Modeling phosphorus in the environment - Chapter 5 potx

A complex array of physical, abiotic, and biotic processesoccurs within the channel and hyporheic zone of streams, and these complex processes, as well as land-use impacts, influence bot

in Streams: Processes and Modeling

Considerations

Brian E Haggard

University of Arkansas, Fayetteville, AR

Andrew N Sharpley

U.S Department of Agriculture-Agricultural Research Service, University Park, PA

CONTENTS

5.1 Introduction 105

5.2 Abiotic and Biotic Processes 106

5.3 Phosphorus Spiraling 109

5.3.1 Determining Phosphorus Spiraling 109

5.3.2 Stream Properties and Phosphorus Spiraling 113

5.4 Algal and Microbial Processes 114

5.5 Stream Sediments and Phosphorus 115

5.5.1 Sediment Source Effects 115

5.5.2 Sediment and Equilibrium Phosphorus Concentrations 117

5.6 Impact of Stream Processes on Eutrophication 120

5.7 Modeling Phosphorus Transport in Stream Channels 122

5.8 Conclusions 124

References 125

5.1 INTRODUCTION

Modeling of phosphorus (P) transport from the landscape to aquatic systems represents

a number of complex processes, including rainfall–runoff patterns, manure and fertilizer application, soil–water–P interactions, and crop and forage growth Numerous process-based models are available to accomplish this task, and many of these models are discussed in other chapters of this book These models integrate large amounts of

information to simulate catchment-scale P transport from the landscape They havebeen reasonably successful in predicting catchment-scale to edge-of-field losses of Pand how nutrient and land management affects these losses However, a major gap inpredicting the response of receiving water bodies is the simulation of fluvial processesoccurring during P transport from landscape (edge of field) to receiving waters.Although much research relates land management to edge-of-field P losses, littledata are available on the fate and transport of this P once in fluvial systems, thoughthis can affect the amount ultimately entering a water body (McDowell et al 2004).Most importantly, tributaries can act as sinks and sources of P and can influence theeffectiveness of watershed best management practices (BMPs) and the response time

of impacted water bodies to land remediation (Meals 1992) The modeling of P transportthrough streams must account for variability in flow, P sources, and in-channel processes(Hanrahan et al 2001); however, the complexity and spatial variability of these in-stream processes limit the ability of catchment-scale models to simulate P transport instreams (Sharpley et al 2002) A complex array of physical, abiotic, and biotic processesoccurs within the channel and hyporheic zone of streams, and these complex processes,

as well as land-use impacts, influence both P concentrations and loads during stream transport Furthermore, these channel processes can greatly influence the short-and long-term impact of P inputs on the degree of eutrophic response of receiving waters.The proportion of P transported during storm and base-flow conditions varieswith catchment P sources and the importance of stream-channel and riparian pro-cesses Though long-term (e.g., annual) inputs and outputs of P to streams may besimilar, short-term (e.g., daily and monthly) transport of P from effluent discharges,drainage fields, and other upstream sources may be heavily buffered by stream-channel processes Streams that have been enriched with P will often act as short-term transient storage zones, releasing dissolved P back into the water column whenaqueous concentrations are low (Ekka et al 2006; Haggard et al 2005) The release

down-of P from sediment to overlying waters can delay or even mask P loss reductionsfrom catchment-based BMPs (Meals 1992; National Research Council 2000) Despite the importance of these in-stream processes, stream effects on discreteand edge-of-field P inputs are not adequately simulated in many process-basedmodels of catchment P transport This chapter describes the important fluvial pro-cesses that influence the form and amount of P transported in streams and approaches

to model P transfer from the landscape downstream

5.2 ABIOTIC AND BIOTIC PROCESSES

Aquatic systems — in particular stream reaches and networks — may alter thetiming, magnitude, and bioavailability of P transport from the landscape furtherdownstream (Meyer et al 1988; Sonzogni et al 1982) Several processes — such assediment sorption and desorption — occur that may influence P transport throughaquatic systems (Froelich 1988; Klotz 1988; Taylor and Kunishi 1971), precipitationand dissolution (Fox 1989; House and Donaldson 1986), microbial and algal uptake(Elwood et al 1981; Hill 1982), and riparian floodplain and wetland retention(Kronvang et al 1999; Mitsch 1992; Novak et al 2004) Many of the abioticprocesses are influenced or mediated by biota; for example, coprecipitation of

