 0.08 This study Riparian wetland, Cooke, 23 2.8 0.353 0.0523 Raisen and Mitchell, 1995 Australia Raisen et al.,1997 Riparian wetland, Cheraw State 0.038 0.067a Casey and Klaine, 2001 Park Golf Course, SC River-fed created wetland, Lake 11-38 1.4-2.9 Phipps and Crumpton, 1994 County, IL Mitsch et al., 1995 52 River-fed created wetland, 58-66 5.2-5.6 Spieles and Mitsch, 2000 Columbus, OH Nairn and Mitsch, 2000 Mississippi River diversion, 31 Lane et al., 1999 Caernarvon, LA Mitsch et al., 2005 a The Cheraw State Park site data are given in phosphate-phosphorus loading.

Table 2.5 Annual and per pulse removal of nitrate (as N) and total phosphorus in selected created river diversion and riparian

wetlands.

CHAPTER 3

THE EFFECT OF REMOVING HYDROLOGIC PULSING ON A RIVER

DIVERSION RIPARIAN WETLAND

3.1 Abstract

While hydrologic pulses are key to wetland function and often enhance the productivity and contribute to the stability of ecosystems, the overall effects of these pulses on biogeochemical processes in riparian river diversion wetland ecosystems have not been clearly demonstrated in ecosystem studies. One year of pulsing and one year of steady-flow inputs of nutrient laden river water were compared in a 3-ha created oxbow.

The wetland received 8 flood pulses in 2004 for a total of 27 m yr-1 of inflow and 20 m yr-1 of inflow in the steady flow year 2005. The retention rate was higher for nitrate- nitrogen, total nitrogen, and total phosphorus during the pulsing year then during the steady flow year. The only nutrient species that did not have a difference in retention rate in the two years was soluble reactive phosphorus. There were differences in spatial dynamics of most of the nutrient species in pulsing and non-pulsing years and between wet and dry seasons. In all cases, however, total nitrogen increased through the emergent marsh and then decreased across the open water basin. There was greater avian use of

53 the created oxbow wetland during the pulsing year than in steady flow conditions. The guild of bird species with the greatest difference in presence between the pulsing and steady-flow conditions was shorebirds.

Keywords: Avian use, created wetland, nitrate, phosphorus, wetland design

3.2 Introduction

Flood pulses have both positive and negative effects on the overall biotic function of wetland ecosystems (Day et al.,1977, 1995; Mitsch and Ewel, 1979; Mitsch et al.,

1979; Mitsch and Rust, 1984; Megonigal et al., 1997; McDougal et al., 1997; Galat et al.,

1998; Hein et al., 1999; Day et al., 2000; Reyes et al., 2004; Lane et al., 2004; and Mitsch et al., 2005c). But the overall effects of pulses on biogeochemical and biological processes in these riparian river diversion wetland ecosystems are not clear, nor have they been clearly demonstrated in ecosystem studies (Mitsch and Day, 2006). The role of pulsing in river ecology has evolved from the River Continuum Concept (Vannote et al.,

1980) to the Flood Pulse Concept described by Junk (1999) and Tockner et al. (2000) which emphasizes the exchange between a river and its floodplain as the primary factor affecting the function of both systems.

Stabilizing water levels or managing them outside the range of historic fluctuations eliminates the dynamic patterns that allow a diversity of wetland species and communities to thrive (Bedford, 1996). Minimizing the fluctuations in Great Lakes fringing wetlands through active management (Wilcox, 1993; Wilcox and Whillans,

1999) has run counter to the maintenance of hemi-marsh conditions that benefit most

54 wetland bird species (Gibbs et al., 1991; Steen et al., 2006). However little literature exists studying whether this same relationship exists in riparian wetlands that have a flood driven pulsing hydroperiod. Understanding the relationships between birds and wetland forcing functions is useful as aquatic birds have been identified as indicators of wetland processes (Fernandez et al., 2005), especially water eutrophication processes

(Rutschke, 1987).

