The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Articles | Volume XL-7/W3
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XL-7/W3, 15–18, 2015
https://doi.org/10.5194/isprsarchives-XL-7-W3-15-2015
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XL-7/W3, 15–18, 2015
https://doi.org/10.5194/isprsarchives-XL-7-W3-15-2015

  28 Apr 2015

28 Apr 2015

Satellite-based assessment of grassland yields

K. Grant1, R. Siegmund2, M. Wagner2, and S. Hartmann1 K. Grant et al.
  • 1Institute for Crop Science and Plant Breeding, Bavarian State Research Center for Agriculture (LfL), Freising, Germany
  • 2GAF AG, Munich, Germany

Keywords: Radar, Change detection, Sentinel-1, Grassland, Cutting date, Forage yield model

Abstract. Cutting date and frequency are important parameters determining grassland yields in addition to the effects of weather, soil conditions, plant composition and fertilisation. Because accurate and area-wide data of grassland yields are currently not available, cutting frequency can be used to estimate yields. In this project, a method to detect cutting dates via surface changes in radar images is developed. The combination of this method with a grassland yield model will result in more reliable and regional-wide numbers of grassland yields. For the test-phase of the monitoring project, a study area situated southeast of Munich, Germany, was chosen due to its high density of managed grassland. For determining grassland cutting robust amplitude change detection techniques are used evaluating radar amplitude or backscatter statistics before and after the cutting event. CosmoSkyMed and Sentinel-1A data were analysed. All detected cuts were verified according to in-situ measurements recorded in a GIS database. Although the SAR systems had various acquisition geometries, the amount of detected grassland cut was quite similar. Of 154 tested grassland plots, covering in total 436 ha, 116 and 111 cuts were detected using CosmoSkyMed and Sentinel-1A radar data, respectively. Further improvement of radar data processes as well as additional analyses with higher sample number and wider land surface coverage will follow for optimisation of the method and for validation and generalisation of the results of this feasibility study. The automation of this method will than allow for an area-wide and cost efficient cutting date detection service improving grassland yield models.