dissolved P with calcite may be biologically mediated during active photosynthesis(Neal 2001) Furthermore, sediment-associated biotia (i.e., microbial organisms)often account for large fractions of P uptake during sediment–P sorption experiments(Haggard et al 1999) For instance, several studies have found that aquatic biotaaccounted for 30 to 40% of P uptake and release in wetland and stream sediments(Khoshmanesh et al 1999; McDowell 2003; McDowell and Sharpley 2003) Incontrast, other work has suggested that the microbial community associated withstream sediments played only a small role in P sorption and buffering capacity (Klotz1988; Meyer 1979) Clearly, the temporary storage of P from these in-channelprocesses does alter the transport characteristics of P from different landscapepositions through streams to a given outlet within a catchment.

The importance of these in-channel processes varies with discharge regimes inindividual streams, where these abiotic and biotic processes are most likely to bemost important during relatively low- or base-flow conditions Several studies haveshown that P retention occurred in stream reaches during low- or base-flow condi-tions (Dorioz et al 1998; Hill 1982; House and Warwick 1998) Under these flowconditions P may be temporarily retained within a stream channel, but increasingdischarge during episodic storm events may resuspend phosphorus and transportparticulate phosphorus (PP) further downstream The resuspension of sedimentsmight also influence dissolved P concentrations in the water column of streamsduring these higher flow events (House et al 1995; Koski-Vähälä and Hartikainen2001) A large fraction of the P load transported downstream may be resuspendedfrom bottom sediments (Svendsen et al 1995)

In contrast, McDowell et al (2001) observed the opposite trend of dissolved Pretention by channel sediments during storm flow and release during base flow Reflect-ing the dynamic nature and site specificity of in-channel processes, catchment hydrol-ogy, and sources of sediment and P, McDowell et al (2001) described the mechanismscontrolling P release from soil and stream sediments in relation to storm and base flow

at four flumes along the channel of a 40 ha, second-order agricultural catchment(Figure 5.1) Base-flow dissolved P concentrations (average of 1997 to 2004) weregreater at the catchment outflow (0.042 mg L−1 at flume 1) than at the uppermost flume(0.025 mg L−1 at flume 4), whereas the inverse occurred during storm flow (0.304 mg

L−1 at flume 4 and 0.128 mg L−1 at flume 1) Similar trends in total P concentrationwere observed However, it is questionable whether short-term pulses in dissolved Phave much ecological impact in streams (Humphrey and Stevenson 1992)

During storm flow, in-channel decreases in P concentration were indicative ofdilution of P originating from a critical source area above the uppermost flume(flume 4), where an area of high soil P intersected an area of high erosion andoverland flow potential (Figure 5.1) During base flow, the increase in P concentra-tions downstream was clearly controlled by channel sediments, such that the Psorption maximum of the uppermost flume (flume 4) sediment (532 mg kg−1) wasfar greater than the outlet flume (flume 1) sediment (227 mg kg−1) Paralleling thesetrends, the sediment equilibrium phosphorus concentration (EPC0) of sediment atflume 1 was greater than at flume 4 (0.034 to 0.004 mg L−1) Sediment EPC0 trendswere highly correlated to base flow dissolved P concentrations (0.025 mg L−1 atflume 4 and 0.042 mg L−1 at flume 1)

FIGURE 5.1 The distribution of Mehlich-3 soil P (>100 mg kg−1 ), erosion (>6 mg ha−1 yr−1 ), and mean dissolved P concentration in storm and base flow (mean of 1997 to 2004) and P sorption–desorption properties of channel sediment at four flumes in FD-36, Pennsylvania.

