Nutrient assimilation and root porosity responses of aquatic macrophytes to application of dairy effluents using constructed wetlands by Bryce Reed Romig A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Land Rehabilitation Montana State University © Copyright by Bryce Reed Romig (1993) Abstract: This study was developed to determine the effect of two liquid dairy waste loading concentrations on wastewater nutrient removal by four north-temperate wetland species, Nebraska sedge (Carex nebraskensis Dewey), beaked sedge (Carex rostrata Stokes), broad-leaved cattail (Typha latifolia L.), and panicled bullrush (Scirpus microcarpus Presl); and, to evaluate the role of root aerenchyma in wastewater treatment. Microcosms were planted with these four species and were treated with two concentrations (30% and 70%) of liquid dairy waste. Unplanted microcosm units receiving only H2O served as controls. Plants were grown for one month prior to applications of wastewater. . Soil waters were sampled from ceramic cups embedded in the soil medium at three soil depths (15, 30, 45 cm) prior to each application of effluent. Water was analyzed for pH, total Kjeldahl-nitrogen (TKN), nitrate nitrogen (N03-N) , and chemical oxygen demand (COD). Plants were harvested and measured for biomass production, root porosity, and total nitrogen at the end of the treatment cycle. Results showed differences (p < 0.05) in plant root porosity that corresponded to different levels of effluent concentration. No correlation was observed between percent root porosity in the wetland species and the level of contaminant removed. No differences were observed in soil pore water TKN at the 30 and 45 cm depths; however, significant differences were observed at the 15 cm depth. Soil surface TKN also was significant with respect to effluent concentration and plant treatment. Microcosms receiving the low concentration (70% dilution) showed stabilization of N level over the two month period while units receiving the high concentration (30% dilution) showed increased levels of nitrogen (N) . The results of this study suggested that volume of root porosity was not a determinant factor in plant efficiency of waste modification. The effects of plant uptake on nutrient removal from applied wastes also were inconclusive. However, wastewater modification occurred chiefly in the upper 15 cm of soil and was affected by plant rooting depth and effluent concentration. These results suggest wetlands are capable of removing N and organic constituents from livestock wastes provided threshold loading rates are not exceeded.  ( NUTRIENT ASSIMILATION AND ROOT POROSITY RESPONSES OF AQUATIC MACROPHYTES TO APPLICATION OF DAIRY EFFLUENTS USING CONSTRUCTED WETLANDS by Bryce Reed Romig A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Land Rehabilitation MONTANA STATE UNIVERSITY Bozeman, Montana June 1993 ® COPYRIGHT by Bryce Reed Romig 1993 All Rights Reserved rflS7i ii APPROVAL of a thesis submitted by Bryce Reed Romig This thesis has been read by each member of the thesis committee and has been found to be satisfactory regarding content, English usage, format, bibliographic style, and consistency, and is ready for submission to the College of Graduate Studies. TuU U . 1443 Date ^ ~i?g L, v— tv • 'TTt Co-Chairperson, Graduate dommittee#c Dat Z r). /7?? Co-Chairperson, Graduate Committee Approved for the Major Department 7U?3-?3 aai' Date Head, Major Department Approved for the College of Graduate Studies fr/Sk 3 Date Z Graduate Dean iii STATEMENT OF PERMISSION TO USE In presenting this thesis in partial fulfillment of the requirements for a master's degree at Montana State University, I agree that the Library shall make it available to borrowers under the rules of the Library. If I have indicated my intention to copyright this thesis by including a copyright notice page, copying is allowable only for scholarly purposes, consistent with "fair use" as prescribed in the U.S. Copyright Law. Requests for permission for extended quotation from or reproduction of this thesis in whole or in part may be granted only by the copyright holder. Signature / / f t /0. /911Date VITA Bryce Reed Romig was born in Albuquerque, New Mexico, the second son of Donald A. Roitiig and Edith L. Powers on the 2Oth day of August, 1964. Raised in Santa Fe, New Mexico, Bryce graduated from Santa Fe High School in 1982 to enter the University of Idaho in Moscow, Idaho the fall of the same year. He graduated cum laude in May of 1986 with a Bachelors degree in Forest Resources Science from the College of Forestry Wildlife and Range. Post-education work in mine reclamation activities in south-central Idaho and management of a native seed outlet for the plants of the southwestern United States culminated in the pursuit of a Master's Degree in Land Rehabilitation from Montana State University. V ACKNOWLEDGEMENTS I acknowledge the financial assistance of the Montana Water Resource Center in conjunction with a grant from the United States Geological Survey. Many thanks to Dr. Robyn Tierney, Animal and Range Science Department, Montana State University for encouragement and assistance throughout my graduate activities. Thank you to the other members of my graduate committee Dr. Frank Munshower, Dennis Neuman, and Dr. Jon Wraith. To Edith Powers, Donald Romig, Ann Girand, Ken, Doug, Kate and extended family, thank you for undying support and inspiration. I also wish to extend a personal note of thanks to Peggy Wright, whose lessons will remain the strongest. vi TABLE OF CONTENTS Page ACKNOWLEDGEMENTS........................ . ............ V TABLE OF CONTENTS................. vi LIST OF TABLES ........................................ viii LIST OF FIGURES................. xii A B S T R A C T .......................... xiii O B J E C T I V E S ............... I 'REVIEW OF L I T E R A T U R E ..................... 2 Livestock Wastes Impacts on Health and. Environment ................................... 2 Current Management of Livestock Wastes ........... 3 The Wetland Alternative . . . 5 Aquatic Plants in Wastewater Treatment ........... 6 The Wetland System as a Wastewater Treatment Reactor ....................................... 7 Wetland Soils and Soil Redox Status ........... _. 10 Aerenchyma and the Formation of Cortical Air ̂ Space .......................................... 12 Aerenchyma F u n c t i o n .......... • 14 Plant Gas Exchange Mechanisms ......... . . . . . 16 Rhizosphere Oxygenation . . . . . . . . . . . . . 19 Nitrogen Dynamics in Saturated Soils . . . . . . . 21 M E T H O D S.................................................. 26 Experimental Design . ............ 26 Statistical Analyses .............................. 26 Microcosms......................................... 28 Soil Collection and Preparation . . . 29 Oxidation-Reduction Potential .................... 32 Plant Collection and Preparation ................. 33 Effluent Collection and Preparation ............ 34 Sampling and Analysis ............................ 35 Effluents . . . . . .......................... 35 Microcosm Soil Water Sampling ............... 37 Plant S a m p l i n g ............................... 37 Soil Sampling ..................... 38 Quality Control/Quality Assurance ........... 38 RESULTS ................. 40 General Observations ........................ . . . 40 System parameters ............................. • • 41 Microcosm Oxidation-Reduction Potential . . . . . 42 vii Plant Characteristics.......................... . 42 Plant Biomass ................ 42 Plant Nitrogen Concentration ............... 44 Plant Assimilated Nitrogen .................. 47 Plant Nitrogen Summary ...................... 47 Root Porosity.......... 48 Liquid Dairy Effluents ............................ 51 Soil Water Samples................................. 55 Soil Water pH . . . .... ..................... 55 Soil Water C O D .............................. 55 Soil Water NO3 - N ............................ 59 Soil Water T K N .......... 59 Soil Nitrogen L e v e l s ............................... 65 D I S C U S S I O N ...................._......................... 68 Oxidation-Reduction Potential . . . . 68 Plant Characteristics .................... 70 Plant Nutrient Assimilation ................. 70 Plant Root Porosity . . . ......................... 71 Liquid Effluents in the Non-Wetland Condition . . 72 Soil Water Characteristics ........................ 73 Soil Water p H ............................ 73 Soil Water C O D .............................. 74 Soil Water NO3 -N .............................. 76 Soil Water T K N .............................. 77 Soil Nitrogen L e v e l s ......... 78 CONCLUSIONS.......... 80 LITERATURE CITED ....................................... 83 I A P P E N D I C E S ................. 99 APPENDIX A - Plant Descriptions . .............. 100 APPENDIX B - COD Regression.............. 103 APPENDIX C - ANOVA O u t p u t ....................... 105 Page viii LIST OF TABLES Table 1. Experimental design for redox measurement in greenhouse m i c r o c o s m s ............................. 2. Waste application schedule for wetland wastewater treatment study (1992) ............... 3. Sample analyses conducted for influent and effluent water parameters ........................ 4. Microcosm characteristics and baseline soil and interstitial water parameters ........... 5. Microcosm oxidation-reduction potential (mV) by depth, concentration and presence or absence of Carex rostrata (Caro) ............. 6. Effect of effluent concentration on mean (± S.E.) root biomass (g) ̂ . I. Effect of effluent concentration on mean (± S.E.) aboveground biomass (g) ................. 8. Effect of effluent concentration on mean (± S.E.) total plant biomass ( g ) ............. • • 9. Effect of effluent concentration on mean (± S.E.) nitrogen concentration in root biomass (mg g j ................. .. . . .................. 