Thu, Aug 18, 2022: 8:15 AM-8:30 AM
518C
Background/Question/MethodsChange from native terrestrial ecosystems to novel grazing areas for ungulates has facilitated the invasion of non-native grasses and the decrease in native vegetation cover. Reduced vegetative cover and increases in sheet erosion are related to plant and ungulate invasion. Damage caused by ungulates, and access roads maintained to manage them, increases mass wasting potential thus priming the landscape for erosion. A major conservation challenge is collecting landscape-level information about land cover change and the mass movement of soils associated with acute, pulsed rainfall events. Our goal was to conduct high-precision change detection of land cover and surface features between adjacent grazed and non-grazed plots and between a two-year period that included several mass sediment flux events associated with tropical storms. To detect terrestrial erosion and land cover changes we used Unmanned Aerial Systems (UASs) to collect imagery of the plots in the watershed of Kaʻamola, Molokai, Hawaii in 2021 and surveyed the grazed plot in 2019 and 2021. Images were processed to generate orthomosaics and digital elevation products that were classified using a rulset based classification in the eCognition software. Spatial analyses in ArcGIS Pro were used to identify change in land cover and topographic indicators of erosion.
Results/ConclusionsWe found that the management of watershed elements for livestock, with associated infrastructure, has greater proportional areas of bare soil, rock, and grass land cover types when compared to the non-grazed plots. Proportionally, the grazed plots have more than double the area of grass and rock land cover types as well as almost 50% more bare soil cover present as compared to the non-grazed plots. Our preliminary results also indicate that slope and adjacent land cover type are both strong indicators of future erosion. By coordinating with land managers, we are now able to guide both restorative and preventative management decisions to mitigate erosion and sediment flux off of the watershed and onto the reefs. This is a workflow that can be applied to time series sets of imagery to better guide preventative action planning at a high level of spatial precision and accuracy. This approach to detecting change within a two-year period and between grazed and non-grazed land showcases the advantages of using UAS derived imagery products to highlight key correlates of erosion used in guiding management decisions.
Results/ConclusionsWe found that the management of watershed elements for livestock, with associated infrastructure, has greater proportional areas of bare soil, rock, and grass land cover types when compared to the non-grazed plots. Proportionally, the grazed plots have more than double the area of grass and rock land cover types as well as almost 50% more bare soil cover present as compared to the non-grazed plots. Our preliminary results also indicate that slope and adjacent land cover type are both strong indicators of future erosion. By coordinating with land managers, we are now able to guide both restorative and preventative management decisions to mitigate erosion and sediment flux off of the watershed and onto the reefs. This is a workflow that can be applied to time series sets of imagery to better guide preventative action planning at a high level of spatial precision and accuracy. This approach to detecting change within a two-year period and between grazed and non-grazed land showcases the advantages of using UAS derived imagery products to highlight key correlates of erosion used in guiding management decisions.