Simulation of Phosphorus Transport in Vegetative Filter Strips. Dowon Lee. Dissertation submitted to the Faculty of the

Transcription

1 Simulation of Phosphorus Transport in Vegetative Filter Strips by Dowon Lee Dissertation submitted to the Faulty of the Virginia Polytehni Institute and State University in partial fulfillment of the requirements for the degree of Dotor of Philosophy in Environmental Siene and Engineering APPROVED: T. A. Dillaha, Co-Chairman ;JI. H. Sherrard, Co -Chairman J. A. Burger C. W. Randall C: J. R. Webster January 9, 1987 Blaksburg, Virginia

2 Simulation of Phosphorus Transport in Vegetative Filter Strips by Dowon Lee T. A. Dillaha, Co-Chairman J. H. Sherrard, Co-Chairman Environmental Siene and Engineering (ABSTRACT) This study investigated the effetiveness of vegetative filter strips (VFS) in removing phosphorus from surfae runoff. Dissolved and partiulate nutrients were treated separately due to differing transport and removal mehanisms. Nutrient transport in VFS appeared to be a funtion of runoff rate, onentration and size distribution of suspended solids, and biologial fators that influene hydrologi and hemial proesses in filter strips. Three sets of experimental field plots were onstruted to simulate VFS. Eah set onsisted of three plots ontaining sediment and nutrient soure areas and., 4.6, or 9.1 m grass filter strips. Artifiial rainfall was applied to the plots, and surfae runoff, soil, and plant material samples were olleted and physially and hemially analyzed. The VFS redued surfae runoff, suspended solids, and phosphorus losses. Most removal of sediment and phosphorus was aomplished in the first few meters of the VFS. The filter strips did not remove phosphorus as effetively as sediment, due to their ineffetiveness for filtering dissolved phosphorus and sediment-bound phosphorus assoiated with fine partiles. The VFS often inreased orthophosphorus losses in surfae runoff. Laboratory bath experiments of phosph~rus desorption reation suggested that plant residues, living plant anopy, and soil omponents of the strips ould release dissolved phosphorus to surfae runoff. A modified Elovih equation and a diffusion-ontrol model were used to desribe the phosphorus release from the plant and soil materials. A omputer model, GRAPH, was developed to simulate phosphorus transport in VFS by inorporating phosphorus transport submodels into the VFS model in SEDIMOT II, a stormwater and sediment transport model. The model onsiders the effets of advetion

3 proesses, infiltration, biologial uptake, phosphorus desorption from the soil surfae to runoff, the adsorption of dissolved phosphorus to suspended solids in runoff, and the effets of dynami hanges in the sediment size distribution on hemial transport. GRAPH was verified using the results of the physial plot simulations. The model's preditions and observed phosphorus transport ompared favorably. Sensitivity analysis suggested that sediment and phosphorus removal was sensitive to the input parameters in the order: filter length and width, grass spaing, and filter slope and surfae roughness. Inreased filter width and length and aboveground biomass inreased orthophosphorus loss from VFS.

4 Aknowledgements I would like to thank Professors Theo A. Dillaha and Joseph H. Sherrard for their kind suggestions and guidane. I am, in partiular, indebted to them for introduing me to the dissertation topi. I am also grateful to Professors James A. Burger, Chin Y. Kuo, Clifford W. Randall, and Jakson R. Webster for their advie and enouragement. I was introdued to soil siene by Dr. Burger, hydrology and hemial transport models in waterways by Dr. Kuo, and systems eology by Dr. Webster, who also helped me maintain my major researh interest, nutrient yling in eosystems. Dr. Randall has materially and spiritually supported me while I have been enrolled in the Environmental Siene and Engineering Program. Appreiation should also go to Dr. Andrew N. Sharpley of the Water Quality and Watershed Researh Laboratory, Oklahoma, to Drs. David C. Martens, Raymond B. Reneau and Jak C. Parker of the Agronomy Department, and to Dr. Saied Mostagimi of the Agriultural Engineering Department for useful suggestions throughout my researh. I also aknowledge Dr. Luian W. Zelazny from the Agronomy Department for kindly providing laboratory proedures for the segregation of soil partile size frations and phosphorus adsorption reations. I wish to also thank Mr. Jan Carr for providing assistane with omputations and the Agriultural Engineering Department for the use of its failities. Aknowledgement is also Aknowledgements iv

5 extended to Miss Helen Castros and Mr. Craig Eddleton, laboratory speialists in the Soil and Water Quality Laboratory of the Agriultural Engineering Department for their assistane in the analyses of soil and water samples, to Dr. Donald W. MKean, International Student Coordinator for advies on writing tehniques, and to the Virginia Water Resoure Researh Center for the two-years of finanial support provided during this study. The support of Southern Regional Projet S-164 also is gratefully aknowledged. Deep gratitude is expressed to my parents Hwansu and Deokee Lee for reserving my responsibility as the eldest son, and to father-in-law Cheolkwon Kim for his attention and support. Thanks are due to my brother and sisters for their attention to my parents while I have been enrolled as a student in the United States. This thanks may be little understood in the United States. In Korea, traditionally, the eldest son should stay with his parents and take are of them. Finally, I wish to express appreiation to my wife Hyeonsuk, daughter Chang ha, and son Yeoam for their sympatheti understanding and sarifie during the ourse of this work. Aknowledgements v