10. Effect of effluent concentration on mean (± S.E.) nitrogen concentration in aboveground biomass (mg g ) ................................... II. Effect of effluent concentration on mean (± S.E.) total nitrogen in root biomass (mg) . . . 12. Effect of effluent concentration on mean (± S.E.) total nitrogen in aboveground biomass (mg) 13. Effect of effluent concentration on mean (± S.E.) total nitrogen uptake (mg) ............. 14. Mean (± S.E.) percent root porosity of four wetland plant species receiving biweekly applications of dairy effluent ......... Bage 32 3 5 36 42 I 43 43 45 45 46 46 49 49 50: 50 ix LIST OF TABLES— Continued Table Page 15. Composition of liquid dairy effluents (n=2) applied to microcosms on a biweekly basis . . . . 52 16. Change in effluent characteristics in non-wetland condition. Data are for low concentrate preparation (n=2) . . . . 53 17. Change in effluent characteristics in non-wetland condition. Data are for high concentrate preparation (n=2) . . . . 54 18. Mean (± S.E.) microcosm pH by effluent concentration for five sampling periods (n=45) . . 56 19. Mean (± S.E.) microcosm pH by soil depth for five sampling periods (n=45) ................. 56 to O Mean (± S.E.) microcosm pH by plant species for five sampling periods (n=36) ................. 57 21. Mean (± S.E.) microcosm COD (mg O2 I 1) by concentration for three sampling periods (n=45) 58 22. Mean (± S.E.) microcosm COD (mg O2 I 1) by soil depth for three sampling periods (n=45) . . 58 23. Mean (± S.E.) microcosm COD (mg O2 I 1) by plant species for three sampling periods (n=36) 58 24. Mean (±S.E.) microcosm NO3 -N (mg I 1) by concentration for five sampling periods (n=45) . . 60 25. Mean (± S.E.) microcosm NO3 -N (mg I 1) by soil depth for five sampling periods (n=45) ........... 60 26. Mean (± S.E.) microcosm NO3 -N (mg I 1) by plant species for five sampling periods (n=36) . . 61 27. Mean (± S.E.) microcosm TKN (mg I 1J by concentration for five sampling periods (n=36) . . 61 28. Mean (± S.E.) microcosm TKN (mg I 1) by soil depth for five sampling periods (n=36) . . . 63 Mean (± S.E.) microcosm TKN (mg I 1) by plant species for five sampling periods (n=36) . .29. 63 X 30. Mean (± S.E.) total nitrogen (%) in microcosm soil by plant species after completion of effluent application (n=3) 67 31. Mean (± S.E.) total nitrogen (%) in microcosm soil by effluent concentration upon completion of effluent application (n=3) . . 67 32. Regression output for COD s t a n d a r d s ................................ 104 33. Analysis of variance output for redox measurements in planted and unplanted wetland microcosms . . 106 34. Analysis of variance output for biomass of plant r o o t s .............................. 106 35. Analysis of variance output for aboveground b i o m a s s .......................................... 106 36. Analysis of variance output for total plant b i o m a s s .......................................... 106 37. Analysis of variance output for nitrogen concentration in root b i o m a s s ................. 107 38. Analysis of variance output for nitrogen concentration in aboveground plant biomass . . . 107 39. Analysis of variance output for total nitrogen content of root b i o m a s s ........................ 107 40. Analysis of variance output for total nitrogen content of aboveground biomass . ............... 107 41. Analysis of variance output for total plant nitrogen uptake ................................. 108 42. Analysis of variance output for root porosity of four wetland plant s p e c i e s .......... 108 43. Repeated measures analysis of variance among subject effects on pH of soil water s a m p l e s ......................................... 108 LIST OF TABLES— Continued Table rage Table Rage 44. Repeated measures analysis of variance within subject effects on pH of soil water s a m p l e s ......................................... 109 45. Repeated measures analysis of variance among subject effects on COD of soil water s a m p l e s ......................................... 109 46. Repeated measures analysis of variance within subject effects on COD of soil water s a m p l e s ......................................... 109 47. Repeated measures analysis of variance among subject effects on nitrates in soil water samples............................... H O 48. Repeated measures analysis of variance within subject effects on nitrates in soil water samples.............................. H O 49. Repeated measures analysis of variance among subject effects on TKN of soil water s a m p l e s ................................ H O 50. Repeated measures analysis of variance within subject effects on TKN of soil water s a m p l e s ................................... H l 51. Repeated measures analysis of variance between subject effects on soil nitrate concentration................................... H l 52. Repeated measures analysis of variance within subject effects on soil nitrate concentration................................... H l 53. Repeated measures analysis of variance between subject effects on soil TKN levels . . . 112 54. Repeated measures analysis of variance within subject effects on TKN levels .............. 112 xi LIST OF TABLES— Continued LIST OF FIGURES Figure Fage 1. Side view schematic of microcosm experimental unit construction, including placement of porous ceramic c u p s ..................... 27 2. Placement and randomization of microcosms in g r e e n h o u s e ..................................... 30 3. Root porosity (%) versus effluent concentration for four wetland plant species.......... 48 4. Soil water TKN level (mg I 1) versus depth for three effluent concentrations ................. 64 5. Soil water TKN level (mg l’1) versus sampling date for three effluent concentrations ............. 66 xii xiii ABSTRACT This study was developed to determine the effect of two liquid dairy waste loading concentrations on wastewater nutrient removal by four north-temperate wetland species, Nebraska sedge (Carex nebraskensis Dewey), beaked sedge (Carex rostrata Stokes), broad-leaved cattail {Typha latifolia L.), and panicled bullrush (Scirpus microcarpus Presl); and, to evaluate the role of root aerenchyma in wastewater treatment. Microcosms were planted with these four species and were treated with two concentrations (30% and 70%) of liquid dairy waste. Unplanted microcosm units receiving only H2O served as controls. Plants were grown for one month prior to applications of wastewater. . Soil waters were sampled from ceramic cups embedded in the soil medium at three soil depths (15, 30, 45 cm) prior to each application of effluent. Water was analyzed for pH, total Kjeldahl-nitrogen (TKN), nitrate nitrogen (NO3'-N) , and chemical oxygen demand (COD). Plants were harvested and measured for biomass production, root porosity, and total nitrogen at the end of the treatment cycle. Results showed differences (p < 0.05) in plant root porosity that corresponded to different levels of effluent concentration. Nd correlation was observed between percent root porosity in the wetland species and the level of contaminant removed. No differences were observed in soil pore water TKN at the 30 and 45 cm depths; however, significant differences were observed at the 15 cm depth. Soil surface TKN also was significant with respect to effluent concentration and plant treatment. Microcosms receiving the low concentration (70% dilution) showed stabilization of N level over the two month period while units receiving the high concentration (30% dilution) showed increased levels of nitrogen (N) . The results of this study suggested that volume of root porosity was not a determinant factor in plant efficiency of waste modification. The effects of plant uptake on nutrient removal from applied wastes also were inconclusive. However, wastewater modification occurred chiefly in the upper 15 cm of soil and was affected by plant rooting depth and effluent concentration. These results suggest wetlands are capable of removing N and organic constituents from livestock wastes provided threshold loading rates are not exceeded. I OBJECTIVES This study was developed to investigate liquid wastes generated by dairy cattle and the treatment of these wastes using constructed wetland treatment systems. First, the removal of contaminants from simulated constructed wetland systems planted with four wetland plant species was studied following application of two concentrations of liquid dairy effluents. Second, plant rooting depth and root aerenchyma volume were studied to evaluate the role of rhizosphere oxygenation on transformation of N species and removal of chemical oxygen demand (COD) under saturated soil conditions. The objectives of this research were: 1) to evaluate the effect of two feedlot effluent loading concentrations on nutrient uptake by four temperate wetland species; 2) to determine the role of root aerenchyma in the removal of pollutants from dairy runoff. Two hypotheses based on these objectives were developed. I) Removal of water borne pollutants by wetland systems increases with the size and volume of cortical air tissue in the roots of wetland plants. The level of pollutant removal by an individual wetland plant increases with increased rhizosphere oxygenation. 