(Adapted from R.W McDowell, A.N Sharpley, and G Folmar, J Environ Qual 30,

In general, P concentrations and loads increase with increasing discharge, especiallyconcentrations and other sediment-bound constituents often display hysteresis instreams, where the chemograph increases rapidly and peaks on the rising limb of thestorm-event hydrograph (Richards et al 2001; Thomas 1988) Thus, P concentrationsare often greater on the rising limb when compared with concentrations measured at asimilar discharge on the falling limb of the storm-event hydrograph (Sharpley et al.1976) Although P concentrations are typically greater during this flow regime, theimportance of in-stream P retention is minimized because of sediment resuspensionand scouring within the channel Many streams export a very large fraction — greaterthan 80% — of P loads during episodic storm events (Green and Haggard 2001; Pionke

et al 1996; Richards et al 2001), whereas Novak et al (2003) observed that greaterthan half of dissolved P export occurred during base-flow conditions A single largestorm event may often transport a large portion of the annual P load in many streams.Though stream reaches may show significant P retention during base-flow conditions,

P inputs would typically equal outputs on large time scales, such as annual export.However, P deposition on riparian floodplains may be a significant P sink during stormevents and stream-bank overflow (Kronvang et al 1999)

5.3 PHOSPHORUS SPIRALING

As P enters streams from discrete and diffuse sources, P cycles from the dissolvedinorganic form to the particulate form (abiotic and biotic) and back into the dissolvedinorganic form many times while being transported downstream The number ofcycles that may occur within a given stream reach depends on the spiraling length,which is the distance traveled when completing a cycle (Newbold et al 1981; StreamSolute Workshop 1990) The spiraling length is composed of two basic parts: Puptake length and P turnover length (Stream Solute Workshop 1990) The uptake

length, S w, is the average distance dissolved inorganic P, PO4, travels downstreambefore it is removed from the water column through various abiotic and biotic process

that occur within a stream channel (Newbold et al 1981) The turnover length, S p,

is the distance traveled in various particulate forms before P is returned to the watercolumn in the dissolved inorganic form (Newbold 1992) The movement of P through

a stream reach is tightly coupled with the downstream transport of water, and each

P cycle begins downstream from the next, producing a spiral through the streamecosystem The use of the spiraling concept and short-term solute injections havebeen increasingly used to estimate P retention efficiency in streams

5.3.1 D ETERMINING P HOSPHORUS S PIRALING

Stream Solute Workshop (1990), Newbold (1992), and Webster and Ehrman (1996)provide valuable guidance on solute dynamics in streams and the experimentalmethods to estimate P uptake length Short-term solute injections of P and a con-servative (hydrologic) tracer are used to estimate P uptake length in a stream reachwhere PO4 concentrations downstream from the injection point typically exhibit anexponential decline (Figure 5.2) The solute injection should be at a constant rateduring storm events (see, e.g., Green and Haggard 2001; Novak et al 2003) Phosphorus

FIGURE 5.2 Dissolved P concentrations in background water samples and water samples

collected during a solute injection experiment on September 2, 1999, at Willow Branch, Oklahoma, with a graphical display of P uptake length calculations (Adapted from B.E.

Haggard and D.E Storm, J Freshwater Ecol 18, 557–565, 2003 With permission.)

Solute injection Background

Distance from first sampling site downstream of injection point (m)

and should last long enough for the hydrologic tracer, such as chlorine (Cl), to reach

a steady concentration at the most downstream end of the stream reach; conductivitymight be used to measure when the steady state is reached downstream Multiplewater samples are collected from sites at increasing distances downstream from theinjection point, and P concentrations resulting from the injection are corrected forbackground concentrations at each site and then for losses due to dilution using ahydrologic tracer (see, e.g., Martí and Sabater 1996) The proportion of P remaining

in the water column is used in a negative exponential relation to estimate the P

uptake rate per unit distance coefficient, k:

where Px is the dissolved P concentration (mg L–1) at a distance x downstream corrected for dilution and background concentration, Po is the dissolved P concentration (mg L−1)corrected for background concentration at the first site downstream from the injection

point, and k is the P uptake rate per unit distance coefficient (m−1) or slope of the linear

relationship in Equation 5.2 The P uptake length, Sw (m) is the inverse of the P uptake

rate per unit distance coefficient (1/k) Phosphorus uptake length has been measured

using radiotracers 32PO4 and 33

4

et al 1981, 1983) and stable PO4 additions (see, e.g., Butturini and Sabater 1998;1996; Mulholland et al 1990; Niyogi et al 2004; Valett et al 2002) Phosphorus uptakelength generally constitutes greater than 90% of the spiraling length in rather pristinestreams (Mulholland et al 1990; Newbold et al 1983), and uptake length has beenused as an indicator of stream P retention efficiency The calculation of P turnover

length, Sp, requires the use of radiotracers and cannot be accomplished with typicalshort-term solute injections using stable PO4