6 Table of Contents Introdution Literature Review , Nutrient Filtering Mehanisms Hydrologi Aspets Chemial and Biohemial Aspets Summary Vegetative Filter Strip Researh Sediment and Phosphorus Transport Models SEDIMOT II Phosphorus Transport Model Experimental Plot Simulation Experimental Design and Sampling Laboratory Tests Runoff Samples Soil and Plant Materials Table of Contents vi

7 Mathematial Model Development Coneptual Model Formulation of the Mathematial Model Dissolved Constituent Transport Sediment-Bound Constituent Transport Potential Appliations Appliation to Grass Filter Strips Formulation of the Finite Differene Equations Results and Disussion Experimental Plot Simulations Runoff Yield Sediment Yield Phosphorus Yield Phosphorus Adsorption and Desorption Kinetis Adsorption Kinetis Desorption Kinetis Model Verifiation Input Data Model Validation Sensitivity Analysis 15 Model Assumptions and Limitations Summary and Conlusions Summary Conlusions Future Researh Needs Table of Contents vii

8 Referenes Summarized Water Quality Data for Feedlot Simulation Program Doumentation Input Data File Doumentation Definitions of Variables in GRAPH Computer Program 155 Example Input Data and Output Vita Table of Contents viii

9 List of Illustrations Figure 1. Nutrient yle in vegetative filter strips Figure 2. Shemati representation of experimental field plots Figure 3. Shemati diagram of phosphorus movement in nutrient soure area and VFS. 38 Figure 4. Shemati diagram showing the input-output of a dissolved hemial in an element of surfae runoff Figure 5. Network for the impliit method Figure 6. Hydrograph of runoff disharge, Test 4 and Run Figure 7. Cumulative sediment partile size distribution in plot runoff Figure 8. Total suspended solid disharges, Test 4 Run Figure 9. Sediment yields for plots QF4, QF5 and QF6 for Test Figure 1. Sediment yields for plots QF4, QF5 and QF6 for Test Figure 11. Sediment onentrations for uniform flow plots QF4, QF5, and QF6, Test 1 Run Figure 12. Sediment onentrations for onentrated flow plots QF7, QF8, and QF9, Test 1 Run Figure 13. Total phosphorus disharge, Test 4 Run Figure 14. Orthophosphorus disharge, Test 4 Run Figure 15. Orthophosphorus loss from plots QF1, QF2, and QF3, Tests 1 and Figure 16. Orthophosphorus loss from plots QF4, QF5, and QF6, Tests 1 and Figure 17. Orthophosphorus loss from plots QF7, QF8, and QF9, Tests 1 and Figure 18. Phosphorus ontent of sediment, plots QF1-9, Test 4 Run Figure 19. Adsorption of phosphate on silt123 and lay123 as a funtion of In t List of Illustrations ix

10 Figure 2. Adsorption of phosphate on silt456 and lay456 as a funtion of In t Figure 21. Plot of phosphorus desorption versus reation time for VFS soils Figure 22. Desorption of phosphate from leaf litter as a funtion of In t Figure 23. Predited and observed hydrographs for plot set QF456 during Test 4 Run Figure 24. Predited and observed total suspended solids for plot set QF456, Test 4 Run Figure 25. Predited and observed onentrations of sediment-bound phosphorus for plot set QF456, Test 4 Run Figure 26. Predited and observed onentrations of orthophosphorus for plot set QF456, Test 4 Run List of Illustrations x

11 List of Tables Table 1. Partile size distribution of undispersed soils in experimental plots Table 2. Experimental plot harateristis and operating onditions Table 3. Cumulative sediment partile size distribution in plot runoff, perent Table 4. Comparison of total and Bray I available phosphorus ontent in plot soils 81 Table 5. Comparison of partile size segregation methods for phosphorus adsorption. 83 Table 6. Time variation of phosphorus adsorption for soil partile lasses Table 7. Time variation of phosphorus release from nutrient soures in VFS Table 8. Elovih adsorption parameters for soil size lasses Table 9. Estimated Elovih equation oeffiients Table 1. Estimated phosphorus desorption oeffiients for Equation [1] Table 11. Elovih desorption parameters for plant and soil materials Table 12. Simulated runoff, total suspended solids, and phosphorus yield, and partile size distribution for plot set QF456 runoff Table 13. Sensitivity of total suspended solids and phosphorus yield to variations in model parameter Table 14. Perent deviation of total suspended solids and phosphorus yield due to hanges in input parameters Table 15. Relative sensitivity of the model to hanges in model paraments List of Tables xi

12 Chapter 1 Introdution Soil erosion is a signifiant fator limiting soil produtivity and an also deteriorate water quality if the assimilative apaity of aquati systems is exeeded. The seletive transport of plant available nutrients with surfae runoff and sediment has been studied extensively due to its importane to both fertility and water quality. Exessive sediment inflow to aquati systems an damage spawning sites and the habitats of aquati organisms, indue turbidity, and thus limit aquati life and diminish the storage apaity of streams and reservoirs. An inreased input of nutrients leads to eutrophi algal blooms in many aquati systems. Nutrient ontributions to water bodies from land runoff have inreased relative to muniipal and industrial sewage effluents as treatment of the latter has improved. Hene, pollution from land runoff, referred to as nonpoint soure pollution, is now reeiving onsiderable attention. In may ases, agriulture is the prinipal ause of nonpoint soure pollution. Intense agriultural land use inreases erodibility by disturbing surfae soils. In addition, agriultural appliation of fertilizers and bioides inreases every year as more land is brought into prodution and existing agriultural land is farmed more intensively. This has resulted in inreased transport of these pollutants in stormwater runoff. Urban development also affets I ntrod utio n 1

