Exchange of Inorganic Phosphate between River Waters and Bed

Sep 7, 2002 - The kinetics of the release of Soluble Reactive Phosphorus (SRP) in oxic and anoxic conditions and uptake in oxic conditions by contamin...
23 downloads 19 Views 137KB Size
Environ. Sci. Technol. 2002, 36, 4295-4301

Exchange of Inorganic Phosphate between River Waters and Bed-Sediments WILLIAM A. HOUSE* AND FRANK H. DENISON Centre for Ecology and Hydrology, Winfrith Technology Centre, Dorchester, Dorset, DT2 8ZD, UK

models on phosphorus transport (12). However, there is a poor understanding of the underlying mechanisms and the reasons that the application of the EPC0 method has been successful in such dynamic systems. The present work uses a fluvarium channel to study the release (in oxic and anoxic conditions) and uptake kinetics of SRP from bed-sediments from a contaminated river and applies empirical models utilizing the EPC0 concept. A new mechanistic model is developed to provide a better understanding of the underlying processes that determine SRP fluxes.

Experimental Section The kinetics of the release of Soluble Reactive Phosphorus (SRP) in oxic and anoxic conditions and uptake in oxic conditions by contaminated river sediments (River Blackwater in Southern England) were measured using a fluvarium channel operated to mimic environmental conditions. Release rates (from 1 to 10 nmol m-2 s-1) and uptake rates were modeled successfully using a Parabolic equation and Diffuse Boundary Layer model. A SRP release experiment over 61 days showed that large gradients in SRP developed in the porewater as a result of diffusion, sorption, and insitu generation of SRP in the anoxic zone. This was modeled using a new Triple Zone Model that incorporated diffusion through a liquid boundary layer and sorption/desorption in oxic and anoxic zones. The results highlighted the importance of the oxic zone in controlling the exchange of SRP between the sediment and water column. The model was also applied to explain why the Equilibrium Phosphate Concentration (EPC0) of the sediment measured in oxic conditions was constant (and equal to the value calculated from sorption isotherm measurements) during 2-day release experiments, and also why it increased in the uptake experiments. Measurements in anoxic conditions showed the importance of the sediment temperature in controlling the flux at the interface.

Introduction The internal cycling of Soluble Reactive Phosphorus (SRP), a biologically available form of the nutrient in lakes and rivers, is known to be important (1-3). A major process in this cycling is the interaction of SRP with sediments, and, particularly in rivers where the ratio of bed area to water volume is relatively high, interactions with bed-sediments exert an important influence (3-7). Although much research on benthic fluxes has been reported on lakes, there are few published data relating to rivers. This is surprising as the majority of treated sewage effluents are discharged to rivers. In particular, the sediments downstream of sewage inflows are generally rich in phosphorus as a result of their historical and current exposure to SRP (8-10). For this reason, there is an urgent need to understand the mechanisms controlling the transport of SRP across the sediment-water interface. Previous work has highlighted the usefulness of the Equilibrium Phosphate Concentration (EPC0) when applied to understanding the direction and size of the SRP flux (811), and this concept has been applied in catchment based * Corresponding author phone: +44 (0) 1305 213634; fax: +44 (0) 1305 213600; e-mail: [email protected]. 10.1021/es020039z CCC: $22.00 Published on Web 09/07/2002

 2002 American Chemical Society

Field Site Visits. Thirteen rivers in England were surveyed to examine trends in SRP downstream of treated sewage inputs. The River Blackwater, a lowland river in southern England, was selected for further study because of its relatively slow flow and accumulation of fine sediments on the bed, a clear trend of reduction in SRP downstream of sewage works in the absence of other inflows, the high degree of impact from several sewage treatment works, and the planned reduction of phosphate concentrations in effluents from the larger sewage works, i.e., over 10 000 PE (People Equivalent). It is situated on the county boundary between Surrey and Hampshire, south of London. All sediments were taken from one site at Aldershot (NGR SU 885538) downstream of a sewage input. The site was visited seasonally on eight occasions between January 1995 and November 1996. On each visit, water samples were taken for the analysis of phosphorus fractions and suspended sediments. Flow was measured using an ultrasonic sensor, and on-site measurements of pH, dissolved oxygen (DO), and temperatures were made using standard procedures (9). Surface sediments (5 cm depth) were sampled into four stainless steel trays (40 × 10 × 5 cm) when possible and otherwise collected using a fine meshed net. The trays were immediately transported back to the fluvarium channel for flux measurements. Chemical Measurements. Total phosphorus concentrations in the sediments were determined by an ignition and digestion (13). The concentrations of total calcium and iron in the digest were measured by atomic absorption (Unicam 929). Total organic matter (OM) was determined by ignition at 550 °C for 17 h. The particle size fractionation was by wet sieving (63 µm to 2 mm) and the pipet method described by Gee and Bauder (14) for particles < 63 µm in size, with bands of 1-2, 0.5-1, 0.25-0.5, 0.125-0.25, 0.063-0.125 mm and 20-63, 2-20, and d0