Because the downstream transport of added PO4 is influenced by stream water

velocity, P uptake length, Sw, is often strongly correlated to stream discharge orwater velocity (see, e.g., Butturini and Sabater 1998; Haggard et al 2001b; Niyogi

et al 2004) Davis and Minshall (1999) suggested that the mass transfer coefficient,

vf, should be used to compare the retention efficiency of different stream reachesand streams; this parameter would help identify underlying abiotic and biotic pro-

cesses influencing P retention The mass transfer coefficient, vf (m s−1), is the verticalvelocity at which dissolved P is removed from the water column by abiotic andbiotic processes occurring within the stream channel (Stream Solute Workshop1990); the mass transfer coefficient is the P uptake velocity within a stream reach.The mass transfer coefficient is a function of the uptake length, average water depthand average water velocity as

PO (see, e.g., Mulholland et al 1985, 1990; NewboldDavis and Minshall 1999; Haggard et al 2001b; Macrae et al 2003; Martí and Sabater

where h is the average water depth (m), u is the average water velocity (m s−1),

Q is the stream discharge (m3 s−1), and w is the average stream width (m) Thus, the

mass transfer coefficient reduces the variability observed in P retention associatedwith stream-water velocity and depth (see, e.g., Davis and Minshall 1999) Doyle

et al (2003) developed the following equation to examine the influence of in-stream

processes (i.e., vf) and hydrogeomorphology (i.e., h and u) on P retention:

where L is the reach length (m) Equation 5.4 is based on mathematical substitution

in the preceding equations in this chapter This approach allowed Doyle et al (2003)

to determine whether a stream reach would be retentive of dissolved P and whetherchanges in channel form and uptake processes would alter dissolved P retention

Hydrogeomorphology (i.e., h and u) will vary spatially along the longitudinal

down-stream gradient and temporally with changes in channel form and discharge of astream (Doyle et al 2003)

The P uptake rate per unit area, U (mg m−2 s−1) may also be calculated as

U = (PbQ)/(Sww) = vfPb (5.5)

where Pb is the average background P concentration (mg L−1) measured before theshort-term solute injection Phosphorus uptake rates are the product of the masstransfer coefficient and ambient dissolved P concentrations, and uptake rates willincrease with increases in ambient concentrations Uptake rates may increase linearlywith ambient concentrations until a threshold concentration is achieved where abioticand biotic processes are saturated (Haggard et al 2005; Niyogi et al 2004).The aforementioned citations would provide the range of P uptake lengths,velocities, and rates However, the greatest variability would generally be observed

in P uptake length and the least amount in P uptake rates Short-term solute additions,such as those used to estimate these parameters, measure gross nutrient uptake andassume that the injection duration is short enough to avoid regeneration, that uptakeprocesses are not saturated by the level of enrichment, and that the change inconcentration through a reach follows an exponential decay (Webster and Erhman1996) Several studies have shown that P uptake length measured using stable Padditions increases as the level of enrichment increases (i.e., ∆PO4↑) (Haggard andStorm 2003; Hart et al 1992; Mulholland et al 1990); the uptake velocity and ratewould decrease with increasing levels of enrichment from short-term additions.Despite some constraints in short-term solute injections, this methodology provides

an extremely valuable means incorporating both abiotic and biotic processes toevaluate whole-reach measures of P uptake processes within the stream channel.Recently, these parameters have been measured in stream reaches where a naturallongitudinal gradient in dissolved P exists (Haggard et al 2001a, 2005; Martí et al.2004; Merseburger et al 2005) Within this context and experimental situation, net

uptake length, Sw-net, net uptake velocity, vf-net, and net uptake rate, Unet, have becomethe usual terminology The parameters measure the net retention efficiency becausethe upstream P sources are continuous (e.g., effluent discharges or other point

sources) and because dissolved P would be retained from and released into the watercolumn continuously by abiotic and biotic processes Reported net uptake lengths,

rates, U, measured in less disturbed stream ecosystems, whereas net uptake velocities,

al 2005; Merseburger et al 2005) These experimental data might be useful inmodeling P retention and release through a stream reach where a significant effluentdischarge or other point source exists