13 the quality of storm runoff and the onentration of suspended solids and hemials in the runoff. Several studies have indiated that pollutant ontributions from urban stormwater runoff may be more signifiant than agriultural soures in many areas. Phosphorus and nitrogen in stormwater runoff from thirteen urban athments representing six homogeneous land uses in the Virginia suburbs of Washington, D.C., were found to be present in suffiient quantity to ause algal blooms (Grizzard et al. 198). Mining and onstrution ativities also have reeived attention due to their pollution potential. These ativities disturb the surfae soil and ause onentrated sediment and hemial transport to loal aquati systems, but they are frequently treated as point soures. Surfae runoff and the dislodgement and transport of soil partiles are the major proesses affeting hemial movement. Sine soil partiles and organi matter are major transport vetors for many hemial speies, erosion must be ontrolled to redue the transport of these pollutants, nutrients, pestiides, toxis, pathogens, and other harmful materials to water bodies. Organi matter is preferentially transported by erosion due to its low density, and the organi matter ontent of eroded material is therefore higher than that of the parent soil (Voroney et al. 1981). Phosphorus is one of the most important and essential nutrients in freshwater bodies. It an ause signifiant water quality degradation if present in exessive amounts. Phosphorus availability usually limits plant growth in unpolluted natural freshwater bodies, and exessive phosphorus has been found to stimulate eutrophiation (Taylor 1967). Dissolved phosphorus is partiularly of importane beause it is more readily available to aquati organisms than partiulate phosphorus. Eroded soil is usually riher in phosphorus than the A 1 horizon beause of the element's strong adsorption to fine mineral soil partiles (Taylor and Kunishi 1971, Taylor et al. 1971) and the nature of the erosion proess whih seletively transport organis, high ation exhange apaity (CEC) lays, and silt size partiles (Singer and Rust 1975). Best management praties for the ontrol of runoff and sediment are often used to redue phosphorus inflows into water bodies. Mass movement is ontrolled by a variety of Introdution 2

14 physial and/or biohemial fators, suh as soil type, topography, vegetation over, season of the year, limate (espeially rainfall), and nutrient availability (Bormann and Likens 1967, Ryden et al. 1973, Gorham et al. 1979). Identifiation of these ontrolling fators is neessary for the suessful implementation of management praties designed to prevent or reverse the detrimental impats of the nonpoint soure pollution. Of the several fators involved, vegetation, by modifying the hydrologi response of soure areas, is the most signifiant fator affeting the movement of water, sediment, nutrients and other materials (Sopper 1971). Permanently vegetated areas are partiularly effetive in nutrient uptake ompared to ropland (Peterjohn and Correl 1984, Nikitin and Spirina 1985). Therefore, vegetative areas, referred to as vegetative filter or buffer strips (VFS or VBS), have been promoted in many areas as a means of removing nonpoint soure pollutants from surfae runoff before runoff reahes water bodies (Sullivan 1986). Vegetative filter strips are planted or indigenous vegetation zones that are loated between pollutant soure areas and water bodies. They are used to filter runoff, trap soil partiles, and protet the soil surfae against loal sour and erosion. Vegetative filters are also benefiial beause they remove fertilizers, pestiides, and miroorganisms from upland runoff that otherwise might reah waterways. In addition, the strips often serve as environmental orridors. They provide valuable food, habitat, and travelways for wildlife. As a result, they permit a greater diversity of wildlife, whih in turn ontributes to a more stable environment, proteting stream water quality and onserving plant and wildlife habitat. These areas also should be onsidered as a landsape omponent sine they are aesthetially pleasing. Vegetative filters, however, may onflit with other land uses sine they an oupy large land surfae areas. Therefore, an appropriate means of determining optimal plaement, dimensions, and arrangement of VFS must be developed if they are to be used effetively and eonomially (Swift 1986). In evaluating the effetiveness of VFS, it is desirable to identify those harateristis whih affet the effiieny of nutrient and sediment redution. One means is to physially or mathematially simulate VFS under a variety of biogeohemial onditions and to evaluate their effetiveness. Unfortunately, urrent nutrient transport mod- Introdution 3

15 els are not appropriate for quantitatively evaluating the effetiveness of existing or planned VFS. SEDIMOT II is presently the only physially based design model available for simulating the response of VFS to sediment transport during storm runoff. The model was developed to simulate surfae runoff and sediment transport from areas disturbed by surfae mining ativities and an desribe deposition and the partile size distribution of eroded sediment as it travels through VFS. SEDIMOT II may be useful in modeling nutrient transport in surfae runoff, for its ability to predit the partile size distribution of transported sediment is essential for modeling nutrient transport. Another useful model is urrently under development at the Water Quality and Watershed Researh Laboratory, Oklahoma. This empirially data-based model desribes dissolved phosphorus transport in surfae runoff, but it does not onsider VFS. The primary goal of the researh presented here is to assess the effetiveness of VFS in removing phosphorus from surfae runoff. The speifi objetives of this researh are: 1. To develop a omputer model for simulating phosphorus trapping in VFS during single storm event by ombining SEDIMOT II and the Oklahoma model. 2. To investigate the nutrient trapping mehanisms of VFS. 3. To investigate the response of VFS to runoff, sediment, and phosphorus inputs using experimental field plot studies. The field tests are required for model verifiation and the development of mathematial phosphorus transport subproess models. 4. To examine existing phosphorus desorption and adsorption models through laboratory studies for use in parameter seletion and model verifiation. 5. To evaluate the sensitivity of the model to input parameters. Introdution 4

