Estimation of grassland use intensities based on high spatial resolution LAI time series
- 1Institute for Applied Remote Sensing, EURAC Research, Bolzano, Italy
- 2Department of Remote Sensing, Institute of Geography, University of Wuerzburg, Wuerzburg, Germany
- 3German Aerospace Center (DLR), German Remote Sensing Data Center (DFD), Oberpfaffenhofen, Germany
Keywords: Leaf area index (LAI), radiation transfer modeling, RapidEye, Alpine area, grassland, management intensity
Abstract. The identification and surveillance of agricultural management and the measurement of biophysical canopy parameters in grasslands is relevant for environmental protection as well as for political and economic reasons, as proper grassland management is partly subsidized. An ideal monitoring tool is remote sensing due to its area wide continuous observations. However, due to small-scaled land use patterns in many parts of central Europe, a high spatial resolution is needed. In this study, the feasibility of RapidEye data to derive leaf area index (LAI) time series and to relate them to grassland management practices is assessed. The study area is the catchment of river Ammer in southern Bavaria, where agricultural areas are mainly grasslands. While extensively managed grasslands are maintained with one to two harvests per year and no or little fertilization, intensive cultivation practices compass three to five harvests per year and turnover pasturing.
Based on a RapidEye time series from 2011 with spatial resolution of 6.5 meters, LAI is derived using the inverted radiation transfer model PROSAIL. The LAI in this area ranges from 1.5 to 7.5 over the vegetation period and is estimated with an RMSE between 0.7 and 1.1. The derived LAI maps cover 85 % of the study area’s grasslands at least seven times. Using statistical metrics of the LAI time series, different grassland management types can be identified: very intensively managed meadows, intensively managed meadows, intensively managed pastures, and extensively managed meadows and moor. However, a precise identification of the mowing dates highly depends on the coincidence with satellite data acquisitions. Further analysis should focus therefor on the selection of the temporal resolution of the time series as well as on the performance of further vegetation parameters and indices compared to LAI.