Monitoring, understanding and modelling carbon emission and fixation fluxes are key actions to guide climate change stakeholders in the application of mitigation strategies. In this study, we use the remote sensing model C-Fix at the local stand scale to improve the integration of algorithms for water and temperature limitation. These new algorithms are applied to estimate net ecosystem productivity in a fully water limited mode. Four different C-Fix scenarios (partially and fully water limited, partially and fully water limited with temperature conversion), to obtain net ecosystem productivity of European forests, were compared with EUROFLUX measurements using different statistics. In the first scenario (i) C-Fix is partially water limited since the NDVI is used as indicator for the fraction of absorbed photosynthetic active radiation (fAPAR). The NDVI depends on plant water availability at the long term (months). In the fully water limited C-Fix scenario (ii) short-term water availability is included taking into account soil moisture content. The third (iii) and fourth (iv) scenarios are respectively the partially and fully water limited scenarios wherein canopy and soil temperature, derived from the atmospheric temperature, are incorporated to account for the temperature driven processes of vegetation and soil. The error propagation by C-Fix was also analysed. Introducing water limitation and converting atmospheric temperature to apparent canopy and soil temperatures, significantly improves validation results based on EUROFLUX data. For pooled data, a final slope, intercept and correlation coefficient of the linear regression between C-Fix and EUROFLUX NEP values is 0.86, 0.09 and 0.64, respectively. When implementing full water limitation a reduction of 42% of the estimated NEP is obtained with respect to partially limited NEP. The error propagation by C-Fix for the partially and fully water limited cases and when applying temperature buffering, are similar. Respectively, an average daily NEP of 0.69 +/- 0.4 and 0.40 +/- 0.3 g C m(-2) d(-1). Due to the combination of simple semi-empirical approaches, inclusion of full water limitation into C-Fix enhances the application field of a remote sensing based model with potential in regional, continental and global dry matter estimations and in carbon uptake/release studies, although a more physically based support is recommended to convert atmospheric temperature to soil and canopy temperature and to estimate fAPAR. (c) 2006 Elsevier B.V. All rights reserved.