16 Chapter 2 Literature Review The design of VFS should be based upon the ability of VFS to remove nonpoint soure pollutants suh as sediment, nutrients, and other hemially and biologially detrimental agents from surfae runoff. Removal mehanisms are, however, not well understood. In this hapter, the potential mehanisms by whih the filters remove sediment and nutrients are reviewed. The review is foused on phosphorus but general nutrient removal mehanisms are disussed. In addition, previous VFS researh and urrent sediment and phosphorus transport models are briefly reviewed. Nutrient Filtering Mehanisms Nutrient budgets in vegetative areas may be determined by mass balane equations that take into aount elemental inputs and outputs and hydrologi, hemial and biologial proesses (Gorham et al. 1979). Beause water is a major transport vehile for nutrients, Literature Review 5

17 hydrologi proesses govern the spatial distribution of nutrients. Chemial and biologial proesses determine the fration of water-soluble and insoluble hemial forms. Jordan and Kline (1972), therefore, suggest that inreased runoff does not neessarily indue orrespondingly high losses of nutrients from nutrient-unsaturated soils. Hydrologi, hemial and biohemial phenomena are important for nutrient onservation on land and the water quality of the orresponding aquati systems (Bormann and Likens 1967, Dillon and Kirhner 1975, Karr and Shlosser 1978, Miller et al. 1979, Shlosser and Karr 1981a, b). Hydrologi Aspets Considerable losses of soil onstituents from land an be attributed to the energy of. raindrop splash and overland flow. These two fators are major onsiderations in the down- slope movement of water, sediment, and nutrients (Singer and Rust 1975). Thus, any mehan ism dereasing the energy of rainfall and surfae runoff ontributes to nutrient onservation in terrestrial eosystems. In general, intereption, infiltration, surfae retention and detention, and evapotranspiration attenuate runoff disharge and thus its kineti energy (Huggins and Burney 1982). Runoff disharge from a landsape is highly variable depending on the amount of leaf area and soil water storage apaity (Fahey and Knight 1986). Waring and Shlesinger (1985) lassified water storages in vegetative systems into the areas of: (1) foliage, branhes, and stems; (2) snowpak; (3) litter layer; (4) soil surfae; (5) vegetation; (6) soil root zone; and (7) subsoil. These storages derease raindrop energy and/or runoff kineti energy through their effets on hydrologi omponents. To begin with, raindrop energy is dissipated when rain is interepted by plant foliage. The total intereption of rainfall onsists of water stored on vegetation and that whih is diretly evaporated from the plant surfae into the air (Leonard 1967). The amount of storage and evaporation during preipitation inreases with an inreased leaf-area index (LAI), de- Literature Review 6

18 fined as the ratio of leaf-area to the soil area it oupies (Marks and Bormann 1972, Wright 1973), when the same geometry and spaing of leaves are assumed. The surfae litter layer also is signifiant beause ats as a sponge retaining and re-evaporating inident water, and thus restriting water movement to the soil surfae where exess water may generate surfae runoff (Moore 1985). The amount and hemial omposition of rainwater is modified when it ontats living and/or dead foliage (Fahey and Knight 1986, Knapp and Seastedt 1986). Living foliage and plant detritus were found to interept approximately 4 perent of rainfall in undisturbed prairie on an annual basis (ited by Knapp and Seastedt 1986). Infiltration and field apaity are diretly proportional to the organi matter ontent of soil, most of whih originates from plant materials. Tiessen et al. (1984) identified relatively oarse plant debris and olloidal and soluble organi ompounds from root exudates, mirobial produts and litter leahates as major soures of soil organi matter. Wishmeier and Mannering (1965) showed that infiltration inreases proportionally to the organi matter ontent of soil. Robertson and Vitousek (1981) presented data showing that the water-holding apaity of soil inreases with the eosystem suession, whih is likely explained by an inrease in soil organi matter ontent (Lola et al. 1984). Suh hydrologi phenomena may be related to the improvement of soil strutural properties. It has been known that organo-mineral assoiates promote soil struture by funtioning as binding agents in soil aggregates (Tisdal and Oades, 1982). Evapotranspiration ours at the earth's surfae in proportion to the air-exposed area whih may be a funtion of LAI in vegetative areas (Waring and Shlesinger 1985, Fahey and Knight 1986). Greenwood et al. (1985) reported that annual evapotranspiration was orrelated with leaf area in Australian Eualyptus forests. Vegetation also tends to roughen the surfae of waterways and thus retard surfae flow, allowing it to move into the soil (Smith 1974). Plant litter also ontributes to surfae roughness and thus redues the transport apaity of surfae runoff (Swift 1986). The roughness may be proportional to LAI or the above-ground biomass, and has been expressed as the Manning roughness oeffiient by engineers (Kouwen et al., 1969). Literature Review 7