2) 2 REVIEW OF LITERATURE Livestock Wastes Impacts on Health and Environment Agricultural developments include areas where livestock are concentrated for feeding and transport. Wastes generated at these facilities are often iu vulnerable hydrologic settings, and may impair water quality. Wastes include feces, urine, unused food products, bedding material, nitrogen (N), phosphorous (P), trace metals from mineral supplements, trace amounts of antibiotics, detergents, and high levels of organic matter (Viraraghavan and Kikkeri 1990). The chemical and physical properties in the solid and liquid fraction of livestock wastes are summarized by the USDA (1978). These properties vary with type of animal and nutritional management (Kirchmann and Witter 1992). The effects of livestock waste on water quality are well documented. Ammonia (NH3) , a primary N species in dairy effluents, is toxic to freshwater aquatic life and fish (EPA 1976). High levels of organic matter and biochemical oxygen demand (BOD) also are associated with fish kills. Organic loading and algal growth contribute to O2 depletion (Richardson and Davis 1987) . Nutrient loading further damages aquatic habitat through eutrophication. The BOD of livestock wastes is several times that of human sewage (Long and Painter 1991). Pathogens in livestock waste can contaminate estuarine habitats and render shellfish and fish unfit for consumption (National Research Council 1972) ., 3 Nitrate contamination of groundwater is a potential health threat in many industrial and agricultural areas (Keeney 1982, Adriano et al. 1971). Additionally, there is significant movement of NO3" through soil profiles and into groundwater under feedlots, corrals, and heavily fertilized fields (Adriano et al. 1971). Manure disposal on fallow croplands also poses a serious hazard to groundwater pollution. Several sources (Barnes and Bliss 1983, National Research Council 1972, Wright and Davison 1964) cite the health risks of NO3" including methanoglobinemia and formation of carcinogenic nitrosamines. Methanoglobinemia results from the reduction of NO3' to NO2". Nitrites oxidize hemoglobin to methemoglobin, a compound that does not carry oxygen (O2). Mammalian infants are particularly susceptible because of low body weight and the presence of nitrate-reducing bacteria in the upper gastrointestinal tract. More complete explanations of NO3" toxicity are offered by Wright and Davison (1964) . The EPA standard for NO3' in drinking water is 10 mg l"1 (EPA 1976). Current Management of Livestock Wastes Manures are beneficial to nutrient cycling in the plant root zone. One method commonly used to dispose of livestock wastes is application of manures onto cropland. Unfortunately, this approach is not a long-term solution to waste disposal. f 4 Land applications of manure increased soil surface organic matter (Mathers and Stewart 1974, Kuo 1981, Chang et al. 1991), subsurface electrical conductivity and sodium (Kuo 1981, Chang et al. 1991), increased subsurface concentrations of NH4+ (Ku o 1981) , and decreased soil pH (Chang et al. 1991). Concentrated manures can exceed the ability of the plant-soil system to cycle nutrients. Thus, selection of appropriate soil types for disposal is critical for land-based disposal of livestock wastes (Kuo 1981). Chang et al. (1991) and Mathers and Stewart (1974) found evidence of leaching of soluble salts and NO3 beyond most plant's rooting depths. They concluded that, in some agricultural soils, potential pollution exists from directly applied manure. Repeated annual applications of manures to soil causes a buildup of salts sufficient to lower soil productivity (Wallingford et al. 1975). Mathers and Stewart (1974) also cautioned against manure applications supplying excess N (salts). Clearly, land disposal of manures as a long­ term solution to livestock waste management is not advisable under most soil conditions. Given the problems associated with concentrated livestock production, livestock wastes must be handled in such a way as to protect surface and groundwater quality. 5 The Wetland Alternative The use of wetland and vegetated terrestrial systems to treat wastewater has been the subject of numerous publications and recent research (Reed et al. 1988). Young et al. (1980) used vegetated buffer strips for reduction of coliform organisms from feedlot runoff. Wetlands also function as filters, anaerobic and aerobic bioreactors, and mediums for exchange and transformation of wastewater constituents. Wetlands have key ecological features which have led many in the past ten years to investigate their potential in providing advanced treatment or polishing of secondary wastewater. Reed and Bastian (1979) indicated that impacts caused by organic loading are relatively small in wetlands receiving secondary wastewater. However, wetlands may be used to improve water quality with respect to N, suspended solids, BOD, P (Kadlec 1987a, Hurry and Bellinger 1990), coliform bacteria (Bavor et al. 1987), metals, and some organic pollutants (StowelI et al. 1981). Wetland plant species concentrate and in some cases assimilate heavy metals and radioactive materials (Wolverton 1987a). Regulatory \ considerations under EPA for treatment of wastewater using wetland systems were presented in Davis and Montgomery (1987). Cost comparisons of constructed wetland systems to contemporary wastewater facilities have been made by Crites and Mingee (1987) , Duffer (1982), Wolverton (1987b), and Richardson and Davis (1987). In general, wetland 6. treatment systems are cheaper to build and maintain than other types of sewage treatment systems (Cooper et al. 1989) . Lower nutrient concentrations than those normally attained in conventional advanced waste treatment processes can be achieved using constructed wetlands (Swindell and Jackson 1990). In referring to wetland research as inflow-outflow "black box" studies, Reddy and DeBusk (1987) observed that the significance of vegetated wetlands in contaminant removal varied with treatment system design, which in turn depended on the desired contaminant removal goals. Further, the variability in N removal rates as reported by different researchers arises from differences in design criteria, substrate of the aquatic system (Weber and Tchobanoglous 1986) and other factors that may effect the growth of microorganisms (Reddy et al. 1989). Aquatic Plants in Wastewater Treatment The role of aquatic plants in wastewater treatment systems has been discussed by many researchers. Ideally, wetland plants should have rapid growth rates, high nutrient contents, and high standing crop (Reddy and DeBusk 1987). DeBusk and Ryther (1987) suggested that large biomass I production, efficient assimilation of carbon, rapid growth rate in dense stands, and ease of vegetative propagation are also important growth characteristics. However, Reddy and DeBusk (1987) argued that nutrient removal via plant uptake is 7 governed more by nutrient concentration in the plant tissues than by plant productivity. Nutrient assimilative capacity of emergent macrophytes also is affected by wastewater composition, loading rate, water depth, sediment characteristics, O2 transfer to root zone, bio- and physio- chemistry at the root-water-sediment interface, plant density, plant cultural practices, and climate (Reddy and DeBusk 1987). Aquatic plant roots provide a medium for microbial growth, filtration of solids, translocation of O2, and improvement of soil permeability (Wood 1990). Bacterial activity is further enhanced by O2 release from plant roots (Armstrong and Armstrong 1988). Boyd and Hess (1970) and Wolverton (1987a) suggested that wetland plants may be harvested as forage. Wood (1990) stated that management of wetland wastewater treatment systems may include selective harvest to control plant community diversity and dominance. Nitrogen removal may be facilitated by harvest of plant biomass or manipulation of hydraulic retention (Whitehead et al. 1987). Gersberg et al. (1986) and Wolverton (1982) also noted that artificial wetland systems may be managed to include biomass production. , The Wetland System as a Wastewater Treatment Reactor A review of biogeochemical processes in wetland ! ecosystems with an overview of littoral O2 and inorganic carbon assimilation is given by Carpenter and Lodge (1986). 8 Bacterial metabolism and physical sedimentation improve water quality in wetlands and function much like conventional activated sludge and trickling filters. Microbially mediated chemical transformations largely dictate the success of, constructed wetland systems (Gersberg et al. 1986). Wetland and aquatic plants are not active in chemical transformations, but provide a physical structure to the aquatic environment that enhance root-microbial symbiosis (Wolverton 1987b) and wastewater treatment capability (Tchobanoglous 1987). Aquatic plants supply O2 to nitrifying microorganisms in microzones adjacent to roots and thus play an important role in N removal (Brix 1987). These rhizosphere interactions are perhaps best described by Reddy et al. (1989) : "Ammonium (NH4*") in the anaerobic zone of the sediment diffuses either into the root zone or the overlying water column as a result of concentration gradients. Some of the NH4"* is taken up directly by aquatic macrophytes and part is oxidized to NO3". The NO3" formed is either taken up by the plant or diffused into adjacent anaerobic zones where it is denitrified." Research has identified additional considerations in the design of wetland systems for water quality improvement. The roles of evaporation (Wood 1990), substrate effects on retention time (Wood 1990, Brix 1987, McIntyre and Rhia 1991), and methods of effluent application (Wood 1990, Rogers et al. 1990) also have been studied. Boyd (1969) found evidence that 9 aquatic macrophyte based treatment systems primarily designed as living substrate for microbial activity were effective in reducing suspended solids and BOD. Soils deep in the profile are thought to have little influence on N removal as one study (Wittgren and Sunblad 1990) found 85% loss of N in the surface 20 cm of soil. Kadlec (1987a) stressed the importance of algas and retention time for particle settling. Gersberg et al. (1986). found that treatment of N in trench type systems by bacterial metabolism requires an additional carbon source. System washout occurs when loading rate exceeds the assimilative capacity of the nitrifying-denitrifying bacteria in the system (Prakasam and Loehr 1972, Barnes and Bliss 1983) . Differences in pollutant removal efficiency caused by temperature variation are reviewed by Virafaghavan and Kikkeri (1990). Brix (1989) stated that plant cover, heat from microbial activity and influent wastewater temperature may help prevent the freezing of systems in the winter. Wood (1990), however, showed the temperature variations in wastewater during winter have little effect on effluent quality. An increase in effluent retention time must accompany a decrease in temperature to obtain, results that are similar to those observed at higher temperatures (Tchobanoglous 1987). 