5.3.2 S TREAM P ROPERTIES AND P HOSPHORUS S PIRALING

The concept of P spiraling, or the distance traveled downstream by one P molecule

as it completes one cycle of uptake and transformations from dissolved to organicforms and back into flow, reveals significant information about the degree to which

P changes during transport in rivers (Elwood et al 1983) Lengths of P spiralingvary from 1 to 1000 m as a function of flow regime, season, bedrock geology, andsediment characteristics (Melack 1995; Munn and Meyer 1990) Similarly, interac-tion of ground water with stream flow within the hyporheic zone can cause increases

or decreases in P concentrations depending on stream-bed upwelling or infiltration

of P-rich stream flow

The first definitive measurement of P spiraling length was reported for a order woodland stream in Tennessee using 32PO4 as the tracer The P spiraling lengthwas 190 m, 165 m of which was in the water, whereas the remainder was in fineparticulate organic matter Other North American workers found that spiraling lengthranged from 23 m in November to 99 m in August when the concentration of coarseparticulate material was less and P was moving largely in dissolved form (Mulholland

first-et al 1985) However, during storms the distance traveled by P in particulate materialcan increase by one or two orders of magnitude above typical uptake lengths (Melack1995) Differences in geology can have a profound effect on P spiraling length.Munn and Meyer (1990) found that a stream with granite bedrock had a spiralinglength of 85 m, whereas in a stream with P-rich volcanic bedrock the spiraling lengthwas 687 m

Factors influencing the P retention and spiraling are notoriously variable along

a stream reach For example, where in-stream geomorphic processes cause sizesorting or where sediments are enriched with P due to local contributions of P-richoverland flow, sediments can represent a significant source of P to stream flow, evenwhen inputs from runoff have ceased Edge-of-field riparian management not onlyimpacts overland flow P removal but also has a strong influence on stream P spiraling(Cooper et al 1987; Hearne and Howard-Williams 1988) Fencing off pastures fromgrazing allows palatable aquatic macrophytes to flourish and decreases P spiralinglength, whereas riparian afforestation can shade out periphyton and macrophytes,thus increasing spiral length The placement of riparian afforestation within a catch-ment is therefore expected to have a major influence on downstream surface water

quality Quinn et al (1993) reported that forested (Pinus radiata) riparian areas at

the headwaters of grassland-dominated catchments in New Zealand adverselyinfluenced downstream P water quality: total phosphorus (TP) concentrations from

grazed and riparian forested catchments of 0.488 and 1.195 mg L−1, respectively.This was attributed to a lack of ground cover within the riparian zone Quinn et al.(1993) suggested that tree planting should be sufficiently sparse to allow develop-ment of vegetative ground cover Fennessy and Cronk (1997) concluded that bufferstrips should be located along headwater reaches where most catchment wateroriginates and that storing water high in the catchment decreases downstream ero-sion In addition, streams and riparian areas in headwaters tend to be narrower thandownstream channels, taking less land out of production while maximizing nutrientretention and removal compared to the targeting of larger channels downstream(Fennessy and Cronk 1997).

5.4 ALGAL AND MICROBIAL PROCESSES

The rate at which algae and other microbial organisms take up and release dissolved

P in the water column represents an important component of P retention withinstream reaches, especially during base-flow conditions A Michaelis–Menton rela-tionship is often used to describe dissolved P uptake as a function of dissolved Pconcentrations, where P uptake becomes saturated as dissolved P concentrationsincrease Several studies have shown that algal and microbial P uptake can becomesaturated at relatively low dissolved P concentrations ( 0.01 mg PO4–P L−1) (Bothwell1985; Mulholland et al 1990), and this aspect of biotic uptake represents a constraint

in typical solute-injection experiments using stable PO4 additions The level of Penrichment required to measure differences in dissolved P concentrations betweensequential monitoring sites downstream from the injection point often exceeds theconcentration where algal and microbial uptake would typically be saturated.Furthermore, typical streams draining agricultural catchments would have dissolved