19 In summary, vegetation redues surfae runoff by dereasing the amount of preipitation reahing the soil surfae, by inreasing infiltration, by roughening the soil surfae, and by ontributing to rainwater intereption and transpiration. Both retardation of flow and derease in runoff disharge redue the kineti energy of runoff, and thus lower its sediment transport apaity. Sediment-bound nutrients an then be removed from runoff in vegetative zones as sediment is deposited due to transport apaity defiits. If nutrients are predominantly sediment-bound, then the deposition proess will largely ontrol the effetiveness of the vegetative area for the removal of nutrients from the runoff. The movement of phosphorus from land to water bodies has been found to be strongly dependent on sediment transport proesses sine it is predominantly sediment-bound (Burwell et al., 1975, Singer and Rust 1975, Mitsh et al. 1979, Sharpley and Syers 1979). Burwell et al. (1975) reported that at least 95% of annual phosphorus losses were aounted for by sediment transport in most soil over treatments. With respet to the annual loss of total phosphorus, 76% was in a partiulate form for a small agriultural athment in New Zealand (Bargh 1978). Rigler (1979) and Johnson et al. (1976) observed that approximately 75% and 78%, respetively, of phosphorus in streams during storm runoff was partiulate. The phosphorus ontent of suspended solids was found to be.12% in a forested and agriultural watershed in New York State (Johnson et al. 1976) and.2-.25% in Dartmoor stream (Rigler 1979). In general, lay-sized partiles have higher phosphorus onentrations than bulk soil beause organi phosphorus and inorgani phosphates are adsorbed by lays (Sharpley 198). Coarse partiulate organi matter (CPOM), however, may also be a signifiant soure of phosphorus. In a woodland stream, CPOM aounted for 6% of the total uptake of phosphorus, FPOM (fine partiulate organi matter) for 35%, and aufwuhs for 5% (Newbold et al. 1983). In desribing the nutrient retentiveness of forests, Knight et al. (1985) have emphasized the hydrologi proesses of rainfall intereption, evapotranspiration, and hydrograph shape. Ryan and Bormann (1982) laim that the size of the nutrient resorption pool is a funtion of forest leaf biomass. Their laim has been supported by Hewlett et al. (1984) who suggests that the mobility of nutrients appears to be restrited by the ability of eosystems to evaporate Literature Review 8

20 water on site, whih is, in turn, assoiated with vegetative over. Soil water storage apaity is assoiated with LAI due probably to the orrelation of LAI and soil organi matter ontent aording to Fahey and Knight (1986). Vegetation dereases runoff, and, in turn erosion and nutrient losses, by enhaning evapotranspiration and infiltration or water storage apaity of a landsape. Finally, Knight et al. (1985) onlude that the leaf area of vegetation, duration of the vernal transpiration, and high arbon/nutrient ratios in the terrestrial forest floor have substantial impats on nutrient onservation in terrestrial eosystems. Chemial and Biohemial Aspets Biohemial aspets are signifiant, espeially when the long-term effetiveness of buffer areas is onsidered. Vegetative systems have several biohemial mehanisms to trap and retain nutrients from surfae and subsurfae flows (Marks and Boramnn 1972, Vitousek and Reiners 1976, Lowrane et al. 1984a, b). Jordan and Herrera (1981) and Herrera et al. (1984) attribute these mehanisms to the strutural and funtional harateristis of the foliage and root system of plants, and the soil or stand in tropial forests. Vitousek and Melillo (1979) list nine physial or hemial mehanisms by whih disturbed forests prevent nitrogen flow into streams: immobilization, prevention of nitrifiation, nitrate redution to ammonium, lay fixation, anion adsorption, plant uptake, insuffiient preipitation, ammonia volatilization, and nitrate redution to dinitrogen or nitrogen oxides. Nutrient onservation an also be attributed to the dynami distribution of diverse bioti and abioti holders (Odum 1969). This attribution is based upon the nutrient-retentive apaity of plants (Muller and Bormann 1976, Feller 1977, Sollins et al. 198, Cronan 198), mirobes (Vitousek and Matson 1984, 1985, Yavitt and Fahey 1984) inluding myorrhiza (Went and Stark 1968a,b), soil fauna, ombined surfae root mats (Stark and Jordan 1978) inluding deaying litter (Fahey 1983, Seastedt 1985), soil organi matter and minerals, and the interations of these omponents. Literature Review 9

21 In soil-plant systems, plants apparently affet the spatial pattern of organi matter, nutrients, and other physial and hemial properties of an eosystem through the yling of nutrients (Barth and Klemmedson 1978). The inorporation of nutrients into growing biomass onsumes water-transportable nutrients in the terrestrial system and thus redues the transport of nutrients into aquati systems (Vitousek and Reiners 1975, Lowrane 1981). Golkin and Ewel (1984) report that phosphorus uptake by herbs on the forest floor is proportional to gross produtivity. Cole and Rapp (1981) attest that nitrogen uptake and requirement of plants are strongly orrelated with biomass prodution, espeially in oniferous speies. The uptake by plant roots is one potential sink of nitrate during its downward movement in the soil (Khanna 1981). Thus spring ephemeral plant prodution has been onsidered as an important nutrient sink (Muller and Bormann 1976, Blank et al. 198). In addition, vegetation leads to lower onentrations of soluble ions available for losses in drainage water by moderating the soil temperature during the growing season, whih redues the deay rates of plant litters (Mark and Bormann 1972). Several studies have shown that plant uptake might not be a diret mehanism for nutrient onservation (Likens et al. 1978, Gholz et al. 1985). More inorgani nitrogen in rainwater is taken up by miorbes on detritus than by living foliage (Seastedt 1985). A simulation model demonstrated that the phosphorus uptake rate is 4 to 5 times higher by miroorganisms than by plants in semiarid grasslands (Cole et al., 1977). Mirobial immobilization initially aounted for removal of phosphorus added in organi wetlands (Rihardson and Marshall 1986). Vitousek and Matson (1984, 1985) report that mirobial immobilization is the predominant fator aounting for lower nitrogen losses from a lear-ut Piedmont site in North Carolina. Yavitt and Fahey (1984) also emphasize that miroorganisms store nutrients, with an estimated 13% and 7.5% of the total nitrogen in the 1 and 2 layers. Trudinger et al. (1979) laim that the mirobial role in the nutrient yle is important beause : (i) miroorganisms make up the bulk of the mass of the biosphere and their rates of growth are generally several orders of magnitude greater than those of the higher organisms; (ii) mirobes arry out many unique reations of geohemial signifiane; (iii) the time period Literature Review 1

