Spring coherence in dissolved organic carbon export dominates total coherence in Boreal Shield forested catchments

Abstract

Abstract The wide range of forested landscapes in boreal environments store and cycle substantial amounts of carbon, although the capacity of these systems to act as either a carbon sink or source is uncertain under a changing climate. While there are clear reports of regional-scale increases in dissolved organic carbon (DOC) concentrations in streams and lakes, there remains substantial watershed-scale variability in these patterns. Coherence is a framework for examining if variables of interest within adjacent spatial units change synchronously or asynchronously through time and has been widely applied in the context of lentic hydrochemistry, and which can shed light on the relative importance of regional- vs. local-scale controls. The objective of this research was to quantify coherence in discharge, DOC concentrations, and DOC loads in forested boreal watersheds, and to what extent coherence varied by season. Coherence was assessed using data from three long-term ecological research sites spanning boreal forest environments (IISD-Experimental Lakes Area, Turkey Lakes Watershed Study, and Dorset Environmental Science Centre) that included 29 829 DOC measurements across 739 stream-years, examining correlation between stream-pairs within each site, but not between sites. Seasonal coherence in DOC export was consistent across the three sites; coherence was significantly greater in spring than all other seasons, and was strongly related to discharge coherence. Currently, the season with the greatest loads (spring) is also the most coherent season, suggesting that annual between-stream coherence may be reduced if spring becomes proportionally less important in hydrologic budgets under a changing climate. This research aids in determining which factors contribute to synchronous watershed behaviour, and which factors may contribute to the timing and extent of individual watershed-scale deviations from landscape-level patterns.