Determination of grassland use intensity based on multi-temporal remote sensing data and ecological indicators

Grassland use intensity and its impact on biodiversity and water pollution is a topic of growing interest. In ecological studies, intensity of use has been assessed by means of three indicators: i) mowing frequency, ii) grazing intensity, and iii) fertilization input. A multidimensional approach is key for the understanding of intensification effects in terrestrial and water ecosystems. Remote sensing is a powerful tool to monitor management indicators. Nevertheless, interdependencies between remote sensing methods and between indicators require new approaches to assess intensity of use. The objective of this study is to monitor ecological indicators of land use intensity based on multispectral imagery using a multidimensional approach. We performed a multi-temporal analysis using a series of RapidEye images within a growing season in the Canton of Zurich, Switzerland, in 2013. We defined mowing frequency classes distinguishing spectral changes between pairs of images. The analysis of the whole image sequence within the growing season helped differentiate grazing intensities. Furthermore, we analysed the suitability of modelled livestock density based on remote sensing derived products to determine fertilizer input. Three grassland management practices were distinguished: i) medium intensive (46%), ii) low intensive (37%), and iii) high intensive (17%). We discuss the combination of high mowing frequency and fields with high grazing intensity to define areas prone to nutrient surpluses. Finally, we demonstrate that the estimation of interrelated indicators of grassland use intensity could be carried out preserving independence between methods.