Excessive nutrients have led to significant eutrophication and subsequent hypoxic conditions in coastal waters throughout the developed world. One of the most notable of these conditions is the 20,000 km2 hypoxic zone in the Gulf of Mexico caused by excessive nitrates in the Mississippi River (Rabalais et al., 1996; 1998; 1999; 2002;

Turner et al., 2005). The retention of nutrients, especially nitrogen, in Midwestern USA wetlands could be of key importance as an economical solution to the Gulf of Mexico problem. The creation or restoration of 20,000 km2 of wetlands and riparian ecosystems in the Mississippi River Basin has been recommended as a means to remove nitrogen from the Basin to alleviate the Gulf hypoxic zone (Mitsch et al., 1999; 2001; 2005b). It is necessary to assess the role of these potential wetlands as nutrient sinks and the importance of hydrologic pulses in this function. Nutrient fluxes come as seasonal pulses related to river and runoff discharge, not as fluxes at constant concentrations. If floods flowing into wetlands are synchronized with high concentrations of nitrate, for example, higher nitrogen retention may result (Nixon et al., 1996; Lane et al., 1999). But if hydrologic pulses occur during periods of low nitrate, lower N retention in the wetland may result (Spieles and Mitsch, 2000).

3.2.1 Goals and Objectives

This project examines the effects of removing pulse fluxes of nutrient-laden river waters into created riparian diversion wetlands on a fourth-order river in central Ohio,

USA, and replacing these natural flood pulses with artificial steady-flow inputs. Wetland function under both pulsing and steady flow conditions was determined through measurements of avian use and fluvial inflow and outflow of nutrients.

The goal of this study is to investigate the effects of removing seasonal flood pulses on wetlands receiving diverted river water in the Upper Mississippi River Basin.

Pursuant to this goal I had the following objectives:

1. Investigate the effect of removing hydrologic pulsing on the ability of a riparian

wetland to remove nutrients;

2. Determine the effects of removing seasonal flood pulses on avian use in the

wetland.

3.3 Methods

3.3.1 Site Description

The wetland in this study is a created river diversion wetland located at the Wilma

H. Schiermeier Olentangy River Wetland Research Park at The Ohio State University in

Columbus, Ohio, USA (latitude 40.021°N, longitude 83.017°E; Fig. 3.1). A river diversion wetland is a wetland on the adjacent floodplain or behind artificial levees that

56 receive water by pumping or flood flows from the main channel of a river (Mitsch and

Day, 2006). From its creation in 1996 until 2004, the 3-ha created riparian wetland

(referred to here as a created oxbow) received, on average, 7 to 8 natural flood pulses per year from the Olentangy River. Inflow from 1998-2004 averaged 20 ± 4 m yr-1. The

Olentangy River provides frequent short (5-6 days of inflow) flood pulses into the created oxbow which typically result in 9-12 days of outflow from the wetland. Water flows into the northern tip of the wetland through a Red Field TideflexTM check valve when the river elevation is higher than the wetland; the valve closes when the river elevation is lower than the wetland water level and water then flows back to the Olentangy River though an outflow control weir (Fig. 3.2; Fink and Mitsch, in press).

3.3.2 Hydrologic conditions

The 3-ha created oxbow was monitored during one year of flood-pulse conditions,

2004, and then subjected to one year with steady-flow conditions, 2005 (Fig. 3.3).

Pulsing conditions during the first year (April 2004 to March 2005) were created entirely by natural flooding of the river itself. Steady-flow conditions during the second year

(April 2005 – March 2006) were created and controlled by a large submersed bypass- pump on the river intake that eliminated floods and created artificial steady-flow conditions. I attempted to provide a similar volume of water to the wetlands in the non- pulsing year with the pumps as was provided naturally by the river during the pulsing hydroperiod year.

Flows into and out of the wetland were measured using a Swofer 2100 current meter for water velocity and an ISCO 730 bubbler module. A simple mathematical

57 model was developed to describe the inflow based on the relative elevation of the river,

the oxbow water surface elevation, and the flooded cross-sectional area of the inflow

pipe. Outflow was estimated from the stage of water within the wetland and the shape of

the outflow weir (USBR, 1997). A daily hydrologic budget was then developed from

daily readings of river elevation and wetland stage.

