Addressing the bottom-up and top-down control in lakes through a dynamic optimization

V. ESTRADA; E. PARODI; M.S. DIAZ

San Luis

Congreso; XVII Congreso sobre Métodos Numéricos y sus Aplicaciones; 2008

INTEC, INTEMA, UNS

The present study has been performed on Paso de las Piedras Reservoir (38° 22´ S and 61° 12´ W) that supplies drinking water for more than 400,000 inhabitants and for industrial purposes. The trophic level of this unstratified water body currently corresponds to eutrophic category due to intense agricultural activities in this region of Argentina and it undergoes recurrent algal blooms dominated by cyanobacteria species (mainly Microcystis aeruginosa and Anabaena circinalis) during warm months (Parodi et al., 2004). El Divisorio Stream and Sauce Grande River are the two tributaries of the lake. An efficient pilot wetland has been built prior to the lake to retain phosphorus and nitrogen from El Divisorio Stream and, in this way, decrease the external loading of nutrients to the lake (Lopez et al., 2007). In this work, we develop a dynamic optimization model (Rodriguez and Diaz, 2007) for Paso de las Piedras Reservoir to determine optimal limiting nutrient inflow and herbivore zooplankton biomass concentration profiles to the lake for algal growth control. Biogeochemical processes that take place in water bodies can be represented through a set of complex nonlinear partial differential algebraic equations resulting from dynamic mass balances that in this study account for phytoplankton, zooplankton, nutrients, dissolved oxygen and biochemical demand of oxygen along the water column. Water samples and in situ determinations have been weekly taken from January to December 2004 at four sampling stations for biological qualitative and quantitative determinations, as well as physicochemical ones (nutrients, temperature, DO, pH, Secchi depth) to do calibration to suit site-specific conditions. Numerical results have shown that the botton-up control is not enough, but necessary, to restore water quality due to internal recycling of nutrients and that in-lake reduction approaches, like increase in herbivore zooplankton concentration, would have a greater impact on improving water quality. The optimization model allows us know how much increase in the concentration of herbivore zooplankton biomass need to control phytoplankton growth and this can be later associated to the quantity of zooplanktivorus fish we must withdraw. The detailed biogeochemical model that has been formulated within a simultaneous optimization framework allows the determination of optimal restoration policies to improve water quality in a highly eutrophic reservoir.