10 Wetland Soils and Soil Redox Status Knowledge of the soil oxidation-reduction condition and the measurement of redox potential (Eh) are valuable tools for interpreting soil nutrient fluxes and changes in nutrient availability. Ponnamperuma (1984) found that the oxidation- reduction potential of interstitial waters in flooded soils and sediments may explain changes in concentration of ecologically important ions (plant nutrients). Armstrong et al. (1990) and Bouldin et al. (1974) related oxidation status of soils to predicting the presence of N species and noted that a mosaic of oxidized and unoxidized zones may be present in wetland soils. Anaerobic conditions occur when the transport processes for O2 fail to keep pace with soil biological and chemical processes (Veen 1988). Variations in redox electrode readings may result because of electrode construction, placement of electrodes, and microsite differences (Cogger et al. 1992). While theoretical justifications of the measurements are made by several authors (Ives and Janz 1961, Bohn 1971), there are no widely accepted methods of measuring and interpreting field redox potentials (Cogger et al. 1992). Several researchers discuss the use of platinum microelectrodes for in-situ redox measurements (Mueller et al. 1985, Faulkner et al. 1989). Redox measurements are more reliably interpreted in evaluating hydric soils (Cogger et al. 1992). Thus, Eh frequently is used as an index of reducing or oxidizing condition in wetlands and wetland sediments. 11 This measure has been particularly useful, albeit controversial, in understanding the role of plants in modification of sediment chemistry. Vegetation shifts between vascular and non-vascular plants leads to corresponding shifts in sediment redox (Jaynes and Carpenter 1986). Haraguchi (1991) presented an argument against the role of plants in influencing sediment redox, and related oxidation-reduction potentials to properties of the substrate. Carpenter et al. (1983) related Eh to root zone oxygenation and corresponding decrease in soil water COQ. Armstrong et al. (1990) demonstrated that Phragmites can raise soil redox considerably during a single growing season. Armstrong (1964) suggested redox condition may play a role in distribution of plant species. While Wium-Anderson and Anderson (1972) found no seasonal variation in redox potential, Cogger et al. (1992) found field redox potentials varied by season and with flooding and draining. Anaerobic soils are frequently phytotoxic or adverse to plant growth. McKee et al. (1988) showed that significantly lower concentrations of sulfide (a phytotoxic anion) correspond to the presence of aerial foots in Rhizophora and Avicennia. Oxygen release in the root zone raises sediment Eh and causes precipitation of ferric and manganese oxyhydroxides on or around plant roots (Wium-Anderson and Anderson 1972). The effects of root zone modification by vascular plant root systems remains important since solutes regulated by redox 12' and pH include potentially toxic forms of aluminum, manganese, zinc, iron, and sulfur (Jaynes and Carpenter 1986). Aerenchvma and the Formation of Cortical Air Space Rhizosphere oxygenation depends on certain anatomical and physiological characteristics of wetland plants that facilitate gas transfer into the root zone. Many wetland plants contain aerenchyma, a cortical air tissue that provides a continuous system of interconnected gas-filled cavities or lacunae (Drew et al. 1981) in both stems and petioles (Dacey 1980, Kawase 1981). Rhizomes and roots possess an extensive system of air-filled cortical lacunae that are often more or less continuous with that of the stems, petioles and leaves (Sculthorpe 1967). Sculthorpe (1967) noted a lack of pit membranes in aerenchyma and suggested that gas may freely diffuse within these tissues. Patterns of aerenchyma development are species-specific (Armstrong 1979, Smirnoff and Crawford 1983). Burdick (1989) studied the effect of root age as a factor in aerenchyma development as originally suggested by Yu et al. (1969). Luxmore et al. (1972) related aerenchyma development to light intensity. Sojka et al. (1972) also showed that air space development is sensitive to temperature. In the early stages of root growth, aerenchyma forms through premature cell lysis about 10 mm behind in the root cortex about 10 mm behind the root apex (Campbell and Drew 13 1983). Aerenchyma formation in many wetland plants responds positively to poorly oxygenated environments (Justin and Armstrong .1987, Kawase 1981). Das and Jat (1977) indicated that reduced substrate aeration can promote breakdown of cortical tissues and increase internal porosity. Justin and Armstrong (1987) noted that some wetland plants do not form aerenchyma but survive in the wetland condition by forming shallow root systems. Aerenchyma formation is a relatively rapid process (Kawase and Whitmoyer 1980, Kawase 1981). Although roots form aerenchyma in conditions of low O2 concentration, Armstrong (1979) suggested that aerenchyma development is not solely a response to anoxia. Atwell et al. (1988) and Drew et al. (1979), however, showed that hypoxia stimulates the synthesis of ethylene and that an increase in internal ethylene concentration is closely related to induction of cell lysis. Both endogenous (Drew et al. 1981) and exogenous ethylene formation (He et al. 1992) have been shown to stimulate formation of aerenchyma. Further, Drew et al. (1981) and Van Noordwijk and Brouwer (1988) found that aerenchyma formation was inhibited when ethylene production was reduced in roots of some plant species. The role of nutrient availability in aerenchyma formation has also been noted. Aerenchyma may form under fully anaerobic conditions with the temporary removal of NO3 and NH4+ (Drew et al. 1989, Konings and Verschuren 1980). Konings and Verschuren (1980) also suggested that cavity formation in 14 aerated roots of maize was retarded by NO3 and NH4* ions. They attributed their observed results to a relationship between nitrate reductase activity and NO3" uptake. Although root porosities for terrestrial plants are lower than for aquatic or wetland plants or plants that have undergone a period of soil saturation, Jensen et al. (1969) reported root porosities of 3.6% for barley, 9.5% for corn, and 36.5% for rice. Armstrong (1979) found that internal air space in roots can be as high as 60% while Bristow (1975) reported 3 0 to 60% air space, in wetland plant roots. In water lilies, lacunae constitute 20 to 40 % of volume in the root and rhizome, and 50% in the petiole (Dacey and Klug 1979). Aerenchvma Function The four primary functions of aerenchyma are: 1) as a plant adaptive response to seasonal flooding or long-term soil saturation 2) as a gas transport mechanism for oxygenation of root tissues and the rhizosphere 3) as a mechanism for providing Q2 to anaerobic soil substrates 4) as an anatomical and morphological property of roots providing structural resistance to saturated soils. Although Kawase (1981) stated that aerenchyma contributes to survival of wetland species, differences in aerenchyma formation have been noted between flood-tolerant and flood- intolerant species (Pearson and Havill. 1988b, Burdick and 15 Mendelssohn 1990) . Crawford (1966) described the distribution of the genus Senecio in response to flood tolerance and anaerobic respiration. Laan et al. (1989) subsequently found aerenchyma development was the primary factor in determining flood tolerance of Rumex species. Smirnoff and Crawford (1983) and Justin and Armstrong (1987) also reported differences in effective root porosities between flood- tolerant and flood-intolerant species. Laing (1940) described the continuity of air space in plants for the supply of O2 and removal of carbon dioxide (CO2) from root tissue. Lacunae supply O2 for respiratory processes j in anaerobic sediment (Sculthorpe 1967, Dacey and Klug 1979), O2 storage (Armstrong 1967a, Sculthorpe 1967) and gas transport (Sculthorpe 1967). Sculthorpe (1967) stated that the measured linear gradients of O2 concentration in plants support the hypothesis that underground organs derive their O2 supply from aerial or floating portions of the plant. Lacunar gas space improves internal ventilation (Armstrong 1979, Drew et al. 1985), and reduces root respiratory demand (Armstrong 1979) . Van Raalte (1944) discussed the dependence of roots on shoots for O2 supply and rhiz©sphere detoxification. Armstrong and Boatman (1967) stated that oxidized iron precipitates around the roots forming an effective barrier against sulfide. ' Oxygen diffusion helps mitigate the toxic effects of reduced soil near the root tip (Armstrong 1967b), and serves to i 16 immobilize phytotoxins. (Armstrong 1972, 1979). The ability to withstand the phytotoxicity of sulfide is a function of a wetland plants' ability to withstand extreme anoxia (Pearson and Havill 1988a). Aerenchyma is not important in oxidation of endogenous sulfide (Pearson