Search for: All records

Biological soil crusts (BSCs) consist of cyanobacteria, algae, fungi, lichens, and mosses, which live within the uppermost millimeters of the soil’s surface where they influence soil stability, nitrogen, and carbon cycles. BSCs are only a few millimeters thick but cover large expanses of ground allowing for possibly significant contributions to the carbon cycle in arid environments. Remote sensing of these organisms has been used to study BCS in a non-destructive manner over the last 36 years. In this project, we aimed to quantify the abundance and distribution of BSCs and to estimate their annual carbon fixation rate at different community stages in two unique research sites (T-East and T-West) in the Chihuahuan desert of southern New Mexico, USA. We first investigated the utility of an unoccupied aircraft system (UAS)-mounted hyperspectral camera to measure the spatial cover of two BSCs functional groups (light and dark BSC) in a shrub-dominated vs. predominantly grassland site. Using a spectral angle mapper (SAM) algorithm we classified the hyperspectral imagery into five cover classes (light BSC, dark BSC, bare soil, grasses, and shrubs). With observations collected utilizing a line point intercept (LPI) method, we quantified the percent cover of BSCs and vegetation within the transects to validate the SAM estimates. Lastly, we incorporated C-fixation rates from local BSC communities to estimate the mean annual C-fixation rates for both BSCs within our research plots. The SAM algorithm overestimated light biocrust (LBC) in T-East by 17.7% when compared with the LPI estimates. Inversely, the SAM underestimated dark biocrust (DBC) for T-East by 9.3% as well as underestimating both Light and DBCs for T-west (23.9% and 6.7% respectively) when compared to the LPI estimates. This produced a higher annual fixed CO2 value for T-East’s LBC of 36.4% and a lower rate for DBC of 37.2% in comparison with the LPI’s observations. The T-West annual fixed CO2 was underestimated by the SAM algorithm by 66.1% for LBC and 42.9% for DBC when compared to the LPI observations. The results indicate the heavily shrub-encroached T-East site had a better accuracy rate than the T-West grassland site for the accurate classification of the five cover classes. It was believed that the vegetation's close proximity to the soil surface and sample timing negatively interfered with the detection rates of light and dark BSCs.