Mapping gilgai micro-relief and its impact on dryland agricultural landscapes using time series of NDVI derived from Sentinel-2 imagery

Remote sensing is a natural tool for mapping soil spatial differences, and it is widely used for digital soil mapping. The new generation of sensors, e.g., those carried by the twin Sentinel2 satellites, can describe crop canopy development with unprecedented high spatial and temporal resolution. These are excellent capabilities to describe the vegetation and have been used in agronomy for mapping within-field differential soil fertility in agricultural areas. These maps, known as Management Zones Maps, are key elements used in precision agriculture to improve the supply of scarce resources such as water, nutrients, seeds, among others. In this chapter, we present a case study of mapping gilgai micro-relief through its impact on crop canopy structure in dryland agricultural paddocks, in the Mallee region, Victoria, Australia. Our methodology uses the time series of NDVI of Sentinel 2 images over 3 years and takes advantage of the methodology for mapping management zones in crop fields. In this way, MZM from Sentinel 2 images exhibit granular structures at high spatial frequency in some paddocks. The granular structure mimics a bi-dimensional lattice very similar to that described by gilgai micro-relief. The granular structure remains undisturbed throughout three consecutive campaigns in the 2019-2021 years, under different crops, which apparently points to the soil variability as the origin of this structure. Therefore, we discuss MZM describes the gilgai subsurface micro-relief. The results enable us to assert that crop and pasture growth is likely to be uneven across a paddock in gilgai areas. This chapter demonstrates how to map this differential behaviour using a time series of Sentinel2 images, providing insights about the gilgai soil structure. These results can reinforce the approach about using vegetation as a proxy of the soil structure, providing remote sensing-based tools able to measure spatial vegetation structure characteristics.