The main goal of this thesis is the establishment of a framework to analyze the Spanish forest ecosystems in terms of their role in the carbon cycle. In particular, the carbon fluxes that they exchange with atmosphere are modeled to evaluate their potential as carbon sinks and biomass reservoirs. Gross fluxes are estimated by a production efficiency model relying on the Monteith’s approach. The emphasis is put in characterizing the water stress effects on the light use efficiency and, eventually, on the GPP. Six alternatives are evaluated. Among them, the ones using the ratio between the MODIS actual and potential evapotranspiration, and the soil moisture from SMOS demonstrate that it is possible to characterize the water stress on GPP using only remote sensing products. Daily images of GPP are calculated and used as a reference in the rest of the thesis. The reference GPP is then used to calibrate and validate a semi-empirical model for the estimation of annual GPP. This model is a simplification of the Monteith approach that relies on a linear relationship between GPP and a PAR-weighted vegetation index (VI). This semi-empirical model can be used to estimate the annual GPP from commonly available VI images and a representative PAR, which does not require actual meteorological data. Therefore, inputs and computing time are considerably reduced. NDVI and EVI are tested. EVI does not need a land cover map, reducing the number of inputs even more. NDVI allows the elaboration of climatic studies that require long time series. Ecosystem respirations are simulated through Biome-BGC. A methodology to calibrate the rooting depth parameter, critical to the simulation of the water balance, is developed. The optimal rooting depth is obtained by comparing the reference GPP with the one simulated by Biome-BGC. The methodology is first tested in 4 validation sites and then extended to the whole study area. As a result, daily GPP maps and an optimal rooting depth map are obtained. Reference GPP and optimized respirations are used to calculate net fluxes. However, both GPP and respiration must be previously corrected. The first one because it represents the contribution of all the vegetation present in the considered area, not only the one from the forests. The second one because Biome-BGC works on equilibrium conditions and does not represent the actual state of the ecosystem. To do so, the necessary information layer, a growing stock map, is produced by the combination of the Third Spanish National Forest Inventory data (more than 50000 plot along the 1997-2007 period) and Landsat-5 TM and Landsat-7 ETM+ imagery (more than 8000 scenes covering the whole study area during the inventory period). Finally, preliminary net fluxes resulting from the corrected GPP and respirations are presented and validated.The main goal of this thesis is the establishment of a framework to analyze the Spanish forest ecosystems in terms of their role in the carbon cycle. In particular, the carbon fluxes that they exchange with atmosphere are modeled to evaluate their potential as carbon sinks and biomass reservoirs. Gross fluxes are estimated by a production efficiency model relying on the Monteith’s approach. The emphasis is put in characterizing the water stress effects on the light use efficiency and, eventually, on the GPP. Six alternatives are evaluated. Among them, the ones using the ratio between the MODIS actual and potential evapotranspiration, and the soil moisture from SMOS demonstrate that it is possible to characterize the water stress on GPP using only remote sensing products. Daily images of GPP are calculated and used as a reference in the rest of the thesis. The reference GPP is then used to calibrate and validate a semi-empirical model for the estimation of annual GPP. This model is a simplification of the Monteith approach that relies on a linear relationship between GPP and a PAR-weighted vegetation index (VI). This semi-empirical model can be used to estimate the annual GPP from commonly available VI images and a representative PAR, which does not require actual meteorological data. Therefore, inputs and computing time are considerably reduced. NDVI and EVI are tested. EVI does not need a land cover map, reducing the number of inputs even more. NDVI allows the elaboration of climatic studies that require long time series. Ecosystem respirations are simulated through Biome-BGC. A methodology to calibrate the rooting depth parameter, critical to the simulation of the water balance, is developed. The optimal rooting depth is obtained by comparing the reference GPP with the one simulated by Biome-BGC. The methodology is first tested in 4 validation sites and then extended to the whole study area. As a result, daily GPP maps and an optimal rooting depth map are obtained. Reference GPP and optimized respirations are used to calculate net fluxes. However, both GPP and respiration must be previously corrected. The first one because it represents the contribution of all the vegetation present in the considered area, not only the one from the forests. The second one because Biome-BGC works on equilibrium conditions and does not represent the actual state of the ecosystem. To do so, the necessary information layer, a growing stock map, is produced by the combination of the Third Spanish National Forest Inventory data (more than 50000 plot along the 1997-2007 period) and Landsat-5 TM and Landsat-7 ETM+ imagery (more than 8000 scenes covering the whole study area during the inventory period). Finally, preliminary net fluxes resulting from the corrected GPP and respirations are presented and validated.