dc(x)/dt ) Ds[d2c(x)/dx2]/(1 + Kd/f) + k0

(11)

where Kd/f is the distribution coefficient for the anoxic part of the sediment, i.e., f > 1, and k0 is the zero-order rate constant describing the generation of phosphate. A zeroorder rate law is assumed as the simplest approximation as the rate is likely to depend on microbial biomass and mineral content involved in reactions that lead to the release of SRP

TABLE 1. Results from the Field-Sampling Program Including the Field Measurements, Analysis of the Surface Sediments and Sorption Isotherm Measurementsa river water sample

date

A B C D E F G H

30.1.95 10.4.95 21.7.95 23.10.95 17.1.96 18.4.96 27.8.96 6.11.96

flow/ m3 s-1

temp/ oC

1.08 0.24 0.11

8.8 14.2 18.1 13.1 9.2 11.3 17.6 13.3

0.34 0.19 0.16

bed-sediment

pH

DO/ %

SRP/ µM

TDP/ µM

TP/ µM

SS/ mg dm-3

OM/ %

TPsol/ µmol g-1

TFe/ µmol g-1

Tca/ µmol g-1

EPC0/ µM

Kd/ dm3 kg-1

6.91 7.68 7.21 7.36 7.45 7.50 7.19 7.40

79 153 90 70 67 82 75 53

20 71 133 126 39 93 87 48

21 93 161 126 39 93 93 51

42 119 180 120 60 103 104 61

88 14 7 2 20 7 7 13

1.6 5.4 1.5 6.4 4.9 1.6 12 11

14 17 1.7 91 86 33 475 280

99 439 188 300 254 132 807 1598

90 168 43 3928 220 99 212 443

6.4 3.1 0.2 1.9 3.0 7.3 19.7 17.6

117 929 2418 724 144 608 572

a Key. DO: dissolved oxygen; SRP: soluble reactive phosphorus; TDP: total dissolved phosphorus; TP: total phosphorus; TP sol, TFe, and TCa are the phosphorus, iron, and calcium contents of the acid digested sediments, respectively; EPC0: equilibrium phosphate concentration; Kd: distribution coefficient describing the sorption affinity of the sediment to SRP; SS: suspended solids; OM: organic matter.

into the porewater. A limitation of the model is that the constant describing the generation on SRP in the anoxic zone is likely to incorporate several abiotic and biochemical processes and so must be determined by fitting the model results to porewater concentrations determined experimentally. In the above equations, Dm ) 7.34 × 10-10 m2 s-1 at 25 °C (23) and was corrected to other temperatures using the relationship Dmη(T)/T which is a constant between 5 and 25 °C (η(T) is the viscosity of water at temperature T). The differential equations were solved numerically using the method of lines with the solution represented by cubic Hermite polynomials (24). The initial conditions for the longer-term fluvarium experiment were that the concentration of SRP in the DBL was equal to the concentration in the bulk solution at the start, the concentration in the porewater of the oxic layer was equal to the EPC0 of the sediment, and the concentration in the porewater of the anoxic sediment was equal to the measured porewater concentration. The values of Kd and EPC0 for the oxic zone were calculated from the sorption isotherm. The only unknowns for a solution are a factor, f (a factor describing the reduced sorption affinity of the anoxic sediment compared with the oxic sediment), and the zero-order generation constant, k0. These were calculated by adjusting the values to obtain the best agreement with measured concentration profile in the anoxic zone. Essentially, k0 controls the asymptotic concentration in the deep sediment and f, the shape of the concentration profile as it approaches the asymptotic concentration. The other parameters, i.e., Kd, EPC0, φ, and δ, were found by independent measurements. The boundary condition at x ) 0 was calculated by fitting the measured SRP concentrations in the bulk solution using cubic splines and then generating 20 equidistant data points within the time interval.

