Variation of Lake Water Chemistry and Spatial Scale

Abstract

Lake chemistry generally reflects the ecological, geological, and climatic processes in and around the lake. This thesis addresses the importance of scale, from small (variability within a lake) to large (catchment and ecoregion variability), on selected chemical metrics. In Paper I, the importance of ecoregions and catchment characteristics for prediction of pH and total phosphorus (TP) were studied. Predictive models were evaluated both on ecoregional (six ecoregions) and whole-country scales. The models were based on catchment-specific variables alone or using a combination of catchment-specific and chemical variables. We found that grouping lakes by ecoregion worked well when within-ecoregion variance was high. For example, partitioning spatial variance by ecoregions worked better for the ecoregions in the south of Sweden than in the north. However, TP and pH were best predicted by other chernical variables, and many of the "better" predictor variables were known to depend on characteristics of the catchment. For example calcium concentration was a good predictor for pH. Future studies should focus on deterrnining the linkage between catchment-scale variability (e.g. geology and weathering) and in-lake chemistry. Paper Il addressed the question of how representative a mid-lake sample is of in-lake conditions. This information is important for interpreting the long term changes within lake monitoring programs. The surface water chemistry of 34 Swedish lakes, spatially-stratified across three main ecoregions and according to residence time and mean depth, was studied. This study also included two seasons; namely a period when biological activity is expected to be high (late summer or August) and a period when hydrological (i.e. wind-induced) disturbance is expected to be high (autumn or October). Where a sample was taken within a lake was found to affect the sample's representativity of lake water surface chernistry. However, a mid-lake sample was shown to be more representative than a random sample. We conclude that mid-lake samples can be considered representative of average surface water chernistry. This information is crucial for how we interpret data taken from monitoring programs that are temporally extensive.

Published in

Rapport / Sveriges lantbruksuniversitet, Miljöanalys
2001, number: 2001:22
Publisher: Institutionen för miljöanalys, Sveriges lantbruksuniversitet

SLU Authors

• Göransson, Elisabet

  • Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences
    

UKÄ Subject classification

Oceanography, Hydrology, Water Resources

Permanent link to this page (URI)

https://res.slu.se/id/publ/132754