Eu Gene Chung, Fabián A. Bombardelli, and S. Geoffrey Schladow
Sediment–water interactions in shallow lakes and reservoirs may become enormously importantwith time, since bed sediments constitute the ultimate repository for nutrients and contaminants, and fine particles and nutrients in the bed may repeatedly be recycled (Luettich et al., 1990). Nutrient-rich sediment layers in eutrophic lakes and reservoirs participate in chemical and biological processes and exchangemasswith thewater column. Therefore, these sediment layers influence the nutrient cycles via the diffusive fluxes of nutrients fromsediments (the process bywhich dissolved 32 nutrients are diffused from the sediments into the water column after being mineralized in the sediment) (DiToro, 2001), as well as via sediment resuspension (the process by which nutrients and particles are released fromthe bottomto thewater column) (Horne and Goldman, 1994). Furthermore, resuspension of toxic bottom sediments is an ecological concern, since toxins and other hydrophobic organic contaminants can accumulate in the biota and food-web of a lake (Gbah et al., 2001).
The relevance of sediment resuspension towater quality in shallow, eutrophic, wind-exposed lakes has led to the development of water quality models describing sediment resuspension as a main source of nutrients (Søndergaard et al., 1992; Dortch et al., 1996; Mian and Yanful, 2004; Chung et al., 2008). Among these sediment-resuspension models, some are formulated through a simple relationship between just wind speed and the sediment entrainment rates (Aalderink et al., 1985; Somlyòdy, 1986), while some others include a measure of the bed shear stress and an indirect reference to the sediment size (Hakanson and Jansson, 1983; Sanford andMaa, 2001;Mian and Yanful, 2004).More sophisticated expressions for sediment resuspension include the characteristics of the sediment in the formulation, expressed through the settling velocity (García and Parker, 1991, 1993). Based on this wide set of formulas, there is a clear question as to which formula provides the most accurate description of the resuspension rates. Moreover, it becomes very important to choose an appropriate sediment resuspension model which can competently predict the dynamic linkages between sediment resuspension andwater quality in shallow lakes and reservoirs.
The Salton Sea, California, USA, is a shallow, highly eutrophic, wind-dominated, and terminal lake in which the nutrient cycles are strongly related to sediment-resuspension events (Chung et al., 2008, in press). Furthermore, the Salton Sea provides a rare example where the physical size will be significantly reduced in response to major inflow diversions that are planned under the Colorado River Quantification Settlement Agreement (QSA). The change of the morphology of the Sea will alter the fetch of the Sea which in turn will likely affect the sediment-resuspension process in the lake (fetch in a particular direction is typically not variable formost water bodies). The impact of sediment-resuspension events on the nutrient cycles of the Sea has recently beenmodeled using a hydro dynamic and water quality model, the dynamic lake model-water quality (DLM-WQ) (Hamilton and Schladow, 1997; Schladow and Hamilton, 1997; Losada, 2001; Chung et al., 2008). This model had been combined with the empirical sediment-resuspension model of Somlyòdy (1986)which makes no provision for variable sediment characteristics (Chung et al., 2008). Even though some water quality model outputs of DLM-WQ combinedwith the Somlyòdy (1986) model captured some of the seasonal trends and short-term variations, variables related to sediment concentrations (particulate phosphorus and nitrogen in particular) did not show short-term variation and needed further refinement. This suggested the need to link DLM-WQ to more sophisticated sediment-resuspension models,which are based on sediment characteristics and the wind induced bed shear stresses and, thus, are able to describe linkages between variable sediment characteristics and water quality in the Salton Sea more accurately.
In this paper, we examine three sediment-resuspension models variously formulated based on sediment characteristics as well as the bed shear stress exerted by wind-induced waves and currents. Both the first and second models have been designed for cohesive sediments using either non-linear (Mian and Yanful, 2004) or linear relationships (Sanford and Maa, 2001) between sediment entrainment and shear stress, respectively. The third model has been originally formulated for non-cohesive sediments and for open channel flows (García and Parker, 1991, 1993), and was recently extended by Chung et al. (2008) for application to lakes. These three sediment models combined with DLM-WQ are applied to the Salton Sea. The output from these resulting models and from an earlier model (i.e., DLM-WQ with the simple Somlyody model) are compared with measured water quality data from the Salton Sea.
There has been considerable debate on the subject of evaluation and validation of numerical models of natural systems, with disagreement on whether model validation is essential or impossible (Mayer and Butler, 1993; Oreskes et al., 1994; Rykiel, 1996; Arhonditsis and Brett, 2005). Here both statistical and graphical evaluation/validation techniques are used for model comparison, including measures of residual errors, sample autocorrelations, t-tests, and box plots (Reckhow, 1980; Abraham and Ledolter, 1983; Jørgensen et al., 1986; Reckhow et al., 1990; Loague and Green, 1991; Power, 1993). Based on these comparisons,we determine the particular sediment-resuspensionmodel combined with DLM-WQ that most closely matches the measured data. We also investigate the effect of the physical changes onwater quality characteristics in the Salton Sea under two possible future configurations using the selected model.
The scientific and technological goals of our work are as follows: (a) assess whether formulas possessing more information on the sediment characteristics aremore representative than standard formulas; (b) determine the most reliable sediment-resuspension model through statistical comparison of model predictions with water quality data in a shallow, eutrophic, wind-exposed lake where sediments play an important role in the dynamics of dissolved and particle nutrients; (c) develop a tool for prediction of future scenarios in shallow lakes. In fact, we show that the model can be used to predict the magnitude of the change of water quality variables due to the physical alteration of the water body. |
