The relationship between measured soil properties, site-specific management zones, and bare soil reflectance : Colorado, USA

Abstract

Soil productivity varies across farm fields and it is influenced by soil physical and chemical properties. Bare soil imagery can be used to delineate areas of homogeneous soil characteristics, based on variations in reflectance. The objectives of this study were to: (i) evaluate site-specific management zones on the basis of spatial variability in measured soil properties, and (ii) determine the measured soil properties whose variability could be best explained by remotely sensed bare soil reflectance data. The study was conducted on three irrigated fields near Greeley, Wiggins and Yuma in northeastern Colorado, U.S.A. Each field had previously been sub-divided into three management zones corresponding to areas of high, medium and low levels of productivity. Each field was divided into grid cells of 0.4 ha each, with one sample point per cell. The soil properties measured were bulk density, cone index, surface color, organic carbon, texture, total pore space, sorptivity; and surface water content. Surface bulk density and sand content were inversely related to the productivity level of the management zones at study sites I (Greeley) and II (Wiggins). At these study sites, organic carbon and silt content were directly related to the productivity level of the zones. At study site II, clay content and cone index at the 20 cm depth had a direct and indirect relationship, respectively, with the productivity level of the zones. The amount of variability of soil properties that was explained by the appropriate spectral bands ranged from 35 to 55% at site I (Greeley), 13 to 73% at site II (Wiggins), and 10 to 52% at site III (Yuma). In the test involving both zones and wavelength bands, some soil properties were related to either zones or bands only, while others were related to both bands and zones. The amount of variability of soil properties explained by either zones or bands, or a combination of both, ranged from 11 to 77% in Wiggins and 17 to 56% in Yuma. The variation in some of the measured soil properties explained the variable productivity of the management zones. The variation of some soil properties across a field can be explained by the variability in reflectance observed on bare soil imagery.