P concentrations that are much greater than 0.01 mg L−1 (see, e.g., Haggard et al.2001b; Macrae et al 2003; McDowell et al 2001; Smith et al 2005), and effluentdominated streams would have dissolved P concentrations that are one to severalorders of magnitude greater than 0.01 mg L−1 (see, e.g., Ekka et al 2006; Haggard

et al 2005; Martí et al 2004; Merseberger et al 2005) It is likely that saturationwill influence algal and microbial uptake kinetics in most agricultural and urbancatchments where catchment-scale P transport is being simulated by process-basedmodels

Though cellular uptake and growth rates are generally saturated at low trations, maximum biomass accrual in streams often occurs at somewhat greaterconcentrations (0.015 to 0.050 mg PO4–P L−1) (Bothwell 1989; Horner et al 1983;Popova et al 2006) This range of dissolved P concentrations might be more typical

concen-of streams draining agricultural catchments, and therefore algal and microbial uptakelikely still plays a significant role in dissolved P retention through stream reaches.However, the importance of algal and microbial P uptake will vary spatially andtemporally with dissolved P concentrations in streams and with the uptake kinetics

of the algal and microbial community Dissolved P uptake rates of algae will varywith light, water velocity, temperature, grazing, and time following disturbanceswithin the stream channel (see, e.g., Dodds 2003; Horner and Welch 1981; Mulholland

et al 1994)

To determine the importance of algal and microbial uptake within individualstreams, some relatively simple experiments could be performed to quantify Plimitation and potential uptake rates Phosphorus limitation of algae in streams isabout as likely to occur as nitrogen (N) limitation (Francoeur 2001), although Plimitation in streams may be more widespread in some agricultural catchmentsbecause of the mobility of nitrate (NO3) Several citations exist in the literature thatdescribe P limitation experimental protocols using nutrient diffusing substrata (see,e.g., Allen and Hershey 1996; Matlock et al 1998; Stanley et al 1990); theseexperiments are usually relatively short in duration (i.e., approximately 10 to 14 days)and inexpensive when measuring chlorophyll a as the response variable on thetreatments Potential P uptake rates from algae and other microbial organisms inthe benthic substrate can be measured using stable PO4 or radio-labeled P; Steinmanand Mulholland (1996) provided some guidance on simple experimental proce-dures to estimate P uptake rates The use of radio-labeled P in these experimentswould allow phosphorus turnover or release rates of the stream algae to be mea-sured as well.

5.5 STREAM SEDIMENTS AND PHOSPHORUS

Stream sediments can act as either a P sink or source for dissolved P in the watercolumn, depending on several abiotic and biotic processes that basically influencestream sediment equilibrium P concentration (i.e., EPC0) In streams (i.e., fluvialsystems with good hydraulic mixing), the availability of P in benthic sediments may

be estimated using EPC0 at zero net sorption or desorption (Froelich 1988; Sharpley

et al 2002) A quasi-equilibrium for dissolved P concentration exists between streamsediments and the water column, where sorption and release rates of dissolved Pare virtually equivalent Stream sediments may have a major influence on dissolved

P concentrations and P retention, especially during base-flow conditions (see, e.g.,Hart et al 1992; Hill 1982; House and Warwick 1998; Klotz 1988; McDowell andSharpley 2003; Meyer 1979) In theory, P will desorb from stream sediments if thewater column dissolved P concentration is less than the sediment EPC0, or alterna-tively, P will sorb to stream sediments if the water-column dissolved P concentration

is greater than the sediment EPC0 (Taylor and Kunishi 1971) However, dissolved

P concentrations in streams are not solely controlled through sediment sorption anddesorption processes, and the relative importance of biotic processes may alter thisequilibrium between the water column and stream sediments

