Environ. Sci. Technol. 2002, 36, 2039-2047
Two-Dimensional Small-Scale Variability of Pore Water Phosphate in Freshwater Lakes: Results from a Novel Dialysis Sampler JO ¨ RG LEWANDOWSKI, KRISTINA RU ¨ TER, AND MICHAEL HUPFER* Leibniz-Institute of Freshwater Ecology and Inland Fisheries (IGB), Mu ¨ ggelseedamm 301, D-12587 Berlin, Germany
Vertical concentration profiles of soluble reactive phosphorus (SRP) in the upper sedimentary zone of freshwater lakes are an important means for studying internal phosphorus (P) loading and to gain insight into early diagenetic processes. The interpretation of such pore water profiles generally neglects the occurrence of horizontal variability at a specific sampling site. To further examine this variability, we have designed a novel twodimensional sampler (2D peeper) consisting of 2280 chambers at a spatial resolution of 9 mm providing a sampling area of 43 × 44 cm. This new device was deployed in three eutrophic lakes in north-eastern Germany. The resulting 2D images of the SRP concentrations, diffusive fluxes, and turnover rates revealed systematic vertical and horizontal structures with local niches of increased phosphorus release. Thus, the extrapolation of P flux calculations based on one-dimensional pore water profiles may lead to a considerable error. The observed small-scale horizontal heterogeneity, probably mainly caused by organisms, was larger in the biologically more active Lake Mu¨ llrose and Su¨ sser See than in the deeper Arendsee where meio- and macrozoobenthos were missing. In all cases, the variability was highest at the sediment-water interface and diminished with sediment depth.
Introduction Concentrations of solutes in pore waters react very sensitively to changes in environmental conditions. Therefore, these concentrations are essential indicators for biogeochemical processes in lake sediments (e.g., refs 1 and 2) and might be used to study sorption and mineralization. Although the amount of phosphorus (P) dissolved in pore waters is small compared to the amount stored in sediment (3), frequently, the greater part of P release to overlying water takes place via pore water. The P release can be measured over a time interval as the difference between two concentrations in sediment core incubations, with flux chambers exposed insitu, or in stratified lakes as hypolimnetic P accumulation. Additionally, P release might be estimated from P concentration gradients at the sediment-water interface using Fick’s first law of diffusion (4). A broad variety of techniques are used to measure pore water gradients. For slicing techniques, undisturbed sediment cores are taken and sectioned, and * Corresponding author phone: +49 30 64181-605; fax: +49 30 64181-682; e-mail:
[email protected]. 10.1021/es0102538 CCC: $22.00 Published on Web 04/05/2002
2002 American Chemical Society
pore waters are separated by centrifugation, vacuum filtration, or pressure filtration (3). In in-situ dialysis, pore water samplers (5), so-called peepers, and gel samplers (6-9) utilize the equilibration process between pore water and sampling medium caused by dialysis. For a few ions, high-resolution in-situ techniques employing microelectrodes are available (10, 11). Large-scale horizontal variations in pore water concentrations may be caused by unevenly distributed inflows (12), as well as by local differences in water depth, the infiltration ratio, and wind exposure (3). According to Håkanson and Jansson (13), lake sediments are more heterogeneous the higher the shore development. Vertical variations of pore water concentrations reflect variations of the chemical milieu and the composition of the solid sediment, which results from the chemical characteristics of the former settled particles and former diagenetic processes (14). Although the existence of microenvironments has previously been proposed (15), lake sediments are traditionally regarded as onedimensional (1D) systems (4), with redox processes being sequentially layered but laterally uniform. This simplified 1D view, at first applied mainly to keep the problem manageable, was artificially reinforced by use of 1D sediment models and 1D pore water concentration gradients, both interpreted as if being valid in a 3D view. The latest developments of techniques for measurements with higher spatial resolution have deeply changed our perception of pore water chemistry. Measurements of dissolved oxygen by microsensors in two (16) or three (17) dimensions and of dissolved metals by gel samplers in two dimensions (2, 18) have revealed considerable small-scale horizontal and vertical variability in pore water concentrations, with most of the turnover occurring at discrete highly reactive sites (19-21). During a preliminary experiment, we deployed four or six 1D peepers at two locations in the sediments of Lake Mu¨llrose, each set within an area of 2 × 2 m, to compare P release from the two different locations. However, the statistical scatter at each location was far too large to compare the two sampling points with each other (coefficient of variation at 28% or 41%, respectively), and the principle shape of the profiles differed even more (see the Supporting Information). To further examine the observed small-scale variability in Lake Mu ¨ llrose as well as in other lakes, we designed a novel 2D peeper and analyzed the 2D distribution of SRP concentrations in pore waters. Although the diffusive gradients in thin films (DGT) technique for phosphate is under development (18) and phosphorus plays a key role in eutrophication, analysis of the 2D distribution of SRP in pore waters has not previously been achieved.