and Havill 1988a), but is important to flood tolerance (Pearson and Havill 1988b). Williams and Barber (1961) stated that aerenchyma functioned as a mechanical property in the protection of wetland plant roots embedded in saturated plastic (i.e. easily deformable) soils. Plant Gas Exchange Mechanisms As early as 1841, Raffeneau-Delile reported pressurization of gasses in the leaves of lotus. Laing (1940) observed concentrations of gasses in the internal atmosphere of the rhizomes and found that O2 decreased from shoot to root with increased concentrations of CO2. He also noted seasonal differences in the gas concentration between above and belowground portions of plants. Scholander; et al. (1955) suggested that pressure changes in the roots of Rhizophora sp. and subsequent O2 movement into roots through lenticels was i • . driven by rise and fall of tides. Plant O2 transport modelling was pioneered by William Armstrong (1967b) who determined that the O2 diffusion rate in plants is not affected by photosynthetic production of O2. Others have noted that gas diffusion is a function of 17 concentration gradient and resistance in lacunae (Sculthorpe 1967) . Gas exchange rates are a function of internal gas partial pressure, route resistances, and cortical gas space. Pick's laws of diffusion were used to describe the rate of gas exchange in the roots of both wetland and non wetland plants (Armstrong 1979). Armstrong also attempted to explain the diffusive properties of O2 in the roots of wetland plants, but recognized that some mass flow caused by temperature and pressure differentials between below and aboveground portions of plants must occur in aerenchyma. .Until 1980, models of gas transport were based on the assumption that the gas phase in the lacunae is stationary and that individual gasses diffuse along concentration gradients (Hutchinson 1957). In 1980, Dacey confirmed that the sun supplies the energy required for a thermo-osmotic pump that drives gas movement in lacunae of water lilies. He found that the flow is not static, and that gas exchange can occur at linear rates of up to 50 cm min"1. He described gas flow from young leaves down petioles, through root tissue, and into older leaf petioles. With the use of tracer studies, gas flow within the petioles of leaves was found to be a linear function of the observed pressure gradient. This observation is in accordance with Darcy's law for flow through porous media. Temperature differences between leaf and atmosphere drive pressurization through two independent diffusive processes, thermal transpiration and hygrometric pressure .(Dacey 1980). Both operate in sunlit conditions as 18 temperatures and vapor pressure gradients (of all gasses including water vapor) are maintained. Variations in gas exchange mechanisms are based on structural differences among plant species, and historical perspectives on the operation of this thermo-osmotic phenomena are discussed by Mevi-Schutz and Grosse (1988). Gas exchange within lacunar spaces of plants not only occurs between the shoot and root, but also from the root to shoot. Several researchers have measured gas transfer in aerenchyma from wetland soils. Dacey and Klug (1979) showed 37% of the methane generated in wetland sediments was removed through the aerenchyma of Nuphar luteum. Reddy et al. (1989) found that a significant portion of denitrified gas formed in the rhizosphere escaped through plant tissues. The role of aquatic plant lacunae has been recognized in promoting carbon cycling (Sebacher et al. 1985) during periods of active plant growth and dormancy (Brix 1989). Sand-Jensen et al. (1982) suggested that CO2 transport is more important than O2 transport. Constable et al. (1992) also stressed the importance of an internal CO2 source in the roots of Typha. One of the more interesting observations made about transport in the internal lacunae of plants is that root length is a function of effective root porosity and length of diffusive O2 path per unit volume of respiring root" (Armstrong 1979) . Armstrong (1967b) stated that lateral roots are primary to O2 diffusion and discussed the theoretical relationship between O2 diffusion rate and root size. 19 Rhizosphere Oxygenation An oxidized substrate is, in most cases ̂ a favorable environment for root growth (van Raalte 1944, Armstrong and Boatman 1967, Teal and Kanwisher 1966, Wium-Anderson and Anderson 1972). Van Raalte (1944), and Teal and Kanwisher (1966) noted that O2 release into a reducing medium occurred when the internal supply of O2 exceeded respiratory demand. Teal and Kanwisher (1966) found comparable ferric hydroxide deposits on the foots of Spartina and suggested that these are produced by O2 release from roots to sediment. Most wetland species modify the soil environment by transporting and releasing O2 into the surrounding substrate (Armstrong 1964, Bristow 1975, Armstrong 1967a). Sarid-Jensen et al. (1982) tested eight plants from the littoral zone of oligotrophia lakes and found evidence of O2 exchange in these lake sediments. Submersed vascular plants characteristic of oligotrophic communities release O2 while mosses lack the root and lacunar systems of vascular macrophytes that facilitate O2 release to sediments (Jaynes and Carpenter 1986). Conversely, Bedford et al. (1991) argued that given microbial, soil oxidative, and root respiratory demands, plants are not capable of oxygenating sediments. Armstrong (1964), used O2 data and redox measurements to explain differences in plant O2 transport in roots. Armstrong (1967a) and McKee et al. (1988) stated that diffusion rate in 20 wetland plant species is species-specific and does not vary within a particular species. Morris and Dacey (1984) and Dacey (1980) also have done extensive work on O2 transport rates in plant root systems. Rate of NH4+ uptake and root respiration are especially sensitive to O2 concentration in the rhizosphere (Morris and Dacey 1984). Rates of gas transfer may depend on size or biomass of the plant and volume of the aerenchyma. Radial O2 loss from roots increases with root porosity and decreases with root length (Armstrong 1972), varying both seasonally (Armstrong et al. 1990) and diurnalIy (Sand-Jensen et al. 1982, Sand-Jensen and Prahl 1982). The effects of soils on plant adaptability and rhizosphere oxygenation appear to be minimal. Leakage of O2 by Typha was found to be site dependant and varied by season. Rhizosphere oxidation did not appear to be related to Eh (Crowder and Macfie 1985) and pH regimes (Crowder and Macfie 1985, Macfie and Crowder 1987). McKee et al. (1988) explained, that the occurrence of aerial roots of Avicennia and Rhizophora, not the reverse, influenced soil oxygenation. Differences in the oxygenated rhizosphere in relation to root size supports arguments citing genetic variation in the O2 transport mechanism operating in wetland soils. The oxygenated rhizosphere is a function of O2 concentration in the root and the root radius, thus favoring large roots (Armstrong 1979). Armstrong (1979), however, showed that roots with a smaller radius had zones of oxygenation similar ' 21 to those of larger roots and suggested that wetland plants with smaller roots may have some competitive advantage. Thin roots have a better O2 supply to,_all root cells than thick roots and thus need less root porosity to provide a transport path for O2. The highest O2 transport rates from aerial plant tissue into the rhizosphere also are associated with plants having small root mass. Older plants have large root systems and higher O2 transport rates (Moorhead and Reddy 1988). Armstrong et al. (1990) demonstrated that greater root numbers, not greater O2 demand, account for sediment oxygenation. Nitrogen Dynamics in Saturated Soils Nitrogen additions to wetland soils consist of N fixation, atmospheric deposition, and biomass decay. . Plant and microbial assimilation (mineralization) ,. NH3 volatilization, nitrification, and denitrification, are other transformation pathways that modify and remove N in wetland treatment systems. Nitrate, the thermodynamically stable form of N in terrestrial oxygenated aqueous systems (National Research Council 1972), may be biochemically assimilated from water by growing plants or may be converted to gaseous N in anoxic situations. In surface waters, main sinks for loss of inorganic N include NH3 volatilization, and algal uptake (Wittgren and Sundblad 1990). Ammonia is the preferred form of N for algae (Richardson and Davies 1987) . Natural 22 concentrations of NO3 and NH3 in soils vary greatly with season and soil aeration. Ammonium diffusion, however, proceeds more slowly than NO3 diffusion because of adsorption to soil particles (Lorenz and Biesboer 1987). Removal of N in wetland wastewater treatment systems requires an understanding of NH3 volatilization, plant assimilation, and nitrification-denitrification processes. Ammonia volatilization is seasonal and inconsistent (StowelI et al. 1981), varies diurnally, and is affected by temperature, pH (Beaucamp et al. 1982, Hoff et al. 1981) and wind speed (Hoff et al. 1981) . At high pH values, bulk loss of NH4+-N occurs as NH3 is volatilized from the water surface. This volatilization is a function of the partial pressure of NH3 and the surface to volume ratio of the water body (Veen 1988). At a pH below 7.5 (Rogers et al. 1991, Wittgren and Sundblad 1990) , volatilization becomes less important. Ammonium is stable at low pH values and at the normal pH of sewage (Davies and Hart 1990); Stengel et al. (1987) and Lorenz and Biesboer (1987) inhibited microbial activity with acetylene and measured N2O, as an indicator of denitrification. Additional reviews of nitrification-denitrification reactions in wetland systems are given by Reddy et al. (1989) , Valiela and Teal (1979) , and Veen (1988). The rates of these reactions are controlled by redox potential, pH, moisture, carbon source, and temperature (Denny 1972, Barnes and BlisS 1983). Denitrification rates 23 are also related to xnicrobially modified soil O2 status (Obenhuber and Lowrance 1991) and carbon content of soil (Burford and Bremner 1975, Bowman and Focht .1974, Firestone 1982). Obenhuber and Lowrance (1991) stated that microbial denitrification can reduce NO3 concentrations provided adequate energy sources and appropriate environmental conditions exist. Adding carbon to an aquifer increases N2O production or denitrification. Denitrification rates may be limited by carbon source, but not universally by soil type (Lorenz and Biesboer 1987). Differences exist in the literature between the importance of denitrification in wetland systems and the role of plant assimilation as the primary N removal mechanism. Most researchers contend that the nitrification- denitrification reactions are key to N removal (Wood 1990, DeBusk et al. 1983, Swindell and Jackson 1990, Morris 1991, Wittgren and Sundblad 1990)• Breen (1990), however, suggested that plant uptake accounted for 80% of N removal. Kadlec (1987b) stated that plants function primarily as uptake and transfer conduits unless there is harvest of biomass. Rogers et al. (1991) concurred; their study showed plant nutrient uptake as the dominant removal mechanism. Heterotrophic fixation of N2, they argued, could also mask denitrification. However, they failed to note research by Oaks and Hirel (1985) and Buresh et al. (1980) that indicated nitrification and bacterial (N-fixing) nitrogenase was inhibited by NH4*. Reddy 24 et al. (1982) found that 50% of N2 loss in simulated wetland systems was through means other than plant uptake. Denitrification rates in natural wetlands can exceed natural N inputs by a factor of three (Zak and Grigal 1991). Weber and Tchobanoglous (1986) stated that when plant density is near a maximum, plant uptake of N is assumed to be minimal and nitrification-denitrification is thought to be the major pathway of N loss. Veen (1988) evaluated the change in root activity with respect to a reduction of root respiration in terrestrial plants known to be unequally sensitive to O2 stress. He found that lower root O2 concentrations reduced NO3 uptake. These findings conflict with those of Lambers et al. (1978) who showed that nitrate reductase activity enhanced in roots under anaerobic conditions. Other work by Guyot and Prioul (1985) indicated that plant tolerance to O2 deficiency improves with additional NO3.. Changes in soil structure, hydrology, plant community composition and microbial activity affect the speciation and dynamics of nutrients present in wetland soils. Nitrate will not adsorb in soils below pH 5.5 while NH4+ is strongly held on the cation exchange at this pH. In undisturbed wetlands with mineral soils the pH equilibrates toward a neutral value. The pH of more acidic soils increases in flooded conditions and enhances the reduction of Fe (III) to Fe (II) . Conversely, the pH of alkaline soils decreases in flooded 25 conditions with an accumulation of respiratory CO2 (Ponnamperuma 1972). Further modification Of pH is caused by microbial reactions in the sediments when nitrifying bacteria produce hydrogen ions that lower pH (Davies and Hart 1990). Acidification blocks the N cycle by inhibiting nitrification and leads to an accumulation of NH4* (Rudd et al. 1988). Veen (1988) attributed an increase in nutrient solution alkalinity to reduced root O2 concentration. Bicarbonate production of the root system is a reflection of the NO3 reduction rate of the root system (CO2 generation) . A study by Jaynes and 'Carpenter (1986), using pots planted with mosses and vascular plants, showed that Eh and pH did not vary by sediment type but rather by the influence of the plants. Chen and Patrick (1981) also found that additions of carbon additions reduced Eh and enhanced NO3 reduction rate (denitrification). 26 METHODS Experimental Design Forty-five microcosms (Figure I) were planted with four wetland plant species - Nebraska sedge (Carex nebraskensis Dewey), beaked sedge (Carex rostrata Stokes), broad-leaved cattail (Typha latifolia L.) , and panicled bullrush (Scirpus microcarpus Presl.), and placed in a greenhouse. These species are abbreviated throughout the tables as Cane, Caro, Tyla, and Semi, respectively. Unplanted microcosms are abbreviated as Unpl. Placement of the microcosms within the greenhouse was randomized. The experiment was designed with two levels of repeated measures. The experimental units represent a five (four plants with unplanted controls) by three (control application plus two effluent concentrations) by three (replicate) factorial experiment. The units were treated with either of two concentrations of liquid dairy effluent or a control treatment of water. Each plant species, control, and treatment level (concentration) was replicated in triplicate. The first level of repeated measures consisted of the soil water at three soil depths within each microcosm. The second level of repeated measure occurred over time as each microcosm was measured at two-week intervals. Statistical Analyses The experimental design entailed two levels of repeated measure - depth in the microcosm and changes in effluent 27 parameters over time. All statistical analyses were conducted at p < 0.05. Statistical contrasts of interactions and main effects between plant species, depth and effluent concentration within the treatment cycle were made for changes in TKN, COD, pH and solution NO3* using the PROC GLM procedure in SAS (SAS Inst 1987) . Plant biomass, root porosity, and plant N content were statistically analyzed using the ANOVA procedure in NCSS (Hintze 1990). Comparison of treatment means included ranking of means and Duncan's multiple range test (Mize and Shultz 1985, Ott 1984). PVC pipe Soil level Polyethylene tube with ceramic cup X #4 screw Plexiglass Support brace Polywoven fabric «8'.' T-- ■ ■ Figure I. Side view schematic of microcosm experimental unit construction, including placement of porous ceramic cups. 28 Microcosms The microcosms used in this study were constructed using PVC irrigation pipe (38 cm ID) cut to 60 cm lengths and split laterally to form half-round columns (Figure I). The half- rounds were fitted with a 38 by 60 cm sheet of 0.32 cm plexiglass attached to the PVC pipe with an industrial silicon sealant and screws spaced along each edge. Three holes (1.0 cm diameter) were drilled into the sides of each microcosm at 15 cm increments to allow for placement of lengths of 0.95 cm polypropylene tubing. Tubing was cut to 23 cm lengths and attached to round end ceramic cups (I Bar, High Flow, Soil Moisture Equipment Corp.). The ceramic cups were aligned in the microcosms for center placement in the substrate. The base of each microcosm consisted of a piece of poly-woven ground cloth covering the open end and attached using silicon sealant. This base was not intended for structural support of microcosm contents, but served as a barrier that allowed for exchange of air and water from the microcosms. A standard greenhouse tray was placed to catch water beneath each microcosm. The microcosm units were reinforced with two wood and wire braces (Figure I) for structural support of the plexiglass. The braces were attached at the base and center of each microcosm and tightened using double-threaded rod and butterfly nuts. . A double layer of ground cloth was attached to the plexiglass face of the microcosms to prevent light 29 entry to the rhizosphere. This material could be removed for visual inspection of root growth in the soil column. The ceramic cups were washed with I N HCl solution and then rinsed with distilled water until the pH and specific conductance of the rinse water equaled that of the distilled water influent (Creasey and Dreiss 1988). Cups were stored in a covered box to prevent dust accumulation prior to attachment of the polyethylene tubing. Tubing and ceramics were attached using an epoxy from Soil Moisture Equipment Corp. The epoxy was allowed to dry and cure for a period of 24 hours. Microcosms were assigned numbers and randomly located in the greenhouse (Figure 2). Soil Collection and Preparation Soils used in this study were collected from a natural wetland area located in Gallatin County, Montana (N.W. 1/4 Sec. 25 T. I N R. 3 E, P.M) . This site was an open slough centered in a pasture receiving seasonal use by cattle. Vegetation in the slough included softstem bullrush (Scirpus americanus Pers.), broad-leaved cattail (Typha latifolia L.), plainleaf willow (Salix planifolia Pursh), Nebraska sedge (Carex nebraskensis Dewey), rabbitfoot grass (Polypogon monspeliensis L.), and smooth brome (Bromus inermis Leysser). Soil was collected on June 19 and July 28, 1992. The soil was kept damp in transit and placed in cold storage prior to placement into the microcosms. 39 Plant Q Carex nebraskensis E3 Typha Ia tifo lia [771 U nplanted \ I Scirpus m icrocarpus I I Carex rostrata Treatment 0 70% Dilution 0 30% Dilution B Control Figure 2. Placement and randomization of microcosms in greenhouse'. 