Estimation of Mediterranean forest transpiration and photosynthesis through the use of an ecosystem simulation model driven by remotely sensed data

Aim This paper investigates the use of an ecosystem simulation model, FOREST-BGC, to estimate the main ecophysiological processes (transpiration and photosynthesis) of Mediterranean coastal forest areas using remotely sensed data. Location Model testing was carried out at two protected forest sites in central Italy, one of which was covered by Turkey oak (Circeo National Park) and the other by holm-oak (Castelporziano Estate). Methods At both sites, transpiration and photosynthesis measurements were collected in the field during the growing seasons over a four-year period (1999 and 2001 for the Turkey oak; 1997, 1999 and 2000 for the holm-oak). Calibration of the model was obtained through combining information derived from ground measurements and remotely sensed data. In particular, remote sensing estimates of the Leaf Area Index derived from 1 x 1-km NOAA AVHRR Normalized Difference Vegetation Index data were used to improve the adaptation of the model to local forest conditions. Results The results indicated different strategies regarding water use efficiency, 'water spending' for Turkey oak and 'water saving' for holm-oak. The water use efficiency for the holm-oak was consistently higher than that for the Turkey oak and the relationship between VPD and WUE for the holm-oak showed a higher coefficient of determination (R-2 = 0.9238). Comparisons made between the field measurements of transpiration and photosynthesis and the model estimates showed that the integration procedure used for the deciduous oak forest was effective, but that there is a need for further studies regarding the sclerophyllous evergreen forest. In particular, for Turkey oak the simulations of transpiration yielded very good results, with errors lower than 0.3 mm H2O/day, while the simulation accuracy for photosynthesis was lower. In the case of holm-oak, transpiration was markedly overestimated for all days considered, while the simulations of photosynthesis were very accurate. Main conclusions Overall, the approach offers interesting operational possibilities for the monitoring of Mediterranean forest ecosystems, particularly in view of the availability of new satellite sensors with a higher spatial and temporal resolution, which have been launched in recent years.