Biocrust community survey techniques
In April of 2018, we used a non-destructive method to survey the paired plots in the 25 selected sites, the cover of individual biocrust bryophyte and lichen species, cyanobacterial community types (i.e., “dark” and “light”; see below) were measured, and other ground cover at the soil surface such as rocks, litter, and scat were noted. Cyanobacterial biocrust cover was not classified taxonomically but categorized as lightly-pigmented vs. darkly-pigmented because of the impossibility of species identification of cyanobacteria in the field; a darker level of pigmentation is indicative of successional development and function (Belnap et al . 2008; Lan et al . 2017). To collect cover data, we placed ten 25 x 25 cm gridded point intercept quadrats per plot, recording 20 intersections for each (Jonasson 1983). To place the quadrats, each plot was divided into 12 sections of 0.5 x 0.5 m. We placed quadrats in the bottom left corner of the 10 outer sections. Within each gridded quadrat, we aligned a pin flag with each grid intersection and extended it downward to the soil surface, recording the species or ground cover element intercepted. An intersection accounts for 5% cover within that framework. To account for species or cover types that the pin flag did not hit, but were within the quadrat, we searched the quadrat for additional species and estimated a visual percentage of cover to the nearest one percent. When a moss or lichen species was unknown, we assigned a genus or consistent morphospecies name to separate it from other species and take it into account in our diversity estimates. All evaluations were made when biocrust species were dry and inactive. Because our assessment was non-destructive, specimen collection was not possible, and some species were not identified to the species level. For all 10 sections per plot, we calculated the percentage cover of all visible components. We subtracted the cover of non-colonizable soil surface (i.e., plant litter, rock) and recalculated the cover of all major functional groups: bryophytes, lichens, cyanobacteria, and bare ground, as well as the Shannon diversity index (Shannon & Weaver 1949) based on available habitat. We then averaged the values of the 10 cover frames to calculate cover at the plot level. In the absence of a biocrust diversity survey of the plots at the time of installation, we used the diversity data of the control plots as our best available proxy for the initial biocrust diversity, due to the proximity of the plots and similar environmental conditions.