31 Sampling of the soil for baseline soil parameters was made during the first soil collection period (June 21, 1992). Five soil cores were taken at 0-15 cm and 15-30 cm soil depths. Samples were air dried ground, split and sieved to < 0.2 mm and analyzed for percent organic matter, cation exchange capacity (CEC), alkalinity, and N . concentration. Additional unground samples were used in the analysis of particle size distribution. Organic matter was determined by combustion of a known weight of dry sample at 4500C for 16 hours (Ball 1964). Nitrogen analysis used methods specified by Bremner and Mulvaney (1982). Particle size analysis followed procedures outlined by Gee and Bauder (1986). Cation exchange capacity and alkalinity procedures followed Rhoades (1982) . Refrigerated soils were homogenated by hand to obtain uniform soil consistency and volume in each microcosm. Soils were manually placed in the wetland units on August 3. A small volume (64 cm3) of autoclaved sand was placed around each ceramic pore cup to facilitate removal of soil water samples under negative applied pressure. Soil placement continued until the soil level was 10 cm below the top of the microcosms. A 5 cm head of distilled water was then applied to the top of the soil column to maintain saturation and a reduced state of the soil constituents. Saturation through continuous application of water was maintained throughout the experimental procedures. 32 Twenty-seven platinum electrodes were constructed and calibrated for measurement of soil Eh (Mueller et al. 1985, Faulkner et al. 1989). These electrodes were standardized against an Orion combination redox electrode in a standard ferrous-ferric solution (Light 1972). Electrodes were placed in several wetland microcosms to evaluate differences between planted (C. rostrata) and unplanted units at the three soil depths under control and high concentrate treatments (Table I). Four measurements of millivolt potential were made at one week intervals beginning November 1992. Measurements were corrected to a standard hydrogen electrode (SHE) using the correction value of +200 mV (Orion 1992). Oxidation-Reduction Potential Table I. Experimental design for redox measurement in greenhouse microcosms. Treatment Unplanted controls Carex rostrata Unit Depth (cm) Unit Depth (cm) 15 30 45 15 30 45 9 - - - 37 X - X Control 25 X X X 26 X X - 21 X X X 17 X X X 22 X X X 40 - X • X Low Cone. 14 X X - 4 X X X 42 - X X CM X X - X indicates placement of a platinum electrode at specified depth in unit - indicates no electrode placement at specified depth in unit 33 Plant Collection and Preparation The plant species used in this study comprise a major component of the wetland species composition in north- temperate regions of North America. Plant nomenclature is after Hitchcock and Cronguist (1981). A description of the plant species used, their distribution, and gross morphological characteristics is presented in Appendix A. Whole, intact plants were collected from a wetland at the Nixon bridge of the Gallatin River in Gallatin County, Montana (NW 1/4 S .35 T.2 N R.3 E. P.M.) . Collection of the plants was made on August 21, 1992. Plants were immediately pruned and refrigerated for I week prior to planting. Preparation for planting included tagging all plants with colored telephone wire, and recording fresh weight biomass of each plant. Plants were divided by species into groups with equivalent weights (TOO ± 10 g). Plants were placed in randomly selected microcosms on August 28, 1992. Root crowns were set to their previous soil levels and roots were extended to full length in the microcosms. Roots were positioned close to the upper ceramic cup (15 cm depth) . A constant head (5 cm) of water was maintained after planting. Because the plants were collected in the fall, they were allowed to grow for a period of one month while temperatures in the greenhouse were increased in 5°C increments from 55°C night and day temperature to a 60 °C night and 75 °C day temperature. Supplemental lighting via metal halide lamps 34 (1000 N «111/3 Sylvannia M1000\c\u) suspended 1.5 m above the plant canopy was increased in half hour increments each week to an 8 hour dark cycle over the same period. Effluent Collection and Preparation Fresh liquid manure from a herd of dairy cattle in Gallatin County, Montana was collected one day prior to each effluent application. The dairy cattle diet consisted of alfalfa and barley with some supplementation of trace nutrients and salts. No detergents, antibiotics, or iodinated teat washings were present in the effluent material. Manure (75.7 I) was collected from manure piles freshly scraped from feeding pens. Materials were collected from the entire manure pile at random locations. Manure effluents for the application were immediately prepared by adding 15.1 I of water to every 37.9 I of original manure material. The slurry was then placed in rinsed polyethylene bags, tied, and suspended from a brace for filtration. The filtrate was collected in 18.9 I buckets over a period of. several hours, sealed and cooled overnight. Two concentrations of the manures were made by diluting with distilled water. The two concentrations were 30% (low concentration) and 70% (high concentration) of the original wastewater filtrate. The dilutions were stirred, sub-sampled for laboratory analysis and then applied to the microcosms in 2 I volumes. These applications were made four times on a 35 biweekly basis (Table 2). Placement of the liquid waste was made to avoid stirring of the soil surface by the liquid additions. No direct application of liquid manures was made on the plant surfaces. Any remaining volume in each microcosm was filled with distilled water and a constant water level was maintained between applications. • V . Table 2. Waste application schedule for wetland wastewater treatment study (1992). Period Soil Water Sampling Effluent Application Date Date I Sep. 18 Sep. 28 2 ' Oct. 3. Oct. 13 3 Oct. 18 Oct. 27 4 Nov. I Nov. 10 5 Nov. 15 Sampling and Analysis Effluents Sampling of manure effluents followed the schedule outlined in Table 2. Table 3 is an outline of the chemical analyses performed on both influent and effluent samples. The first two effluents Were sampled at the time of preparation, refrigerated, and sent to the analytical laboratory within 24 hours. Samples were analyzed for NH3-N and NO3 -N. Sub­ samples also were analyzed separately for NH3-N and total Kjeldahl nitrogen (TKN), COD, and pH. Analysis of the third and fourth effluent concentrates consisted of all parameters 36 investigated in the first two sampling periods’ effluents except for the analysis of NO3 ^N. A companion study was utilized to determine the amount of N removal occurring independent of the wetland units. Open containers of effluent were placed in a ventilated hood for two weeks (the period between effluent applications in the Table 3. Sample analyses conducted for influent and effluent water parameters. Parameter Period(s) Reference I Influent TKN1 I - A- APHA 1992 NH3-N 1,2 EPA 1983 (350.2) NH/-N 1-4 APHA 1992 NO3'-N 1,2 EPA 1983 (353.3) COD2 1-4 - APHA, Hach 1992 PH3 1-4 Orion 1992 Effluent TKN 1-5 APHA 1992 NO3"-N 1,2 APHA 1992 N03"-N 3-5 APHA 1992; Willis 1980 NO2"-N I APHA 1992 COD 2,3,5 APHA, Hach 1992 - PH 1-5 Orion 1992 1 - Keltec Auto Analyzer, Perstop Analytical, Inc., Model 1020 2 - Open Reflux Method by block digestion and colorimetry, Hach Corporation 3 - Orion Research pH meter 25OA with pH combination electrode 91-57 microcosms) and subjected to similar temperatures and air flow environments found in the greenhouse. Evaporated volumes of water were replaced. Measurements of pH, Eh, total and NH3-N were made before and after the two week period. Changes in pH, Eh, and total and NH3-N concentration of the liquid dairy effluents were determined. 37 Microcosm Soil Water Sampling Soil water from each microcosm at each of the three soil depths was sampled prior to tlje application of the manure effluents and two weeks after the final manure application (Table 2). Soil water was collected from the ceramic cups by applying a vacuum to the connective ports (Figure I) and drawing aliquots of soil water, into containers. Seven milliliters of each sample were preserved with 0.5 ml 6. N H2SO4. These samples were refrigerated until analyzed for TKN and COD. Five of the seven ml were used for TKN. The remaining 2 ml were used in the analysis of COD. The non- .acidified remains were sent to the analytical laboratory (first two' samplings) for measurement of NO3 -N and NO2 -N using anion chromatography. Soil pore water samples from the latter three sampling periods were analyzed for NO3 -N using a colorimetric technique (Willis 1980) . Nitrite was eliminated from the analytical program since it was not detected above 0.05 mg l"1 in any of the original 135 samples submitted in the first sampling period. Plant Sampling Plant characteristics measured in this study included fresh and dry weight of roots, rhizomes, and shoots; root porosity; and protein content of below and aboveground biomass. Root masses removed from the units were rinsed free of soil. A fresh sub-sample was removed from each microcosm 38 ■ root mass, sealed and refrigerated for root porosity evaluation. Root porosity determinations were made using a gravimetric test (Jensen et al. 1969, Burdick 1989, and Van Noordwijk and Brouwer 1988). All biomass was gravimetrically measured and corrected for percent moisture content. Protein content of ground plant samples was determined by TKN analyses and corrected for dry weight. Protein analyses were conducted on duplicate air dried samples. Soil Sampling The microcosms were disassembled at the termination of the experiments. At this time, soils were sampled at the surface of each unit and at the three soil depths adjacent to the ceramic cups. Samples were air dried, ground, and passed through a 2 mm screen. The samples were then analyzed for TKN and NO3.-N. Quality Control/Quality Assurance -To improve the precision and accuracy of data collected in this study, a quality control/quality assurance program was implemented. Measurements of TKN, nitrate, and COD in manures and soil water samples included a 1:20 field replicate sampling (blind field replicates for samples sent to the respective labs). A cross contamination blank and a bottle blank were included for each set of soil water samples. Accuracy of both the TKN and NO3"-N concentration was 39 determined using EPA nutrient standards in the final two sample sets (Periods 4 and 5) . Data precision for COD analysis was based on regressions of known standards (Appendix B) . 