22 over whih miroorganisms have olonized earthly environments is four to five times greater than that oupied by higher organisms, and (iv) the mirobial world embraes a wider range of environments than that for plant and animals. Reently the effet of soil mirofauna on nutrient yling proesses has been studied. It has been suggested that soil fauna, suh as miroarthropods, maroarthropods, snails, nematodes, and oligohaets retard the rates of nutrient loss from plant litter by stimulating mirofloral growth, whih, in turn, leads to nutrient tie-up (Doue and Crossley 1982). Elliott et al. (1984) observed that high soil protozoan ativity was orrelated with large delines in mirobial arbon and phosphorus and with inreases in mineral nitrogen and extratable inorgani and organi phosphorus. They further noted that protozoan grazing resulted in a high biomass C:N ratio due probably, to a shift from a fungal to a baterial dominated miroflora. Mass and nutrient losses from forest litter are enhaned by the omminution of miroarthropods, and their feeding ativities may stimulate mirobes to inrease the nutrient retention apaities of forest litter (Seastedt and Crossley 198, Parker et al. 1984). Mirobial. feeding by soil fauna has been reported to result in inreased mirobial ativities (Baath et al. 1981, Woods et al. 1982, Parker et al. 1984). In these studies, mirobial feeding nematodes redued nitrogen loss through leahing (Baath et al. 1981) and aused inreased uptake of nitrogen by plants, whose growth was aelerated by the inreased mineralization due to bateria and by NHt-N exretion by nematodes (Ingham et al. 1985). Aording to these studies, soil mirofauna funtion as nutrient-holders and also promote the ativities of other holders. Deomposing leaf litter may also play a signifiant role in nutrient trapping in vegetative areas. Approximately 6% of the total phosphorus uptake in Walker Branh, a small.firstorder woodland stream in east Tennessee, was assoiated with leaf detritus (Newbold et al. 1983). Sorption onto substrates in streams is predominantly a biologial proess regulated by the quantity of mirobes present, whereas physial sorption is generally less than 2% of the phosphorus transloation onto benthi substrates in streams (Gregory 1978). Deaying leaf litter supplies energy or ats as a arbon soure to algae, bateria, and fungi, and these Literature Review 11

23 miroorganisms assimilate phosphorus from the water (Gregory 1978, Meyer 198). In a researh of radioative trae, inorgani phosphate in water was taken up by mirobes assoiated with deomposing leaves rapidly, often in as short a time as 5 minutes. As muh as 95% of the radioative phosphate was removed within 1 m of stream length (Ball and Hooper 1963, ited by Smith 1974 and Gregory 1978). Even though hydrologial and biologial situations on land are not idential with those within streams, the phosphorus ontent of dead leaves has been found to inrease while leaves are deaying in the terrestrial environment (Gosz et al. 1973, Williams and Gray 1974, Fahey and Knight 1986). The phosphorus soure may ontain preipitation, upland inflow, and fungal transloation from the humus and upper mineral soil layers (Fahey and Knight 1986). This may be explained by the findings that the addition of organi arbon soures, suh as surose, dried grass, or ellulose to surfae soil, aelerated mirobial ativity and subsequently inreased mirobial and organi phosphorus (Hannapel et al., 1964, Chauhan et al. 1979, 1981). Similarly, deaying leaf litter in vegetative filters may at as a good arbon soure, attrating organisms and ontributing to the formation of a biosurfae on the filters, whih an remove phosphorus from runoff and upper mineral soil. The biosurfae is defined as the biologial surfae whih is exposed to runoff water. If dead litter promotes phosphorus transloation from upper mineral soil to mirobes, phosphorus ontent in mineral soil is relatively low and thus it would indiretly enhane the adsorption of phosphorus onto the mineral from storm runoff. Consequently, it is promising to note that the potential of mirobial uptake of phosphorus an be estimated by the quantity and quality of dead litter. On the other hand, it has been suggested that leaf litter may also ontribute to soluble phosphorus in runoff from woodland and farmland (Taylor et al. 1971, Singler and Rust 1975). Under ertain onditions, surfae runoff an exude a substantial amount of nutrients from vegetative over and litters (Timmons et al. 197, Burwell et al. 1975). Covington (1981) suggested that plant debris might behave as a sink for nutrients after a forest was disturbed, whereas it funtioned as a soure during the rapidly aggrading phase. This phenomena was explained by noting that leaf litters with a high arbon/phosphorus ratio at as phosphorus Literature Review 12