Materials and Methods Study Sites. Three lakes located in north-eastern Germany were the subject of our research. The Arendsee is about 50 m deep and stratified during summer, whereas Lake Mu ¨ llrose and Su ¨ sser See are polymictic lakes less than 10 m deep. All three lakes show typical eutrophication phenomena, such as high primary productivity and mass developments of cyanobacteria which caused the responsible authorities to carry out restoration measures. Morphometric, trophic, and sediment characteristics of the three lakes are listed in Table 1. Bathymetric maps and position of sampling points are illustrated in Figure 1. In the Arendsee, to shorten the adaptation time to a decrease of external P loading, sediments from calcite-rich deposits (Seekreide) near the shore were distributed to cap the natural bottom sediments with an average of 2.6 cm of VOL. 36, NO. 9, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2039
TABLE 1. Morphometric, Trophic, and Sediment Characteristics of the Arendsee, Lake Mu1 llrose, and Su1 sser See morphometry area (km2) mean depth (m) maximum depth (m) shore development water residence timeyear mixis trophic state sedimentb TP (mg g-1 dw) Ca (mg g-1 dw) Fe (mg g-1 dw) Al (mg g-1 dw) loss on ignition (%) a
Arendsee
Mu1 llrose
Su1 sser See
5.14 28.6 48.7 1.1 114 mono- to dimictic eutrophic
1.27 4.9 7.5a 2.0 0.4 polymictic hypereutrophic
2.68 4.3 8.2 2.0 1.2 polymictic hypereutrophic
1.32 239 7.1 8.0 26.5
1.59 122 23.6 5.9 34.0
2.49 134 16.1 19.1 15.5
Water depth of the sediment trap at the mouth of River Schlaube 11.2 m. b Sediment (0-5 cm) at sampling point.
FIGURE 1. Bathymetric maps with water depth in meters and positions of sampling points of Arendsee, Lake Mu1 llrose, and Su1 sser See. calcareous mud. However, Seekreide showed no substantial P uptake, SRP transport through the Seekreide layer was controlled mainly by diffusion, and consequently, capping of lake sediments did not significantly prevent P release (22). Lake Mu ¨ llrose was dredged between 1992 and 1998 using a swimming suction dredge; at the sampling point, dredging was carried out in 1994. In the catchment of Lake Mu ¨ llrose, no external measures have been performed to date. It is controversial whether the trophic state of Lake Mu ¨ llrose improved because of dredging or not. In Su ¨ sser See, phosphorus precipitations with large amounts of aluminum sulfate were carried out almost annually from 1977 to 1992, producing a sediment layer with a substantial portion of aluminum-bound phosphorus. Some external measures were conducted in the past decade to decrease P loading. However, the water quality has not improved substantially. 2040
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 9, 2002
Sampling and Chemical Analysis. We constructed a 2D peeper according to the principle of Hesslein’s (5) 1D peeper, with 48 rows and 49 columns to investigate the horizontal and vertical heterogeneity of the sediment (Figure 2). The volume of the drill holes in the Perspex plate was 650 µL, the diameters of the holes were 7 mm, the depth of the holes were 17 mm, and the distances between the dialysis chambers were 2 mm. Thus, the peeper provides a horizontal and vertical resolution of 9 mm. The sampler was initially filled with oxygen-free distilled water and coated with a polysulfone membrane with a pore size of 0.2 µm (HT-Tuffryn 200; Pall Gelman Laboratory). The peeper was fixed in a frame and exposed vertically in the sediment with about half of the peeper chambers in the sediment and the rest in the overlying water. All investigations were conducted during turnover between October 2000 and May 2001. Fourteen days after
FIGURE 2. Two-dimensional pore water sampler (“2D peeper”). exposure, the peeper was pulled out, stored at about 10 °C under a nitrogen atmosphere to inhibit oxidation, and brought to the laboratory, where the chemical analysis of the samples was carried out within 24 h. The position of the sediment-water interface can be recognized by the peeper position in the frame, by discoloration of the membrane, as well as by epiphytes and sediment on the peeper. SRP was determined photometrically following a modified version of the molybdenum blue method (23) which was scaled down to a final volume (sample + reagent) of 300 µL. This modification allowed for the use of microtiter plates and as such for a parallel processing of 96 samples. Repetitions showed a relative standard deviation
