Background The regulatory risk assessment of pesticides requires the assessment of exposure of aquatic ecosystems in small surface waters adjacent to agricultural fields. This exposure is predicted using simulation models, for which an important input parameter is the degradation rate in water. In regulatory dossiers, the decline rate in water from outdoor mesocosms is often available, but this rate encompasses more processes than degradation. Therefore, a procedure was designed for estimating the degradation rate in water that was suitable for mesocosm studies with limited datasets, e.g. datasets lacking site-specific sorption coefficients and relevant sediment properties. The procedure, based upon inverse modelling with TOXSWA, has been tested on a dataset for prosulfocarb in stagnant ditches. Results A degradation half-life in the ditch water of 2.9 days (20 °C) was found. This short half-life was to a significant extent accounted for by biodegradation rather than hydrolysis or photolysis. This half-life was considerably shorter than the half-life in the water layer of two laboratory water–sediment system experiments. Conclusion The estimation procedure resulted in a unique half-life for the degradation rate in water. Such outdoor mesocosm studies seem to be better suited to assessing the degradation rate in water in ditches than the conventional water–sediment studies.