5.5.1 S EDIMENT S OURCE E FFECTS

Sediments within the fluvial system are derived either from overland flow or streambank erosion Sediments derived from stream banks will largely consist of subsoilthat is relatively depleted of P The proportion of sediment within the fluvial systemwill depend on the age of the channel network For example, in areas with recentgully formation (channel rejuvenation), subsoil material will dominate (Olley et al.1993) Furthermore, P derived from subsoil materials in these systems will be lessreadily available to be released to water, and the subsoils will likely represent a net

sink for P (McDowell and Sharpley 2001) As a result of the erosion of subsoils,which are often dominated by silt-sized particles, the predominant form of P transport

in these fluvial systems is PP, whereas in sandy catchments most P is transported

in dissolved form (Baldwin et al 2002)

In a much larger catchment, McDowell et al (2002) examined the processescontrolling sediment P release to the Winooski River, Vermont, the largest tributary

to Lake Champlain (Figure 5.3) Iron-oxide strip P (algal-available P) of the river

FIGURE 5.3 The impact of land use and physical transport processes in P in fluvial sediments

with the Winooski River watershed, Vermont (Adapted from R.W McDowell, A.N Sharpley,

and A.T Chalmers, Ecol Eng 18, 477–487, 2002 With permission.)

0 10 20 30 40

–1 )

Before confluence

runoff from dairy farm

All other sediments

Kilometers

100 80 60 40 20

10

sediments adjacent to agricultural land (3.6 mg kg−1) was significantly greater (p <

.05) than that of sediments adjacent to forested land (2.4 mg kg−1) This was especiallythe case in the tributary adjacent to a dairy farm where overland flow was directlycontributing to the enrichment of fluvial sediment Notably, impoundment (731 mg

kg−1) and reservoir sediments (803 mg kg−1) had greater TP concentrations than riversediments (462 mg kg−1) This was attributed to more fines (< 63 µm) in impound-ments and reservoirs than in river sediments Furthermore, turbulence at the conflu-ence of two tributaries resulted in the shift of the particle size toward coarser particles.This also increased the release rate of these sediments and their sensitivity to incom-ing P sources Consequently, impoundment and reservoir sediments had lower abil-ities to release P to solution in the short term, thereby acting as sinks for sedimentsrich in P provided the system was maintained in an aerobic state However, ifsediments, like soils, become anaerobic through oxygen depletion and stratification,

P release increases via iron and sulphate reducing bacteria (Boström et al 1988;Lovley et al 1991)

The results of the research on the Winooski River catchment demonstrate thatfluvial hydraulics has a strong influence on the properties of sediment within riversystems The input and delivery of fine sediment enriched with P was influenced byadjacent land use The fluvial sediment, particularly at the outflow of the river intoLake Champlain, represents a P storage pool, which has long-term potential torelease a large amount of P to overlying waters In the short term, however, riverflow and the physical properties of the sediments will influence the amount ofsediment P leaving the catchment Thus, in connecting sediment P loss from thelandscape to channel processes, variability in flow, local sources of P, and sedimentproperties must be taken into account, particularly near the point of impact Because

of these complexities, channel processes and changes in P forms and loads are notcurrently simulated in many models that estimate P loss from catchments (Hanrahan

et al 2001)

5.5.2 S EDIMENT AND E QUILIBRIUM P HOSPHORUS C ONCENTRATIONS

The first experiments evaluating sediment and soil EPC0 were conducted over

30 years ago (Taylor and Kunishi 1971), and the PO4 buffer mechanism in streams,wetlands, and estuaries has been intensively investigated since the 1980s (see, e.g.,Ekka et al 2006; Froelich 1988; Haggard et al 1999; House et al 1995; Klotz 1988;McDowell et al 2002; McDowell and Sharpley 2001) The methods to estimateEPC0 are fairly straightforward, where sediments are mixed with solutions containingdiffering initial PO4 concentrations and then P sorbed (mg P kg−1 dry sediment) isplotted as a function of either initial or final dissolved P concentration (mg P L−1)

in the aqueous solution (Figure 5.4) (Froelich 1988; Haggard et al 2004; Klotz1988; Taylor and Kunishi 1971) However, several experimental variables can sig-nificantly influence the sediment EPC0 determined in the laboratory These equili-bration experiments would be best performed on fresh, wet sediments because dryingsediments can alter the P sorption characteristics where sediment has a lesser Pbuffering capacity (Baldwin 1996) Another important consideration is that thechemical composition of the water used in the extraction does influence the estimation