Facilitating Synergies Between Free EO Datasets for Precision Farming

Abstract:

Available online tools for space-borne monitoring of vegetation phenology often focus on one satellite sensor. However, combining various sensors allows for synergies between different datasets and alleviation of shortcomings related to sensor-specific compromise between spatial and temporal resolution. We demonstrate how these concepts have been implemented in SatAgro, an online precision agriculture platform for individual farmers. In this project, we combine MODIS and Landsat data to allow user-friendly daily access to information about plant phenology, coupled with Landsat-based insight into spatial detail at the land parcel level. We developed an approach to combine MODIS Terra and Aqua data which utilizes quality information contained in MODIS product with pixel-based NDVI to select the more reliable retrieval. Phenology-related statistics are estimated not only based on time series of vegetation indices, but also ancillary meteorological data. Landsat data is processed using LandCor atmospheric correction framework, which considers within-scene atmosphere variability, surface anisotropy, and varying acquisition geometry. The effect of particular improvements in data pre-processing is demonstrated through estimates in spatio-temporal consistency of phenological evolution across space and time and between the two sensors. Lastly, we discuss the potential to integrate further sensors in the platform, in particular from Sentinel satellites.