A Primary Production Algorithm Developed for Lake Constance

Kurzfassung

A primary production algorithm based on remote sensing data as inputs is under development for inland waters. For this task, Lake Constance represents a good test site since there exists an extended data base of various in situ measurements for the last two decades. A widely used open ocean primary production model will be adjusted to the conditions in inland waters by combining experimental data and sensitivity analyses. Main factors influencing the primary production are the chlorophyll concentration at the surface, light, temperature, depth of the mixed layer, vertical distribution of water constituents, primary production parameters, and phytoplankton taxonomic composition. The first three factors can be derived from existing satellite data like MOS, NOAA/AVHRR, and METEOSAT. For chlorophyll and temperature as well as wind speed derived from Radar data the synergetic use of the sensors on board of ENVISAT is planned. The data analysis showed that at noon 44 % of the variance in primary production can already be explained by the surface chlorophyll concentration. From surface data, chlorophyll can be well extrapolated down to 10 m water depth which corresponds fairly well with the main part of the euphotic zone during the vegetation period in Lake Constance. The depth-resolved model for the euphotic zone can be improved by using the optical depth instead of the geometrical depth. Temperature as influencing factor is relevant mainly within the upper four meters where primary production is sensitive to temperature. It is studied whether it can be derived together with the depth of the mixed layer from the surface temperature and the two week history of weather data (air temperature, wind speed, and global irradiance) using a neural network approach.