Structural-functional model to remote sensing of vegetation physiognomies seasonal variation based on life form

There are in the literature, many attempts to classify vegetation-using indices derived from orbital sensors. The relationship between remotely sensed data and the vegetation structure has been the object of study both for experts in environmental remote sensing and plant ecology. One difficulty to establish these relationships is due to the vegetation dynamics. In cases where the vegetation undergoes seasonal variation in its green biomass, as occurs in the Brazilian Cerrado, misclassification may take place. To minimize such effects, a semi-empirical model of NDVI seasonal Cerrado variation was developed. The objective of this study is to improve understanding of the seasonal variation of Cerrado vegetation indices derived from orbital sensors, using a model that couples reflectance with the proportional contribution of its life forms seasonally. The Cerrado tract used as a case study is located in São Paulo State (21°37'30" S, 47°37'30" W) and ranges from grassland to forest sub-types. Two approaches were conducted to model the Canopy: one was a simplified seasonally structure canopy and another a not temporal relatively complex canopy structural 3D model. The procedure followed an aggregation method, using the PROSPECT model to generate transmittance and reflectance of green leaves and the SAIL model to generate canopy reflectance. The results obtained were compared with 9 Landsat-TM images. The NDVI obtained from those Landsat-TM images show a high coincidence with the curves generated by the model, throughout the range of plant physiognomies. During the growing season, the grassland showed values smaller than those predicted by the model but the remaining subtypes had an encouraging level of coincidence with the model. On second approach, the SPRINT model was used to model a gradient of physiognomies sampled in the field over a 39 permanent plots (Leaf Area Index and phytossociological parameters). This method of analysis of seasonal variation by modeling NDVI derived from empirical models and meteorological data, and the relationship between spatial explicit structure of the vegetation and orbital remote sensing provide a measurable condition to quantify the relative seasonal variation of the Cerrado physiognomies by orbital sensor.