Results and Discussion Background. This is a hard water river with dissolved calcium concentrations in the range of 1.6-2.4 mmol dm-3, total alkalinities of 2.6-3.8 mequiv dm-3 during the eight visits. Only once (Table 1) was the river in storm conditions, with a high suspended solids concentration and TP concentration about double the SRP but with similar TDP and SRP concentrations. TDP concentrations were similar to SRP showing that dissolved organic and inorganic polymeric phosphates were not important. However TP was generally greater than the SRP, particularly at the higher flows, indicating the importance of transport by suspended solids. The DO saturations were usually < 100% with the lowest value of 53% recorded in the autumn of 1996. An exception was in April 1995 when the recorded value was 153%. Sediment Analysis. The sediments varied in composition with no clear seasonal trends in the OM, TPsol, TFe, and TCa

FIGURE 1. Seasonal changes in the particle-size distribution of the sediment. contents. The sorption isotherms produced EPC0s less than 20 µmol dm-3 and generally much lower than the SRP concentration in the river water (Table 1). There was no significant correlation (CL 90%) between the EPC0s and the SRP concentrations. The finding that the EPC0s were less than the SRP concentrations on all occasions indicates that the sediments acted as a sink for SRP. The only significant correlation (CL 99%) with Kd was with the TCa content. The particle-size analysis (Figure 1) show that finer material settles from the water column during low-flow making the silt fraction an important part of the surface sediment in the summer. Generally the surface sediment ( 30 mm. An optimum value of 2.5 × 10-5 µmol dm-3 s-1 was obtained. The factor, f, relating the distribution coefficients in the oxic and anoxic zones of the sediments, was determined by fitting the curve between the oxic layer and the deeper sediment. A value of ≈20 produced the agreement shown in Figure 3b. The observed decrease in the SRP concentration at greatest depths (> 40 mm) probably resulted from the formation of vivianite in the sediment (25, 26), although others processes cannot be disregarded. There is evidence from XRD results from previous research with sediment collected from this site that vivianite forms up to 20% by mass of the sediment (25). To explore why the Parabolic and DBL models worked at short exposure times, assuming the EPC0 of the surface sediment was constant, the Triple Zone Model was used to generate concentration profiles after 0.1, 1, and 2 days using the parameters allocated above. The results are shown in Figure 4a together with the profile computed after 60 days. Initially, sorption in the oxic zone maintains the EPC0, and the loss of SRP from the sediment occurs by diffusion across the DBL. The oxic layer acts as a buffer zone reducing the flux of SRP from the anoxic zone to the overlying solution. After 2 days, the SRP supplied from the anoxic zone has penetrated further and led to an increase in SRP in the 4300

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 20, 2002

Effects of Reducing the Oxygen Concentration. All experiments showed a release of SRP from the sediment over 48 h (Figure 5). The rates at lower temperatures were significantly lower than at higher summer temperatures (Pearson’s regression coefficient r2 ) 0.94), and they varied between 0.002 nmol m-2 s-1 in the winter and 0.44 nmol m-2 s-1 in the summer. This temperature dependence is consistent with microbially mediated processes. To analyze the results in more detail information on changes that occur at the sediment-water interface during the oxic transition would be required. However, the Triple Zone Model provides an insight into the mechanisms that may be operating in these systems. At higher temperatures, the oxic zone is expected to be thinner as a result of the high sediment oxygen demand caused by microbial degradation of organic matter (27, 28). With the supply of oxygen from the water reduced, the oxic layer will be compressed and present less of a barrier for SRP transport from the deeper sediment. It is possible that the oxic zone disappears completely. During the colder months, the oxic zone expands due to reduced microbial activity, and, when the oxygen concentration in the water is reduced, the oxic layer remains longer preventing the more rapid loss of SRP (18, 28). There is also an increase in pH in the experiments with reduced oxygen (Table 2) that could result in enhanced release of SRP from the sediment as a result of desorption from acidic sites. It is also possible that the increase in sulfate concentration in the water following the oxidation of sodium sulfite could lead to enhanced phosphorus release from the sediment as a result of the interaction between microbially mediated sulfate reduction and iron oxide particles. Re-examination of the SRP Uptake Kinetics. Results from the phosphate uptake experiments showed that the EPC0(DBL) values calculated from the DBL model were higher than the experimental ones. In contrast to the results from the initial SRP release experiments, the measured EPC0s no longer represented porewater concentrations in the surface sediment. This was further investigated using the Triple Zone