24 sinks, while those with low ratios may release phosphorus. Fuller et al. (1956) reported that arbon/phosphorus ratios above 2 initiated the immobilization or biologial fixation of phosphorus. This onlusion was supported by data showing that deomposing plant litters of high arbon/phosphorus ratios redued available phosphorus by immobilizing soluble phosphorus (Gosz et al. 1973; Melillo and Aber 1984). This finding was generalized by Staaf and Berg (1982), who suggested that elements limiting mirobial growth were trapped in deaying plant litter, while unlimited nutrients were released during the whole deomposition period (Staaf and Berg 1982). Phosphorus is the most limiting element for mirobial ativity during the deay phase. However, there still is a ontroversy regarding the major sink of the limiting element. While many studies have onsidered the inrease of mirobial biomass as the major phosphorus sink, others have suggested that a large fration of phosphorus may be immobilized by extraellular reative ompounds, whih are produed as mirobial enzymes depolymerize the detritus substrates (Melillo and Aber 1984). The onversion of surfae runoff into subsurfae flow generally leads to water purifiation (Nikitin and Spirina 1985). When dissolved nutrients are arried into the soil along with infiltrated water, the filtering apaity of subsurfae soil is maximized. Thus soil organi matter may ontribute to nutrient retention by inreasing infiltration rates and water-holding apaities of soils (Wishmeier and Mannering 1965, Allison 1973, Hollis et al. 1977, Robertson and Vitousek 1981). This may be explained by the funtion of organi materials as binding agents. Organi matter tends to promote water-stable aggregates in soils (Tisdall and Oades 1982). The nutrients in perolated water are more likely to be adsorbed onto soil partiles than those nutrients in surfae runoff, espeially organi matter and lay fration (Ryden et al. 1973). Furthermore, most water retained by organi matter is readily available to plants (Hollis 1977). Therefore, organi matter provides favorable onditions under whih plants and mirobes an deplete water and nutrients effetively, and then the water- and nutrient-holding apaity is readily reovered. Barbour et al. (198) summarized the role of organi matter in nutrient retention. "Humi substanes and other organis play a vital role in nutrient storage, forming omplexes with minerals that enhane uptake by plants. Immobilization through deompos- Literature Review 13

25 ition serves to retain nutrients beyond the period of dormany to the time of growth." In general, phosphorus retention is higher in vegetated than in mineral soils (Radwan et al. 1985). Some phosphorus will be adsorbed to the surfae soil of vegetative filter strips while water is running off (Sharpley et al. 1981). It is aepted that soil adsorption aounts for phosphorus retention when biologial retention apaity is saturated or disturbed (Wood et al. 1984, Rihardson 1985, Rihardson and Marshall 1986). The adsorption amount will be ontrolled by soil texture, ph, organi matter, the 'free iron oxides' and 'extratable aluminum' ontents of soil (Mlean et al. 1958), the soil solution temperature, and the runoff rate. Partile size distribution has been reported to aount for up to 9% of the distribution of soil phosphorus (Campbell et al. 1984, O'Halloran 1985). Clay-sized partiles are, in general, responsible for most of the phosphorus ontent of bulk soils. Phosphorus retention in soils has been known to be losely related to the ontent of aid-extratable ferri and aluminum ions in soil (Bohn et al. 1985). Inreasing soil ph dereases exhangeable Al and aetateextratable Al, and thus to some extent, dereases phosphate retention (Lopez-Hernandez and Burnham 1974b). High soil ph is also related to low soluble phosphorus due to inreased reations between phosphorus and Ca (Adams and Odom 1985). The effets of ph on phosphorus solubility are dependent upon phosphorus ompounds (Lindsay and Moreno 196). Aluminum and iron phosphates beome soluble with an inreased ph, while alium phosphates follow an inverse diretion. Organi matter an sorb phosphate or else blok potential phosphate sorbing sites on inorgani partiles, depending upon ertain onditions (Lopez-Hernandez and Burnham 1974a). Hene, soil organi matter an show either a positive or a negative relationship with phosphate adsorption. Lopez-Hernandez and Burnham (1974a) reported that about 85% of the variation in phosphate sorptivity ould be explained by a multiple regression equation involving free iron oxides, extratable aluminum, organi matter, lay ontent and ph. However, lay ontent and ph made a omparatively small ontribution in some British and tropial soils. Usually, organi matter, lay ontent, and soil aggregates are outstandingly important in the transport of adsorbed hemials. Literature Review 14

26 Summary The proesses affeting nutrient inflows into vegetated soil may inlude inflows from upland areas through surfae and subsurfae waters, soil weathering, nitrogen fixation, partile impation, and gas adsorption (Gorham et al. 1979). Nutrient output proesses onsist of the removal of dissolved nutrients by surfae runoff, erosion, subsurfae drainage, and gaseous losses inluding volatilization and denitrifiation (Burwell et al. 1977). Thus, the fate of nutri;.. ents in vegetative areas an be oneptualized as shown in Figure 1. It is believed that eah omponent of an eosystem plays a role as a nutrient holder, and the nutrient-retentive apaities of whole systems are synergistially promoted by the the interplays of the omponents. However, if gaseous outputs do not exeed their inputs, the exess of nutrients over the retention apaity of terrestrial systems are destined to eventually flow into aquati systems via surfae and/or subsurfae waterways. It is important to note that when forest eosystems reah steady state prodution, their ability to trap and retain nutrients from runoff and subsurfae water may be negligible (Vitousek and Reiners 1975, Omernik et al. 1981, Lewis 1986). There are no permanent nutrient sinks in nature. Existing nutrient sinks are transformed into nutrient soures as irumstanes hange. In this ontext, so-alled best management praties (BMP's) ould be defined as means of promoting the apaity of nutrient sinks in terrestrial systems and of reduing the transformations of sinks to soures. Thus there is a suggestion that appropriate harvest of plants prevents the transformation of vegetal omponents into inorgani nutrients, whih are vulnerable to leahing. Lowrane (1981) laimed that nutrient retentiveness by riparian eosystems ould be ontinued if biomass aumulation and harvest were maintained with a minimum of soil disturbane during the driest months. Literature Review 15

