Stream Nutrient Retention Efficiency at an Enriched System in the Eucha – Spavinaw Basin B.E. Haggard 1, E.H. Stanley 2, D.E. Storm 3 1 USDA – ARS Poultry Production and Product Safety Research Unit, Fayetteville, AR, USA 2 University of Wisconsin, Center of Limnology, Madison, WI, USA 3Oklahoma State University, Biosystems Engineering Department, Stillwater, OK, USA Introduction In the last decade, the effect of application of animal manure to pastures on nutrient concentrations and fluxes in surface runoff have been the focus of scientific investigation in the Ozark Plateaus, USA. Nutrient concentrations and fluxes in Ozark streams often increase with the proportion of pasture or agricultural land use in the catchment (Haggard et al. 2003a, Petersen et al. 1999), and these observations have overshadowed the potential impacts of municipal wastewater treatment plants (WWTPs) inputs on aquatic systems. Despite nationwide efforts to reduce WWTP nutrient inputs, WWTPs continue to pose a threat to regional water quality. Wastewater treatment plants and other point sources contribute almost 50% of the total nutrient load to some aquatic systems in this region (Haggard et al. 2003b). The effects of WWTPs on stream nutrient concentrations and cycling are substantial; the distance required to temporarily retain over 63% of WWTP nutrient inputs has been reported to be as long as 30 km (Haggard et al. 2001a, Martí et al. 2004). Effluents can cause sediment deoxygenation several kilometers downstream (Rutherford et al. 1991) and decreases in the phosphorus (P) buffering capacity of benthic sediments (Dorioz et al. 1998, House and Denison 1998, Haggard et al. 2001a). In short, WWTPs can have significant effects not only on nutrient loads, but on nutrient transformations and general limnological conditions in the stream that may persist for several km. Unfortunately, we have limited measures of nutrient uptake using spiraling methods (Stream Solute Workshop 1990) in highly enriched systems, especially in streams receiving nutrient enriched WWTP effluent. Study Site Description The focus of our study was Columbia Hollow, a 3rd order tributary to Spavinaw Creek in the Ozark Plateau of northwest Arkansas, USA (Fig 1). Spavinaw Creek drains into Lakes Eucha and Spavinaw, which are drinking water supply reservoirs in northeast Oklahoma. This reservoir system is the primary municipal water supply to the City of Jay, Oklahoma, and surrounding communities; it also supplies over half of the municipal drinking water to Tulsa, Oklahoma. Columbia Hollow receives effluent from a municipal WWTP with secondary treatment in the City of Decatur, Arkansas. This facility receives wastewater from a poultry processing plant and a residential population of approximately 1,000 inhabitants. Mean discharge from the Decatur WWTP is ca. 5,000 m 3 /d (60 L/s), and the effluent has limits of 15 and 10 mg/L on ammonium (NH 4 -N) and nitrate (NO 3 -N), respectively. Currently, no regulations exist for P. The Decatur WWTP contributes almost 25% of the annual P load to Lake Eucha (Storm et al. 2002) and is approximately 9 km from Spavinaw Creek and over 30 km from Lake Eucha following the stream channels. Within the study reach, Columbia Hollow is a typical Ozark Mountain stream with a chert gravel bed and karst topography in the upland areas. The catchment area above the most downstream site was 18 km 2, with 73% of the land used as agriculture and pasture. Field Methods We sampled 7 sites at Columbia Hollow almost monthly from June 1999 through February We estimated discharge at each site from depth and velocity measurements at approximately 0.3 m intervals across a transect perpendicular to streamflow. At each site, we measured electrical conductivity, temperature, and pH at a single point; and surface water samples were collected at 3 points perpendicular to streamflow. Water samples were filtered immediately on site using 0.7  m pore diameter Whatman GF/F glass fiber filters and acidified with H 2 SO 4 to pH <2. Water samples were put on ice, stored in the dark, and analyzed within 48 h of collection. Upon return to the laboratory, dissolved inorganic nutrient (SRP, NH 4 -N and NO 3 -N) and chloride (Cl-) concentrations were determined. Study Objectives In a stream receiving municipal WWTP effluent discharge, we evaluated net nutrient retention efficiency using metrics from the nutrient spiraling concepts and calculations of nutrient transformation. Specifically, we evaluated: 1)The ability of the study reach to retain soluble reactive phosphorus (SRP) 2)The nitrification and uptake of ammonium (NH 4 ) within the study reach 3)The ability of the study reach to retain nitrate (NO 3 ) Table 1: Average monthly TP yields from 2001 and 2002 for the War Eagle Creek SWAT model (method 1) and the SWAT model linked to the QUAL2E model (method 2) and respective measured values Fig 1: Map of the Columbia Hollow catchment (GPS: Lat , Long ), Eucha-Spavinaw Basin and water-quality sampling sites. Nutrient Spiraling Calculations We used nutrient spiraling methods (Stream Solute Workshop 1990) to estimate transport and retention of nutrients from the WWTP. Short-term solute additions are often used to estimate the nutrient uptake length (Sw), the average distance a nutrient molecule travels downstream in the dissolved form before uptake by benthic abiotic and biotic processes. Nutrient uptake length is an indicator of stream nutrient retention efficiency; the shorter the length the greater the efficiency and vice versa. We used this same principle of determining downstream declines in nutrient concentrations to estimate a whole-reach measure of nutrient retention. However, because the nutrient addition from the WWTP was substantial (see Results) and continuous, the observed longitudinal pattern in nutrient concentrations downstream from continuous WWTP inputs is the net result of nutrient uptake and release processes. Therefore, we used downstream declines in nutrient concentrations to estimate a different parameter: the net nutrient uptake length (S net ; Haggard et al. 2001a). Changes in the dilution-corrected nutrient concentration downstream from the WWTP were used to estimate S net on each sampling date using the following equations: C x = C o e -kx ln(C x /C o ) = -kx S net = -1/k where C x is the dilution-corrected nutrient concentration at distance x (km) from the WWTP (mg/L), C o is the nutrient concentration at the most upstream site below the WWTP (mg/L), and k is the nutrient change coefficient (1/m). Simple linear regression was used to determine if the relation between dilution-corrected nutrient concentrations and distance downstream (i.e., S net ) was significant at  = The mass transfer coefficient (v f ) is the vertical velocity at which a nutrient molecule travels from the water column to the stream substrate (Stream Solute Workshop 1990). The parameter is related to S net through average depth and velocity: v f -net = h·u/S net = Q/(S net ·w) where h is the average water depth (m), u is the average water velocity (m/s), Q is the average discharge (m3/s), and w is the average stream width (m). Net nutrient uptake rate (U net, mg m -2 s -1 ) of nutrients added by the WWTP can also be calculated from S net as follows: U net = (C o ·Q)/(S net ·w) = v f-net ·C o Because in this study we calculate S net, we have re-defined these terms as net mass transfer coefficient (v f-net ) and net nutrient uptake rate (U net ) for this study. We estimated the level of nutrient addition (i.e.,  SRP,  DIN,  NO 3 -N and  NH 4 -N) as the difference between concentrations upstream and downstream from the WWTP. Linear nitrification rate (K NIT, 1/m) was modeled using equations and methods described by Bernhardt et al. (2002) in which K NIT was determined by fitting the model to longitudinal changes in NO 3 -N concentration. Nitrate uptake per m (K NO 3 -N ) and K NIT were simultaneously solved for using MS Excel Solver and minimizing the sum of squares between predicted and observed NO 3 -N concentration. MS Excel Solver was constrained so that K NIT, K NO 3 -N and NH 4 -N uptake per m (K NH 4 -N ) were greater than or equal to zero. The fraction of NH 4 -N nitrified and modeled nitrification rate (g m -2 d -1 ) were estimated using the following equations: %NH 4 -N Nitrified = k NIT /k NH 4 -N Modeled Nitrification Rate = (k NIT /k NH 4 -N )·U net-NH 4 -N Results Soluble Reactive Phosphorus Concentration and Retention The Decatur WWTP effluent discharge increased downstream SRP concentration (p<0.001), resulting in up to 50 fold increase. The greatest SRP concentrations was 9.9 mg/L in February SRP concentration decreased with increasing distance from the WWTP input, but concentrations 3 km downstream were still 30 fold greater than at the upstream site. S net -SRP lengths are substantially greater than that observed in other streams in the basin that drain forested or agricultural catchments (0.2–0.9;Haggard et al. 2001b). U net -SRP was high at Columbia Hollow, while the demand (v f-net ) for this nutrient relative to its supply was low compared to unenriched streams. A P exchange mechanism existed in the study reach that apparently helped maintain high ambient SRP concentrations when WWTP inputs were low (  SRP <3 mg/L) and may be related to benthic sediment equilibrium P concentrations. Ammonium Transformation and Retention NH 4 -N concentrations below the WWTP were 10–100 fold greater than upstream concentrations (p<0.01). NH 4 -N concentration showed a consistent longitudinal decrease along the study reach. At the most downstream site, NH 4 -N concentration declined to <1 mg/L and was <0.1 mg/L on most sampling dates. The longitudinal decrease in NH 4 was significant, and S net ranged from 0.4–1.4 km. Modeled nitrification rates were 7–31 g NO 3 -N m -2 d -1 and several orders of magnitude greater than rates in prairie and agricultural stream sediments (Kemp and Dodds, 2002) or a desert stream with extensive hyporheic sediments (Jones et al., 1995). Modeled nitrification often exceeded NH 4 uptake suggesting additional reduced N sources were nitrified in addition to NH 4 inputs from the WWTP. Apparent retention of NH 4 was counter balanced by increases in NO 3, indicating overall retention of DIN was minimal on most sampling dates. Nitrate Concentrations and Retention DIN was dominated by NO 3 upstream from the WWTP, ranging from 3.0–5.5 mg/L. In contrast to NH 4, NO 3 -N concentrations gradually increased downstream of the WWTP on all sampling dates. Downstream changes in NO 3 resulted in negative S net -NO 3 -N, v f-net and U net. NH 4 -N concentration showed a consistent longitudinal decrease along the study reach. The transformation of NH 4 to NO 3 made Columbia Hollow a significant downstream source of NO 3, as was the case for SRP.SUMMARY Point source nutrient loading to Columbia Hollow has created an unusual mixture of biogeochemical characteristics that includes high areal uptake rates that contrast with extremely low v f-net for nutrients, a P exchange mechanism that apparently helped maintain high ambient P concentrations, and distorted N cycling in which absolute and relative rates of transformation are substantially different from those in unenriched streams. Thus, rather than acting as a nutrient sink, Columbia Hollow appears to act more as a short-term storage zone for added P and a transformer of added N. Limited retention emphasizes that point source inputs may have long-term and large-scale effects on water quality (see also Mart  et al. 2004), and that the key to managing point sources will be to understand the biogeochemical distortions created by focused N and P loading and the circumstances that foster a return to more normal nutrient cycling conditions. References cited in text are available upon request, and a more comprehensive manuscript has been submitted to the Journal of the North American Benthological Society. Corresponding author: Brian Haggard, Research Hydrologist, USDA – ARS PPPSRU, 203 Engineering Hall, Fayetteville, AR USA; tele: ; fax: ;