Model and data from experiment A. Unfortunately no information on the concentration profile of SRP in the porewater was obtained so independent estimation of k0 and f was not possible. Other parameters were available from measurements, i.e., the DBL thickness, δ ) 0.35 mm, was assigned from flow measurements, temperature of 10 °C, EPC0 ) 6.4 µmol dm-3, and Kd ) 117 dm3 kg-1. The firstorder rate constant and factor, f, were assigned as above, and the concentration profiles of porewater SRP were generated after 0.2, 1, and 2 days. The results in Figure 4b illustrate how diffusion through the DBL leads to increased concentrations of SRP in the porewater and to a surface EPC0 of ≈24 µmol dm-3 which is close to the value obtained from the DBL model (19.9 µmol dm-3). Development of the concentration profile in the DBL and surface sediment at < 2 mm is not influenced by the choice of the Kd for the anoxic sediment or k0. Increases in concentration at the surface will depend on the EPC0 of the oxic layer and the supply of SRP from both the anoxic zone and the overlying water. Further validation of the model is needed using different sediments with measurements of the oxic layer depth during the experiment and testing whether the simple diffusion/ sorption/desorption and SRP generation in the anoxic zone, described by zero-order kinetics, is adequate.

Acknowledgments We thank the Department of the Environment (DOE Grant EPG/1/9/09) and the Natural Environment Research Council for their financial support. I also thank Barbara Gainswin for her input.

Literature Cited (1) Holden, A. V. Verh. Internat. Verein. Limnol. 1961, XIV, 247251. (2) Holtan, H.; Lamp-Nielsen, L.; Stuanes, A. O. Hydrobiologia 1988, 170, 19-34. (3) Bowes, M.; House, W. A. Hydrol. Proc. 2001, 15, 261-280. (4) Johnson, A. H.; Bouldin, D. R.; Goyette, E. A.; Hedges, A. M. J. Environ. Qual. 1976, 5, 148-157. (5) Hill, A. R. Freshwat. Biol. 1982, 12, 293-304.

(6) House, W. A.; Denison, F. H. Water Res. 1998, 32, 1819-1830. (7) Dorioz, J. M.; Cassell, E. A.; Orand, A.; Eiseman, K. G. Hydrol. Process. 1998, 12, 285-309. (8) Richardson, C. J.; Vaithiyanathan, P. Soil Sci. Soc. Am. J. 1995, 59, 1782-1788. (9) House, W. A.; Denison, F. H. Water Res. 1998, 32, 1819-1830. (10) House, W. A.; Denison, F. H. Sci. Tot. Environ. 1997, 205, 2549. (11) Reddy, K. R.; O Connor, G. A.; Gale, P. M. J. Environ. Qual. 1998, 27, 438-447. (12) Cooper, D. M.; House, W. A.; Reynolds, B.; Hughes, S.; May, L.; Gannon, B. Sci. Tot. Environ. 2001, in press. (13) Andersen J. M. Water Res. 1976, 10, 329-331. (14) Gee G. W.; Bauder J. W. Methods of Soil Analysis; Klute, A., Ed.; American Soc. Agronomy, Soil Science Soc. America: WI, 1986. (15) Murphy J.; Riley J. P. Anal. Chim. Acta 1962, 27: 31-36. (16) House, W. A.; Denison, F. H.; Smith, J. T.; Armitage, P. D. Environ. Pollut. 1995, 89, 263-271. (17) Wang, W.-C. Water Resour. Bull. 1974, 10, 662-671. (18) Slomp, C. P.; van Raaphorst, W. Chem. Geol. 1993, 107, 477480. (19) House, W. A.; Denison, F. H. Water Res. 2000, 34, 1187-1200. (20) Berner, R. A. Early Diagenesis; Princeton University Press: Princeton, NJ, 1980. (21) Krom, M. D.; Berner, R. A. Limnol. Oceanogr. 1980, 25, 797806. (22) Gunnars, A.; Blomquist, S. Biogeochemistry 1997, 37, 203-226. (23) Li, Y.-H. and Gregory, S. Geochim. Cosmochim. Acta 1974, 38, 703-714. (24) Sincovec R. F.; Madsen N. K. Math. Software 1975, 1, 232-260. (25) Woodruff, S. L.; House, W. A.; Callow, M. E.; Leadbeater, B. S. C. Freshwat. Biol. 1999, 41, 73-89. (26) Nembrini, G. P.; Capobianeo, J. A.; Viel, M.; Williams, A. F. Geochim. Cosmochim. Acta 1983, 47, 1459-1464. (27) Cai, W.-J.; Sayles, F. L. Mar. Chem. 1996, 52, 123-131. (28) Edwards, R. W.; Rolley, H. L. J. J. Ecol. 1965, 53, 1-19.

Received for review February 25, 2002. Revised manuscript received July 2, 2002. Accepted July 22, 2002. ES020039Z

VOL. 36, NO. 20, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4301