Imaging analysis of biological soil crusts to understand surface heating...

Potter, C., and J. Weigand (2018), Imaging analysis of biological soil crusts to understand surface heating properties in the Mojave Desert of California, Catena, 170, 1-9, doi:10.1016/j.catena.2018.05.033.
Abstract: 

Biological soil crusts (biocrusts) are composed of a combination of cyanobacteria, green algae, lichens, and mosses that can protect arid soil surfaces from erosive forces. The presence of biocrusts in desert regions is influenced by a complex interaction of climatic, edaphic, and topographic variables, but their distribution and coverage densities in southern California deserts are not precisely mapped. Visible satellite bands (red-green-blue; RGB) and infra-red image resolution are generally too coarse to distinguish between biocrusts and their interstitial sand substrates, limiting most remote sensing analysis of these soil biotic assemblages. The purpose of this study was to better understand the relationships between biocrust cover, surface coloration and roughness, and the thermal infra-red (TIR) emission properties of biocrusts to determine if new types of imaging of desert surfaces could be used to more accurately detect biocrust presence. Based on the results from megapixel digital photos and TIR imaging for nearly 130 different, well-developed biocrust surfaces in the Mojave and Lower Colorado Deserts of California, we designed laboratory heating experiments to test the effect of biocrust cover on the heating properties of sandy wash habitats. Biocrust samples collected from the field were subjected to artificial heating and cooling to simulate the Mojave Desert diurnal cycle. We used image segmentation algorithms to separate true-color biocrust images from their sandy substrates, and to map the heating and cooling patterns of these surfaces with different levels of biocrust coloration and roughness. The major findings of this study showed that biocrust patch images in the field had Red texture mean values between digital numbers of 95 and 131 with moderate positive tail shewness levels, indicative of a highly rugose surface of the biocrust cover. Such patches were also well-compacted overall, with moderate skewness toward negative tails. In controlled laboratory temperature experiments to simulate the desert diurnal cycle, surface wetting with a fine mist of water resulted in significantly lower peak surface temperatures of biocrust samples compared to dry biocrust heating results of the same crust and sand samples. The findings of this study can improve future mapping for lichen-dominated biocrust surfaces in the Mojave Desert.

PDF of Publication: 
Download from publisher's website.