Temporal Coherence Between Lake and Landscape Primary Productivity

Understanding the patterns and drivers of primary productivity is a major goal of ecology, but little is known about whether the primary productivities of different types of ecosystems—here, lakes and the landscapes in which they are embedded—fluctuate in related ways through time. Due to shared climatic variation and well-known connections between lake and terrestrial ecosystems, such as nutrient and resource subsidies, we hypothesized that interannual fluctuations in aquatic and terrestrial primary productivity indices could be coherent. We also expected that lake and watershed characteristics could modify the strength and nature of primary productivity relationships. We applied wavelet coherence analyses to time series of lake chlorophyll-a and satellite-derived NDVI to examine coherence between lakes and land, and used random forest regression and generalized additive models to evaluate why coherence varies among lakes. There can be substantial coherence between lake and terrestrial primary productivity, but the strength and phase (direction and time lag) of this relationship vary widely, and there were marked differences between short (2–4-year periods of oscillation) and long (> 4-year periods of oscillation) timescales. Across all timescales, variables associated with the connectedness of lakes to their watersheds were consistently the important explanatory variables of the strength and phase of coherence. The patterns observed in this study suggest the importance of cross-ecosystem flows, as opposed to shared climatic variation, in determining temporal coherence between lakes and the landscape.