Forecasting changes to ecological communities is one of the central challenges in ecology. However, nonlinear dependencies, biotic interactions and data limitations have limited our ability to assess how predictable communities are. We used a machine learning approach and environmental monitoring data (biological, physical and chemical) to assess the predictability of phytoplankton cell density in one lake across an unprecedented range of time scales. Communities were highly predictable over hours to months: model R2 decreased from 0.89 at 4 hours to 0.75 at 1 month, and in a long-term dataset lacking fine spatial resolution, from 0.46 at 1 month to 0.32 at 10 years. When cyanobacterial and eukaryotic algal cell density were examined separately, model-inferred environmental growth dependencies matched laboratory studies, and suggested novel trade-offs governing their competition. High-frequency monitoring and machine learning can help elucidate the mechanisms underlying ecological dynamics and set prediction targets for process-based models.
Phytoplankton cell densities by scanning flow cytometry and environmental variation across depths from Lake Greifensee, Switzerland, in 2014 and 2015
This is a high-frequency dataset consisting of estimates of phytoplankton cell density every 4 hours during the summer and autumn of 2014 and 2015, measured using scanning flow cytometry. It also includes associated environmental variables, some of which were measured at longer time scales, such as once to twice a week (as in the case of zooplankton and dissolved nutrients).
Dataset_ELE_Greifensee_2014plus2015_TS_cytobuoy_env_logabund.csv
Phytoplankton cell densities by microscopy and environmental variation from Lake Greifensee, Switzerland, from 1984 to 2016
This is a dataset consisting of estimates of phytoplankton cell density every month from March 1984 to June 2016, measured using microscopy. It also includes associated environmental variables, most of which were measured at the same location and approximately the same time (as in the case of zooplankton densities and dissolved nutrients). Irradiance measurements have been estimated separately, based on published estimates from a nearby site (see paper for details). For the purpose of analysis, dates have been rounded to the beginning of the nearest month, and where duplicates emerge from this rounding, values have been averaged between the two.
Dataset_ELE_Greifensee_long_term_TS_microscopy_env.csv