Developing MODIS time series for monitoring vegetation conditions: preliminary results

摘要

Satellite remote sensing imagery is being widely used in monitoring human induced or phenological changes in natural and agricultural environments. Inner-annual variations in vegetation conditions can be systematically studied with the use of frequently acquired data, thus offering the potential for operational monitoring of these changes. Such is the case of MODIS sensor, which provides daily images of the land surface and vegetation index products at various spatial resolutions. In this paper, 250 m MODIS daily images were used to generate 16-day composite NDVI time-series for five selected sites in Greece, representing different types of natural and managed vegetation. Specific objectives were to investigate the response of the composites to the seasonal changes of vegetation for the time period from May 2003 to October 2003, and whether variations between vegetation types are detected. Additionally, a comparison was conducted to examine the level of similarity between the 16-day NDVI composites produced and the standard MODIS VI products (MOD13Q1). The results displayed a relatively good response to phenological development and variations were described satisfactorily, while significant differences between the composites occurred for most of the test sites. The noise observed in the NDVI trends hindered a more precise vegetation analysis, however, the reduction or removal of irregularities is a subject for future work.