40 RESULTS Interpretations of statistical outputs were based on the main effect and interaction differences observed initially in the microcosms compared to observed differences in the final measurement period. Results from repeated measures analyses were used to interpret the overall effects of pollutant input to the wetland microcosms. The statistical outputs from ANOVA are provided in Appendix C. Table grand means (X) represent column or rowwise averages across concentration, depth, or plant treatment for the combined samples (n). General Observations Plants grew well in the greenhouse. Root growth was especially concentrated in the upper 20 cm of the soil column. Typha latifolia grew slower than the other plant species. Initially, the microcosms showed high hydraulic conductivity that apparently decreased over the course of the experiment. Water continued to flow out of the sampling ports and from the bases of the microcosms throughout the study. Hydraulic conductivity of the media, overall volume of water applied, hydraulic residence time, and evapotranspiration were not quantified. Carex rostrata and S. microcarpus showed high vigor in all effluent treatments through the treatment cycle. The vigor of C. nebraskensis declined in the latter sampling periods as evidenced by chlorosis and some stem mortality. Prior to wastewater application, all units contained a 41 number of macroinvertebrates and benthic organisms. The origin of these organisms was presumed to be dormant life forms in the collected soil. Several species of aquatic annelids were observed to feed and move through the upper 10 cm of soil. Cladocera species were seen to feed on algae that grew in the water column. Several forms of algae were observed in both planted and unplanted units. Macro­ invertebrates and benthic organisms in units receiving high concentrated wastes were killed following wastewater additions. The same organisms thrived in the control and the low concentrate microcosms. A band of black coloring formed approximately I cm below the soil surface. This phenomenon was presumed to be a function of symbiotic microbial growth between sulfate- reducing bacteria and cyanobacteria under high nutrient concentrations (Hodges 1992). System Parameters The basic characteristics of the microcosms prior to applications of liquid dairy wastes are presented in Table 4. Baseline soil data with the exception of NO3"-N analyses were determined from samples taken at the time of soil collection. The analyses were also performed on soil samples taken at the end of the experiment. Soil was classified as a Typic haplaquoll (Soil Survey Staff 1987) with a sandy clay texture (Gee and Bauder 198 6) . Analyses of soil water samples were conducted prior to pollutant application. 42 Microcosm Oxidation-Reduction Potential Oxidation-reduction potentials (ORP) by depth, concentration level, and presence or absence of C. rostrata are given in Table 5. The millivolt potentials are corrected to the standard hydrogen electrode (+200 mV at 25°C). No .,,.,significant differences were found within or among main effects by plant species, concentration, or depth. Table 4. Microcosm characteristics and baseline soil and interstitial water parameters. Characteristic n. x ± S.E. EC (mS cm-1) 45 1.44 ± 0.20 , PH 129 7.36 ± 0.20 CEC (meq 100 g"1) 2. 29.65 ± 1.15 ' Soil TKN (%)1 20 0.527 ± 0.138 Soil TKN (%)2 60 0.476 ± 0.102 Soil NO3--N (mg kg"1) 60 15.29 ± 27.20 Soil OM (%) 10 9.90 ± 2.23 Soil water TKN (mg I"1) 131 1.961 ± 0.702 Soil water NO3--N (mg I"1) 133 0.436 ± 0.268 Soil alkalinity (mg I"1 as CaCO3) 4 768.13 ± 68.13 Samples made from field collections 2 - Samples taken following disassembly of microcosms Plant Characteristics Plant Biomass The results of the analyses o f . plant biomass are presented in Tables 6 through 14. Plant root biomass of the four species by three effluent treatments is shown in Table 6. Table 5. Microcosm oxidation-reduction potential (mV) by depth, concentration and presence or absence of Carex rostrata (Caro). Depth (cm) (n=4) Concentration (n=6) 15 30 45 Control High Cone. Unpl1 -77.6 ± 51.2 -128.3 ± 54.0 -154.0 ± 69.5 -92.5 ± 53.5 -108.9 ± 48.6 Caro -111.3 ± 53.0 -29.25 ± 89.4 -4.91 ± 35.0 -135.2 ± 21.0 -82.0 ± 40.7 X -96.3 ± 35.4 -128.7 ± 32.3 -79.4 ± 45.7 -111.8 + 29.9 -95.4 ± 30.7 1 - Unplanted microcosms Table 6. Effect of effluent concentration i Table 24. Mean (±S.E.) microcosm NO3 -N (mg I 1) by concentration for five sampling periods (n=4 5). Cone. Sampling Period I* 1 2 3 4 5 High 0.45 ± 0.27 a2 0.70 ± 0.35 a 0.36 ± 0.22 ab 0.18 ± 0.11 a 0.25 ± 0.20 a Low 0.46 ± 0.24 a 0.61 ± 0.33 a 0.51 ± 0.59 a 0.22 ± 0.20 a 0.30 ± 0.27 a Cont. 0.44 ± 0.26 a 0.70 ± 0.21 a 0.33 ± 0.18 b 0.21 ± 0.12 a 0.34 ± 0.18 a 1 - I = Sep. 18, 2 Oct. 3, 3 = Oct. 18, 4 ■= Nov. I, 5 = Nov. 15. - Means in columns followed by the same letter are not significantly different (p < 0.05). d\o Table 25. Me a n (± S.E.) NO3 -N (mg I 1) by soil depth for five sampling periods (n=45) . Sampling Period Depth I1 2 3 4 5 15 cm 0.45 ± 0.27 a2 0.55 ± 0.37 b 0.39 ± 0.23 a 0.20 ± 0.18 a 0.25 ± 0.20 a 30 cm 0.48 ± 0.23 a 0.73 ± 0.29 a 0.41 ± 0.57 a 0.21 £ 0.12 a 0.33 ± 0.19 a 45 cm_________0.42 ± 0.26 a_________0.74 ± 0.19 a 0.40 ± 0.26 a 0.20 ± 0.14 a 0.32 ± 0.27 a 1 - I = Sep. 18, 2 = Oct. 3, 3 = Oct. 18, 4 = Nov. I, 5 = Nov. 15. 2 - Means in columns followed by the same letter are not significantly different (p < 0.05). Table 26. Mean (± S.E.) microcosm NO3 -N (mg I 1J by plant species for five sampling periods (n=36). Sampling Period Plant I1 2 3 4 5 Unpl 0.36 ± 0.29 a2 0.72 ± 0.23 a 0.39 ± 0.25 a 0.26 ± 0.25 a 0.39 ± 0.32 a Cane 0.40 ± 0.26 ab 0.70 ±0.27 a 0.34 ± 0.18 a 0.18 ± 0.09 b 0.29 ± 0.19 ab Scmi 0.46 ± 0.28 ab 0.60 ± 0.41 a 0.38 ± 0.22 a 0.17 ± 0.11 b 0.26 ± 0.15 b Cane 0.49 ± 0.23 ab 0.65 ±0.27 a 0.40 ± 0.28 a 0.18 ± 0.12 b 0.25 ± 0.19 b Tyla 0.54 ± 0.19 b 0.69 ± 0.31 a 0.50 ± 0.72 a 0.22 ± 0.19 ;ib 0.30 ± 0.21 ab 1 - I = Sep. 2 - Means in 18, 2 = Oct. 3, 3 = Oct. columns followed by the , 18, 4 = Nov. I, same letter are 5 = Nov. 15. not significantly different (p < 0.05) . Table 27. Mean (± S . E . ) (n=36). microcosm TKN (mg I'1) by concentration for five sampling periods Sampling Period * Cone. I1 2 3 4 5 High 1.97 ± 0.65 a2 2.33 ± 0.66 ab 3.47 ±0.85 a 3.72 ± 1.95 a 4.34 ± 2.47 a Low 1.98 ± 0.83 a 2.48 ± 0.48 a 3.25 ± 0.63 ab 3.16 ±1.07 b 3.03 ± 0.83 b Cont. 1.98 ± 0.67 a 2.14 ± 0.60 b 2.97 ± 0.67 b 2.57 ± 0.95 C 1.82 ± 0.90 c 1 - I = Sep. 18, 2 = Oct. 3, 3 = Oct. 18, 4 = Nov. I, 5 = Nov. 15. Means in columns followed by the same letter are not significantly different (p < 0.05). H 2 62 higher concentrations of total soil water N over time. Low concentrate units showed an initial increase in soil water N that leveled off in the final two sampling periods. Control units showed a slight decrease in TKN values at the fourth and fifth sampling periods. Table 28 is a presentation of TKN data against depth in the soil profile. Soil water TKN levels Were significantly different in the upper soil zone from those at 30 and 45 cm. At lower soil depths, TKN levels did not differ from controls and those of the measurements made during the initial sampling period. Total N level as a function of depth in the soil profile across plant and concentration treatments is presented in Figure 4. The control treatment showed a lower TKN level at the surface 15 cm, while the low concentrate TKN level was not significantly different among the three depths. High concentrate TKN levels were markedly higher at the 15 cm depth than the 30 and 45 cm depths. Total N levels in the interstitial water samples did not significantly differ at the 30. and 45 cm depths with respect to effluent treatments. A graphic representation of the changes in TKN over time in the 15 cm soil zone is presented in Figure 5. The slope of the relationship between TKN and depth in low concentrate microcosms at the end of the experiment was negative. A positive slope was present in the corresponding curve for the units receiving high concentrate. Total N levels in unplanted units were only different from units planted with T. Iatlfolia • ' ■ .1 J-Table 28. Mean (± S.E.) microcosm TKN (mg I ) by soil depth for five sampling periods (n=36). Depth Sampling Period I1 2 3 4 5 15 cm 1.49 ± 0.47 a2 2.03 ± 0.54 a 3.23 ± 0.94 a 3.87 ± 2.10 a 3.71 ± 2.70 a 30 cm 2.16 ± 0.56 a 2.41 ± 0.51 a 3.44 ± 0.67 ab 2.93 ± 0.78 b 2.68 ± 1.27 b 45 cm 2.33 ± 0.81 b 2.54 ± 0.63 b 2.97 ± 0.44 b 2.58 ± 0.64 b 2.74 ± 0.95 b 1 - I = Sep. 18, 2 = Oct. 3 , 3 = Oct.. 18, 4 = Nov. I, 5 = Nov. 15. 2 - Means in columns followed by the same letter are not significantly different (p < 0.05). ONw Table 29. Mean (± S.E.) microcosm TKN (n=36). (mg I'1) by plant species for five sampling periods Plant Sampling Period I1 2 3 4 5 Unpl 2.21 ± 0.68 a2 2.46 ± 0.56 a 3.45 ± 0.48 a 3.40 ± 1.53 a 3.30 ± 1.53 a Cane 2.01 ± 0.78 a 2.46 ± 0.52 a 3.30 ± 0.89 a 3.41 ± 1.82 a 3.40 ± 2.33 a Scmi 1.95 ± 0.80 a 2.39 ± 0.58 a 3.18 ± 0.94 a 3.29 ± 1.51 ab 3.09 ± 2.03 ab Cane 1.88 ± 0.76 a 2.22 ± 0.70 ab 3.11 ± 0.70 a 3.11 ± 1.44 ab 2.99 ± 2.22 ab Tyla 1.84 ± 0.54 a 2.06 ± 0.56 b 3.10 ± 0.61 a 2.57 ± 0.78 b 2.51 ± 1.15 b 1 - I = Sep. 18, 2 = Oct. 3 , 3 = Oct. 18, 4 = Nov. I,, 5 = Nov. 15. - Means in columns followed by the same letter are not significantly different (p < 0.05).2 64 (Table 29) at the end of the study. Unplanted units and units planted with C. nebraskensis had higher relative concentrations of total N at each sampling period. Units planted with C. nebraskensis had significantly higher total N than the other three plant species. Soil Water TKN (mg M ) Eo CL