The effect of water level fluctuation on benthic macroinvertebrate communities of stony lake littorals: A study of natural lakes and hydroelectric reservoirs of boreal environments

The disturbance of water level fluctuation (WLF) in lacustrine boreal environments is poorly understood. This is surprising given its vast area, circumpolar nature, immense amount of surface water, extensive hydroelectric development and anticipated future climate induced alterations to hydrologic inputs. Further, both the character of WLF, and the effect of WLF on biota inhabiting highly productive littoral areas, which are closely linked at the terrestrial aquatic interface and are therefore highly sensitive to WLF, is not well known. This thesis addresses this lack of knowledge through multiple studies which address the general hypothesis that WLF regulates near-shore littoral benthic macroinvertebrate (BMI) community structure.;The novel results of these studies demonstrate patterns in natural WLF and associated correlations with water quality and littoral BMI communities across multiple lakes within the Laurentian Great Lakes region. The results are congruent with the intermediate disturbance hypothesis and the insurance hypothesis and have direct implications for reservoir management and forecasting climate change impacts.;Using a long-term (1980-2000) data set of natural WLF, water quality and aquatic biota from four remote research areas in the Laurentian Great Lakes region both linear and waveform regression analyses revealed a significant (p ≤ 0.05) decreasing trend in water levels and a ten year oscillation in WLF. Similarly, linear regression analysis demonstrated a significant reduction in yearly amplitude WLF over time. Similarly, BMI taxa richness followed a unimodal response (p = 0.002, r2 = 0.66) with richness decreasing in years when water levels were either higher or lower than the long-term mean. A similar result was found when studying stony littoral BMI communities collected in 2006 of sixteen large boreal shield lakes. Using PCA we derived a new variable D80-D210 (March 31st minus August 1st) as a surrogate for change in mean water level and potential habitat squeeze (loss). We found that, of 73 environmental variables, only three resulted in significant relationship with BMI richness: lower water levels, quantified as D80-D210, have higher macroinvertebrate richness (r2 = 0.38linear ), a classic species-area relationship as BMI richness increases with increasing lake area (r2 = 0.38linear, r 2 = 0.69unimodal), and that as littoral slope increases BMI richness decreases (r2 = 0.32linear). Further, a reference condition approach (RCA) comparing 20 natural lakes and 28 hydroelectric reservoirs demonstrated that BMI taxa richness decreased with increasing WLF magnitude (r2 = 0.47--0.60) especially when WLF deviated from normal (reference) conditions. Water level amplitude was negatively correlated (P < 0.05) with taxa richness in reservoirs but not in natural lakes. We found that thirteen reservoirs fell outside the 95% confidence ellipse for reference sites and had significantly different taxonomic composition than that of the natural lakes. Functional mobility groups (FMG) and functional feeding groups (FFG) were also altered with increasing WLF (non parametric ANOVA, P < 0.05). Swimmers, collectors and detritivores were monotonically positively related to increasing WLF, whereas clams (Sphaeriidae) and filterers were negatively associated with increasing WLF. Crawlers were the only functional group that demonstrated a significant unimodal distribution as predicted by the intermediate disturbance hypothesis. Finally, a depth manipulation experiment demonstrated significantly higher BMI richness, detritivore and burrower abundances in control groups compared to treatment groups and a change in community structure due to depth level.