Agricultural practices changed greatly in the watershed of Acton Lake, a Midwestern US reservoir, during the 1990s. We studied the changes in the streams that drain the watershed and flow into the lake, as well as lake ecosystem responses, over a 22-year period. We hypothesized that decreasing phosphorus (P) inputs from the streams, resulting from better agricultural practices, would lead to decreased phytoplankton biomass in this highly productive lake. In accordance with this special session, we present some of the ecological surprises that occurred during this long-term study.
The most important change in watershed agriculture was a pronounced increase in conservation tillage (15 to 65% of cropland) from 1990-2000, although declines in estimated fertilizer use and livestock (hogs) also occurred. Over the first decade of our study, concentrations of ammonium (NH4), soluble reactive phosphorus (SRP), and suspended sediment (SS), adjusted for stream discharge, declined greatly (5-8% per year) in the lake’s inflow streams; nitrate (NO3) concentration did not change. These changes were not surprising, given the changes in agriculture. However, subsequent changes in streams, and the lake ecosystem responses, were surprising. The first surprise is that stream nutrient dynamics changed direction and magnitude greatly over the second decade, when conservation tillage stabilized at ~65%. During this period, SS and NH4 concentrations continued to decline but at slower rates, while SRP concentrations actually increased; in contrast, NO3 concentrations declined rapidly. The response of the lake ecosystem was also surprising and illustrates resilience against decreasing nutrient inputs. Despite decreased P concentrations in inflows, phytoplankton (which are usually P-limited) increased by ~60% over the first decade. We attribute this to decreasing suspended sediment (SS) concentrations in the lake, which alleviated phytoplankton light limitation. However, increased phytoplankton can occur only with an alternative, increasing P source, which in this case derived from excretion by detritivorous fish (gizzard shad, Dorosoma cepedianum). Gizzard shad feed on sediment detritus and excrete nutrients into water, rendering them available to phytoplankton. During the first decade, gizzard shad biomass (and excretion rates) increased by about 2X, corresponding with increasing phytoplankton. Over the second decade, phytoplankton and fish biomass stabilized. Our results reveal the complex and surprising responses of a lake ecosystem to long-term changes in ecological subsidies of nutrient and detritus, and illustrate how food webs can provide resilience against change. From a management perspective, planning for such resilience may reduce surprises and help develop strategies to improve water quality.