Reduced precipitation can induce ecosystem regime shifts in lakes by increasing internal nutrient recycling

Abstract

Eutrophication is a main threat to continental aquatic ecosystems. Prevention and amelioration actions have been taken under the assumption of a stable climate, which needs reconsideration. Here, we show that reduced precipitation can bring a lake ecosystem to a more productive regime even with a decline in nutrient external load. By analyzing time series of several decades in the largest lake of the Iberian Peninsula, we found autocorrelated changes in the variance of state variables (i.e., chlorophyll and oxygen) indicative of a transient situation towards a new ecosystem regime. Indeed, exceptional planktonic diatom blooms have occurred during the last few years, and the sediment record shows a shift in phytoplankton composition and an increase in nutrient retention. Reduced precipitation almost doubled the water residence time in the lake, enhancing the relevance of internal processes. This study demonstrates that ecological quality targets for aquatic ecosystems must be tailored to the changing climatic conditions for appropriate stewardship.

Keywords: Climate change; Conditional heteroscedasticity; Diatom blooms; Ecosystem regime shift; Global warming; Internal nutrient loading; Lake Sanabria; Long-term monitoring; Nutrient retention; Precipitation decline; Shifting reference states.

Figure 1

Lake Sanabria watershed. (a) Lake Sanabria basin, highlighting the sampling stations used in this study. Place names are indicated in Fig. S1. (b) Lake location within Europe. (c) Lake western, and (d) eastern views. (e) Lake bathymetry map, which can be found enlarged in Fig. S2.

Figure 2

Time series (1986–2018) in Lake Sanabria of water column integrated chlorophyll (a), and chlorophyll (b), and oxygen (c) variation across depth, based on monthly measurements at every 2.5 m depth in the D02 station in the deepest part of the eastern basin.

Figure 3

Changes in the mid-term statistics of chlorophyll at 15 m depth, where chlorophyll maxima usually occur, and oxygen at ca. 50 m depth, near the bottom of the deepest point. A 36-month rolling window was considered to estimate the statistical descriptors (a,c), and the conditional heteroscedasticity tests (b,d): values above the red line indicate significant conditional heteroscedasticity (P < 0.05).

Figure 4

Contrasting land cover situations between 1956 and the present (a) in the Lake Sanabria watershed allows the comparison of phosphorus (b) and nitrogen (c) external load change according to the modeled contribution of primary sources.

Figure 5

Weather trends in the Lake Sanabria area during the last decades. (a) Precipitation at Puente Porto (M02 station), and (b–d) temperature at Puebla de Sanabria (AEMET 2770B station). The latter is located outside the lake watershed but at a short distance. According to the Mann–Kendall trend test, linear tendencies (black line) are significant for precipitation (P = 0.001, n = 666), maximum (P = 0.0006, n = 666), and minimum temperature (P = 0.017, n = 666). The red line indicates a one-year moving average.

Figure 6

Lake Sanabria physical trends during the last decades. (a) Water column heat and (b) temperature patterns. (c) First date and temperature of lake isothermy (Mann–Kendall trend test, n = 31, P = 0.005, P = 0.0008, respectively). (d) Highest monthly buoyancy frequency during the stratification period (Mann–Kendall test, n = 32, P = 0.02). (e) Number of times that water renews annually for some selected depth layers across the water column. The black line indicates the average for 1986–2018 and bars above (blue) and below (orange) deviations.

Figure 7

Sediment indicators of main functional and community changes in Lake Sanabria during the last decades (a) Organic carbon, total nitrogen, and total phosphorus sediment accumulation in the deepest part of the lake. Carbon and nitrogen stable isotopic compositions are also indicated (red line). Note the marked 1959 oscillation corresponding to the Vega de Tera dam collapse that reset the lake's biogeochemical dynamics. See Fig. S7 for the depth-age model. (b) Main chrysophyte cyst changes. (c) Diatom remains (Asterionella ralfsii var americana, Tabellaria flocculosa, Aulacoseira subborealis/pseudodistans, and Cyclotella stelligera), and representative algal pigments (pheophytin-a, bulk phytoplankton indicator; lutein, chlorophytes; echinenone, cyanobacteria; alloxanthin, cryptophytes).