Noncloud Contaminants in Agricultural Soil Monitoring: Quantifying Spectral Distortions From Plastic Covers, Pylons, and Aircraft Overpasses

Abstract

The detection and exclusion of clouds and their shadows have been the primary focus of pixel contamination in spaceborne agricultural soil monitoring. However, contaminants affecting specific parts of agricultural fields, such as stationary features (large pylons and artificial soil cover) and dynamic sources (pylon shadows and passing aircraft with contrails), have been overlooked despite their importance in precision agriculture. This study investigates these underexplored sources of pixel contamination and their implications for agricultural soil monitoring. Using bare soil data from 2017 to 2023 and focusing on the field preparation period from mid-April to late May, we analyzed the effects of artificial soil cover, transmission tower shadows, and aircraft overflights on bare soil reflectance. These pixel contaminants significantly altered surface reflectance compared to clear bare soil pixels, with P $\leq$ 0.0001 for artificially covered, P $\leq$ 0.01 for aircraft-impacted and P $\leq$ 0.05 for tower shadowed pixels. Artificial cover increased the surface reflectance of bare soil by 10% to 50% in the visible and near-infrared bands, with a smaller increase of 5% in the shortwave infrared bands; pylon shadows reduced the surface reflectance by up to 5% within a 10 m buffer around the shadow. Aircraft footprints caused a sixfold increase in reflectance, with contrails affecting large areas and increasing reflectance by up to 30%. Important spectral indices for bare soil analyses were significantly affected by artificial cover, but not always by shadows or aircraft impact. The analysis provides insights into the spectral anomalies caused by pixel contaminants and highlights the need to account for such influences to improve the accuracy of spaceborne agricultural soil monitoring, particularly in small parcels or field zones.