27 INPUTS INTERNAL CYCLE OUTPUTS u land inf low erosion weathering leahing bulk preipitation aseous output absor tion/ impation Inorgani form Organi ~------i residue harvest Figure 1. Nutrient yle in vegetative filter strips: Modified from Vitousek (1981). Literature Review 16

28 As explained, omplex omponents and proesses are assoiated with the nutrient retention mehanisms in vegetative systems. Some of the omponents or proesses in Figure 1 are diretly assoiated with the nutrient filtering funtion of VFS during storms. Others may only indiretly promote the nutrient filtering funtion. Some exert their influene on nutrient trapping only during a storm while others ontinuously deplete the filtered nutrients. For example, hydrologi proesses predominant during rainfall events. In ontrast, bioti omponents onsistently take nutrients up and thus redue nutrient onentrations in the soil solution, whih an trigger more nutrient trapping during storms. Vegetative Filter Strip Researh Several studies have suggested that VFS redue runoff disharge and result in redued pollution problems assoiated with agriultural runoff. For example, Westerman and Overash (198) found that over a 3-month period 21 and 1 perent of the rainfall volume was generated as runoff from an open dairy lot and neighboring pasture, respetively. They attributed the differene in runoff to the high infiltration and surfae storage in the pasture. Sediment transport in VFS has been extensively investigated by a researh group at the University of Kentuky (Tollner et al. 1977, Barfield and Albreht 1982, Hayes et al. 1984). The Kentuky researhers observed that the majority of sediment deposition ourred just upslope of the filter and within the first meter of the filter until the upper portions of the filter were buried in sediment. Subsequent flow of sediment into the filter resulted in the advane of a wedge-shaped deposition of sediment down through the filter. The Kentuky researh reported high trapping effiienies as long as the vegetal media was not submerged, but trapping effiieny dereased dramatially at higher runoff rates that inundated the media. The results of this VFS researh were inorporated into SEDIMOT II, whih will be disussed in the Literature Review 17

29 next setion. Dikey and Vanderholm (1981) observed signifiant redutions in runoff and removal of as muh as 95 % of nutrients and oxygen-demanding materials by vegetative filters. They also found that a hannelized flow redued filter effetiveness with respet to runoff and nutrient redution. Young et al. (198) onstruted field plots on a 4 perent slope with the upper 13.7 m ative feedlot and the lower 27.4 m in a planted area. When a 25-year, 24-hour storm was simulated, on the average, filter strips, redued the total runoff, suspended solids, nitrogen, and phosphorus by 67, 79, 84, and 83 perent, respetively. The VFS also redued NH 4 -N, P 4 -P, and oliform bateria but inreased N 3 -N in the runoff. They suggested that planted vegetation released N 3 -N into the runoff. Edwards et al. (1983) monitored storm runoff from the outlet of a 243-m 2 paved feedlot. They also measured and sampled runoff after the feedlot runoff passed through a shallow, onrete settling basin and two onseutive 3 m long by 45 m wide sod filter strips. The first filter redued the runoff, total suspended solids, nitrogen, and phosphorus by -2, 23, 31, and 29 perent, respetively. The seond filter showed less effetiveness in dereasing the same parameters, with additional redutions of -5, 1, 16, and 14 perent, respetively. The redued effetiveness may be attributed to the seletive filtration of easily trapped materials by the settling basin and first filter strip. The total runoff from the filters was greater than the inoming runoff beause the rainfall rate during runoff events exeeded the infiltration rate of the filters. This rainfall exess, oupled with the added area of the filters, resulted in the inreased runoff. They also suggested that the rainfall in the filter area diluted storm runoff during large storms and that infiltration redued the transported nutrients to downstream areas. Literature Review 18

30 Sediment and Phosphorus Transport Models Examples of models urrently available to simulate storm runoff and/or sediment transport proesses on a watershed or field sale inlude: USLE, MUSLE, ARM, CSU, CREAMS, ANSWERS, FESHM, and SEDIMOT II (see Dillaha 1981, Wilson et al. 1981, Storm 1986). Nutrient transport, however, has not yet been predited with any reasonable level of auray (Sweeney et al. 1985). In this setion, storm runoff, sediment, and phosphorus transport models will be briefly disussed. SEDIMOT II SEDIMOT II (SEdimentology by Distributed MOdel Treatment) is a simple distributed parameter simulation model developed by a researh group at the University of Kentuky. Runoff disharge, sediment yield, partile size distribution, and sediment graphs are predited for eah subwatershed and ombined at the watershed outlets. A major feature of SEDIMOT II is that it uses empirial runoff routing tehniques to redue the input data required by ommon distributed models suh as ANSWERS and FESHM. SEDIMOT II is unique in desribing the responses of grass filter strips to runoff and sediment inflows. Subroutine GRASS, whih desribes the performane of grass filter strips, has been developed by several researhers. Toller et al. (1976) presented design equations relating the fration of sediment trapped in simulated vegetal media to the mean flow veloity, flow depth, partile fall veloity, filter length, and the spaing hydrauli radius of the simulated media. Barfield et al. (1979) developed a steady-state model for determining the sediment filtration apaity of grass media as a funtion of flow, sediment load, partile size, flow duration, slope, and media density. Outflow onentrations were found to be primarily a funtion of slope and media spaing for Literature Review 19