Effect of chlorophyll sampling design on water quality assessment in thermally stratified lakes

In order to adequately assess the ecological status of thermally stratified lakes based on chlorophyll, the sampling must cover all productive layers of the water column. Missing the deep chlorophyll maxima (DCM) that often occur in the meta- or hypolimnion of transparent lakes supported by sufficient illumination and good nutrient availability may cause serious underestimation of the productivity and lead to misclassification of the lake ecological status. There is no commonly accepted sampling design for stratified lakes, and various monitoring guides suggest controversial designs. Our aim was to find some robust criteria to assess the probability of occurrence of a DCM and estimate the differences in measured mean chlorophyll concentrations caused by various sampling designs. Our theoretical model showed that the probability of occurrence of a DCM increases with increasing water transparency and decreasing lake size. Empirical data from Italian and Estonian stratified lakes confirmed the results. Testing of different sampling designs on lakes with full measured chlorophyll profiles available showed that taking only surface layer samples will lead with a high probability to an underestimation of the chlorophyll concentration in the trophogenic layer. In order not to miss the Chl peak in stratified lakes, in most cases it would be more precautious not to limit the sampling to the well-mixed epilimnion but to extend it to the whole euphotic layer. Sampling the epilimnion instead of the euphotic zone could cause up to a 70% underestimation of the chlorophyll concentration, an error that would cause a misclassification of the lake by one or even two status classes in a 5-class assessment system. In most cases, the 2.5 * Secchi depths proved a suitable criterion of the sampling depth and only in the case of surface scums, would sampling of a 3 * Secchi depth layer be recommended in order not to miss the deep chlorophyll maximum.