Mapping Biological Soil Crust Cover in the Kawaihae Watershed

Abstract

Historical land use patterns on Hawai’i Island have created degraded dryland ecosystems that are at high risk for erosion. In places such as the Kawaihae watershed in leeward Kohala, the impacts of sediment deposition from the watershed have detrimentally affected coastal marine ecosystems by decreasing habitat quality and burying important cultural sites. Due to the extensive and long-term effects of erosion, it is important to understand and protect non-traditional agents that may help to hold sediments in place. Biological soil crusts are communities of photosynthetic microorganisms that grow over mineral soil in arid and semi-arid ecosystems and are known to increase soil stability. Despite their potential to mitigate erosion, the distribution of biocrusts in degraded drylands on Hawai’i Island is unknown. We mapped biocrusts in the Kawaihae watershed, a semi-arid landscape prone to erosion, using imagery collected by small unmanned aerial systems (sUAS) at three spatial resolutions (1.15, 2.05 and 2.80 cm/pixel). Using a pixel-based methodology, we produced classifications with overall accuracies ≥85% at all three resolutions. As biocrust development is associated with increasing soil stability, we also explored this relationship in the Kawaihae watershed. We identified 3 different biocrust levels of development (LOD) and conducted soil aggregate stability testing at all development levels. We found that there was a significant increase in soil stability between soils without surface biocrusts (LOD score of 0) and those with biocrusts at any development level (LOD 1-3). More highly-developed biocrusts imparted greater soil stability than less-developed biocrusts, but the impact on soil stability reached a ceiling beyond biocrust LOD 1. In addition, we applied our mapping methodology to investigate the direct impacts of biocrusts on soil loss. We overlaid our classified maps with digital surface models (DSMs) from data sets covering a 2.75-year time span. We found trends of varying soil loss between biocrust and bare soil areas, but more field work is needed to verify our results. We also explored the effects of grazing animals on biocrust cover by comparing classified images of a grazed site and a grazing-exclusion site. We found differences in biocrust coverage between the sites, including differing proportions of the land cover types present, but additional field data collection is necessary prior to drawing definitive conclusions. Overall, our project provides a new biocrust mapping methodology that could be used by researchers and land managers globally and adds insight into the role of biocrusts in erosion prevention in the Kawaihae watershed and similar arid/semi-arid landscapes.