A general management model to optimize lake liming operations: Fid-Bau Portal

A general management model to optimize lake liming operations

Håkanson, Lars

One of the most common ways to combat the ecosystem effects of acid rain is liming, which can be performed in many ways, such as lake liming, wetland liming, full‐scale drainage area liming, and by dosers in tributaries. This study presents results from extensive liming operations in many Swedish lakes. A practically useful management model for planning and predicting the outcome of liming is presented. The model is based on a validated, general dynamic model for radiocaesium, which has also been used with very good results for radiostrontium, different metals and phosphorus. In this work, the model is applied for calcium from liming. Lake liming can be regarded as a tracer experiment to test the dynamic model. A few of the lakes have been used for the calibrations of the model, and then the model has been validated (blind‐tested) using 68 lakes for pH and 17 lakes for calcium concentrations. Results were very good or good in 97% of the cases for pH and 94% of the cases for calcium. Controls in the two lakes where the dynamic model did not give good pH predictions revealed that these two lakes were fertilized and limed a few years before this study started. If those conditions are accounted for, the model would give good results in these cases also. The model also includes submodels to calculate initial pH, and how this is related to catchment area and lake morphometric conditions, as well as target pH and how this is related to initial pH, and the large‐scale, long‐time influences of deposition of sulphur and other acidifying compounds from the burning of fossil fuels. In order to make the model practically useful, only variables readily accessible from maps and standard monitoring programmes have been accepted as driving variables (such as lake area and mean depth). This work also presents critical model tests (sensitivity and uncertainty analyses using Monte Carlo techniques). These tests have demonstrated that the critical part determining the predictive success of the model is related to the final calculation step where the calculated calcium concentration is transformed into a change in lake pH. It has also been demonstrated that more detailed characterizations of the catchment areas by means of drainage area zonation or Geographical Information System techniques will not increase the predictive accuracy of initial pH values, and that most lakes in the northern part of Sweden were excessively limed to pH values higher, or much higher, than the ‘natural’ or target pH. This is a waste of resources as there are many lakes, especially in the southern part of the country, which need to be limed.