Efficient integration of increasing shares of renewable resources into electricity supply requires information about spatial and temporal variations in the availability of energy carriers. Regarding biomass, the spatial variation is crucial for the estimation of the regional potential and costs of biomass for electricity generation. Described in this paper is the examination of the applicability of net primary productivity data for improving the quality of a high resolution inventory of potentials of biomass energy. The net primary productivity is simulated with remotely sensed input data taking into account site-specific information such as precipitation, irradiation and temperature. For the examples of forest wood and surplus straw, two different methods of inventorying were applied and compared: 1) disaggregating biomass potentials derived from statistical crop data using a land cover map and 2 additionally weighting the distribution with the net primary productivity, We found that in the case of surplus straw, this method yielded better results than only disaggregating the potential with the land cover data. No clear conclusions could be drawn from the statistical test applied regarding forest wood, because the sample size of reference values was too small. Based on a critical reflection of the data which is input to the NPP simulation, it was concluded that the distribution of forest wood potentials should not be weighted with the NPP data unless more appropriate satellite input data is available. (C) 2008 Elsevier Ltd. All rights reserved.