Monitoring agricultural field trafficability using Sentinel-1

The use of heavy mobile machinery in agriculture for tillage and harvesting is now indispensable since it facilitates farming over large areas. However, one of the impacts of regular and prolonged use of heavy mobile machinery is soil compaction. To help minimize this harmful effect, trafficability of agricultural fields needs to be determined. Soil moisture acts as one of the dominant controls for field trafficability. Therefore satellites such as Sentinel-1, which is one source of spatio-temporal soil moisture information, could be useful in assessing trafficable conditions. One limitation of satellite-derived soil moisture is that only the upper surface layer is mapped. In this study, we determined the feasibility of Sentinel-1 surface soil moisture to monitor trafficability over 2016-2017. We first determined coupled conditions when surface soil moisture is a good indicator for values at the subsurface. We applied a probabilistic approach to determine trafficability using extensive in situ measurements of penetration resistance and surface soil moisture over a variety of crops. Trafficability is expressed as the probability that penetration resistance will exceed a threshold, for a given soil moisture value. Furthermore, we investigated the variability encountered from these measurements to gain insights on other temporal controls. Our results show coupled conditions for soil moisture >= 0.19 cm(3) cm(-3) and there is an almost 1:1 correspondence between surface and subsurface values. For decoupled conditions, values at the subsurface can be two times more than the surface. An increase in penetration resistance variability coincided with the maturity of crops for cultivated fields. Aside from soil moisture, root growth may have a significant impact on the temporal variability of soil's penetration resistance. The status of trafficability can be monitored through the high temporal resolution of Sentinel-1. However, aggregation to coarser resolutions maybe necessary as its original 10 m resolution may be suboptimal, based on validation against in situ measurements. Days favorable for traffic were observed in early spring. This information can aid farmers in the timing of tillage activities or for water managers in deciding to adjust water levels to meet agricultural demands.