Using of MODIS NDVI Time Series for Grassland Habitat Classification and Assessment A. Halabuk 1 , M. Mojses 1 1 IInstitute of Landscape Ecology, Slovak Academy of Sciences, Branch Nitra, Akademická 2, 94901 Nitra, Slovakia Email: Andrej.Halabuk@savba.sk 1. Introduction Satellite images with high spatial resolution are widely used for habitat assessment and surveillance. Their low temporal resolution however, limits their ability for regular monitoring of habitats in adequate time span. Recently, after the launch of the Terra and Aqua satellites with the MODIS (Moderate Resolution Imaging Spectroradiometer) on board, new approaches started to be more frequently used for habitat classification and monitoring. One of the most widespread approaches use time series analysis of vegetation indices (e.g. NDVI – Normalized Difference Vegetation Index) that reflects the temporal profile of vegetation greenness on the land surface. Semi-natural grasslands in agricultural landscape bear high biodiversity values and there is still lack of precise information on their spatial extent and status at pan European scale. One of the main criteria for good status of the semi-natural grasslands is their extensive usage, e.g. regular cutting and/or grazing. In order to get such information, detection of site management is needed for longer periods (e.g. 10 years) in order to reveal trends for possible abandonment or intensification. Because of gradual availability of time series products from sensors such as MODIS and under a great expectation of upcoming Sentinel3 mission, we analyzed here suitability of MODIS NDVI time series at 250m spatial resolution for classification and assessment of grassland habitats in Slovak heterogeneous landscape. Particularly, we focused on detection of cutting practices, overgrowing, flooding, overgrazing, which are all considered as important determinants of grassland habitat quality. 1.1 NDVI time series for grassland classification and assessment Grasslands with similar physiognomy may have different temporal pattern of NDVI affected by a broad range of natural or human driven factors. Multitemporal analysis of NDVI time series iteratively explores the main determinants of seasonality and uses this information for the subsequent classification of grasslands to determine their vegetation type, status and functioning. Grassland classification may include both full coverage classification of grassland areas or exploration and classification of main grassland types based on sampled areas. For example Aragon and Oesterheld (2008) used a combined approach using information of spatial arrangement (from single date HR Landsat TM image) and information on functional properties (NDVI dynamics derived from multitemporal MODIS 250m NDVI series) to map grassland vegetation communities in Argentinean flooded Pampa grasslands. The authors successfully classified 5 grassland vegetation types with an overall accuracy of 76% and documented that grassland vegetation communities significantly differ in their seasonal and interseasonal pattern of NDVI. Hill et al. (1999) classified a pastoral landscape in eastern Australia resulted into 8 broad categories: sown perennial pastures, sown perennial pastures with woodland, sown annual pastures, mixed pasture and cropping, native pastures, native pastures with woodland, degraded or revegetated areas and forest. Paruelo et al. (2001) used NDVI dynamics as a descriptor of ecosystem functioning and widely applied this Page 71