Calibration of the zooplankton community size spectrum as an indicator of change in Canadian Shield lakes

Abstract

Developing the crustacean zooplankton community size spectrum into an indicator of change in lakes requires quantification of the natural variability in the size spectrum related to broad-scale seasonal, annual, and spatial factors. Characterizing seasonal patterns of variation in the size spectrum is necessary so that monitoring programs can be designed to minimize the masking effects that seasonal processes can have on detecting longer-term temporal change. We used a random effects model to measure monthly, annual, and interlake variability in the slope (i.e., relative abundance of small and large organisms) and centered height (i.e., total abundance) of the crustacean zooplankton normalized abundance size spectrum from 1981 to 2011 among eight Canadian Shield lakes. Consistent with theoretical predictions, the slope was a relatively stable characteristic of the zooplankton community compared with the height, which varied significantly among lakes. We identified a seasonal signal in height and slope and used a mixed effects model to characterize the linear rate of change from May to October; there was an overall decline in height and an overall increase in slope. Seasonal variance was greater than annual variance for both the height and the slope, suggesting that long-term monitoring of lakes and interlake comparisons using zooplankton size spectra should be based on temporally standardized sampling protocols that minimize the effects of seasonal processes. We recommend sampling the zooplankton community in midsummer because this results in size spectrum estimates close to seasonal mean values.

Size spectra analysis of a decade of Laurentian Great Lakes data

“…However, the use of height alone can make comparisons across systems difficult because the range of size bins can vary across lakes and even sites (as it does in our dataset), and thus the midpoints are not always comparable. Therefore, we fixed height to a specific size group (size bin: 2 À23 g; wet weight: $1.2 Â 10 À7 g, primarily small zooplankton like rotifers) to make multi-lake comparisons, as in Barth et al (2019) (Fig. 1).…”

“…Size spectra slopes are postulated to reflect energy transfer efficiency across trophic levels, and changing slopes may indicate ecosystem perturbations across entire food webs (Sprules 2008;dos Santos et al 2017;Barth et al 2019). However, slopes may be less responsive than other metrics (e.g., height) to overall ecosystem productivity.…”

“…For our data, MLE may give higher influence to small size groups because the smallest bins contain such an overwhelming number of individuals (e.g., a billion cells). Second, OLS has been the primary method used both historically (Sprules et al 1991;Vidondo et al 1997;Hocking et al 2013) and in more recent studies (Yurista et al 2014;McGarvey and Kirk 2018;Barth et al 2019); thus, using OLS allows for direct comparisons of our data with this literature. Third, we found limited effect of slope between lakes, and only large differences Fig.…”

“…A shallower slope than theoretical suggests more larger sized animals than expected and the converse if slopes steepen. For instance, deviations from this anticipated slope have been reported in systems with large influxes of migratory species (Duplisea and Kerr 1995), substantial incorporation of allochthonous material into food webs (Perkins et al 2018), overfishing (Robinson et al 2017), or recent invasions of non-native species (Kopf et al 2019;Barth et al 2019). The intercept and (or) height (predicted number of organisms in the midpoint) of size spectra have been postulated to reflect system production (e.g., higher intercepts or heights reflect greater abundance at a given body size; Jennings and Blanchard 2004;Guiet et al 2016).…”