To quantify the differences between the pumped steady flow and the natural river

pulsing, the Richardson-Baker flashiness index for streams was applied to the wetland

(Baker et al., 2004).

(1)

Where q is the daily inflow rate. The index measures oscillations in flow relative to total

flow, and as such, provides a useful characterization of hydrologic inputs. This index was calculated on a monthly time step to allow a statistical comparison of flow stability between the pulsing and non-pulsing years.

3.3.3 Water Quality

Dawn and dusk inflow and outflow 500 mL grab samples were taken between

April 2004 and March 2006 on days when there was flow into or out of the created oxbow wetland. During the year with the pulsing hydroperiod (2004), when the

Olentangy River stage reached or exceeded 220.9 m MSL and was higher than the surface water elevation of the wetland, river water was diverted into the oxbow wetland.

More detailed sampling was conducted during some storm events using ISCO 6874 autosamplers at the inflow and outflow points of the created oxbow. Autosamplers

58 collected hourly 500 mL samples during high flow periods. Grab samples were also

taken at 11 locations around the wetland to determine the spatial distribution of nutrients

and suspended sediments (Fig. 3.1). During the pulsing year, these samples were taken

every 4 days during eight flow events; during the steady-flow year, these samples were

taken twice per month. The grab samples (n = 700 over the two water years) were

- analyzed for nitrate + nitrite (NO3 ), total Kjeldahl nitrogen (TKN), soluble reactive phosphorus (SRP), total phosphorus (TP), and total suspended solids (TSS).

3.3.4 Chemical Analysis

Nitrate + nitrite, soluble reactive phosphorous (SRP), total Kjeldahl nitrogen

(TKN), and total phosphorus (TP) were analyzed using Lachet (2000) modifications of

standard methods (USEPA 1983, APHA 1998) with a Lachet QuickChem flow-injection

analyzer (FIA+ 8000). Manual and auto-sampler samples were split into filtered (0.45

µm) and unfiltered subsamples within 48 hours of sampling, and analyzed for appropriate

chemistries. TP and SRP were analyzed with an ascorbic acid and molybdate color

reagent method. TP samples were digested in a block digester using 0.5 mL of a digestion

o solution made up of H2SO4, K2SO4 and mercuric sulfate and heated for 1 hour at 160 C

and 1.5 hours at 380oC. TKN was determined using the salicylate and hypochlorite

method. Nitrate + nitrite was determined by the sulfanilamide method after reduction in

a cadmium column. Throughout this paper references to nitrate or nitrate-N should be

interpreted as references to nitrate + nitrite. Total nitrogen (TN) was estimated to be the

sum of TKN and nitrate + nitrite-N. TKN is a measure of ammonia-N and organic-N.

Measurements of ammonium ion concentration in this wetland are negligible (Fink and

59 Mitsch, in press). Therefore the TKN fraction of the TN can be considered a measure of the organic-N concentration in the wetland surface water. Total suspended solids were determined using an empirical correlation with turbidity developed for the Olentangy

River (Harter and Mitsch, 2003). Turbidity was measured in the laboratory using a Hach

2100 N turbidimeter.

3.3.5 Avian use

Avian use of the created oxbow was observed twice-per-month during the pulsing hydroperiod (2004) and steady-flow (2005) years Birds were observed from the same

60 locations as the grab samples were taken (Fig. 3.1), and were only recorded when they

were observed visually. A bird was considered to be using the wetland and recorded only

if it was observed feeding, hunting, sitting, swimming, or walking through the wetland.

Avian species that were only observed flying over the wetland area were not counted.

Because the birds were not tagged, it was not possible to differentiate between

individuals, therefore only the species of bird observed was recorded and not the number

of birds observed.

3.3.6 Statistical Methods

The mean, standard deviation, and correlations between nutrient and other environmental values were calculated using Microsoft Excel and SPSS 11.0. Direct comparisons of single variables between the first and second water years of the study were done using a homoscedastic, two-tailed Student’s T-Test for equal means. All tests were conducted at a 95% confidence interval (α = 0.05).