Similarity in spatial structure constrains ecosystem relationships: Building a macroscale understanding of lakes

Abstract

Aim: We aim to measure the dominant spatial patterns in ecosystem properties (such as nutrients and measures of primary production) and the multi-scaled geographic driver variables of these properties, and to quantify how the spatial structure of pattern in all of these variables influence the strength of relationships among them. Location and time period: We studied > 8500 lakes in a 1.8 million km2 area of Northeast USA. Data comprised 10-year medians (2002-2011) for measured ecosystem properties as well as long-term climate averages and recent land use/land cover variables. Major taxa studied: We focus on ecosystem properties at the base of aquatic food webs, including concentrations of nutrients and algal pigments that are proxies of primary productivity. Methods: We quantified spatial structure in ecosystem properties and their geographic driver variables using distance-based Moran Eigenvector Maps (dbMEMs). We then compared similarity in spatial structure for all pairs of variables to the correlation between variables to illustrate how spatial structure constrains relationships among ecosystem properties. Results: The strength of spatial structure decreased in order for: climate, land cover/use, lake ecosystem properties, and lake and landscape morphometry. Having a comparable spatial structure is a necessary condition to observe a strong relationship between a pair of variables, but not a sufficient one; variables with very different spatial structure are never strongly correlated. Lake ecosystem properties tended to have an intermediary spatial structure compared to that of their main drivers, likely because climate and landscape variables with known ecological links induce spatial patterns. Main conclusions: Our empirical results describe inherent spatial constraints that dictate the expected relationships between ecosystem properties and their geographic drivers at macroscales. Our results further suggest that understanding the spatial scales at which ecological processes operate is necessary to predict the effects of multi-scaled environmental changes on ecosystem properties.

Publication
Global Ecology and Biogeography
Date