Results/Conclusions Long-term erosion bridge measurements, consistent with other data from this watershed, reveal that microtopographic changes in the spatial distribution of erosion and aggradation are a function of: a) ground-cover type/pattern, b) precipitation amount/intensity, and c) vegetation changes. Microtopographic data show that net erosion within the watershed essentially only occurs where abiotic processes now dominate. These abiotic zones, ordered from highest erosion rates to lowest, are: channels (-3.56mm annual average); rills (-3.38mm/yr); bare Alfisols (-1.22mm/yr); bare pumice (-0.95mm/yr); and bedrock (no change). High-intensity summer monsoon rain is the dominant erosive driver on this hillslope, and ancillary data show that the single most intense rainfall event of a summer can cause nearly all the erosion measured that season. In contrast, watershed zones dominated by biotic processes exhibit dramatically lower erosion rates. Patches with cryptobiotic soil crust display no or little soil loss (-0.18mm annual average), as cyanobacteria present in these crusts bind soil particles. Large litter patches and vegetation clumps of all sizes shield the underlying soil, preventing extensive erosion; the upslope side of these patches, as well as fallen logs, act as sediment “sinks” and catch sediments eroded from upslope “source” areas. The relative areas and spatial patterning of sediment sources and sinks (particularly including connectivity) result in a complex and dynamic flux of water and sediment on this hillslope, reflected in these microtopography observations. Fluctuating environmental conditions or disturbances can force sediment sinks (mainly vegetation) to shift toward lower abundance and/or connectivity – conversely causing increased connectivity of source areas, triggering accelerated erosion. Microtopography changes confirm that ongoing erosion is degrading the old, shallow (~50 cm thick) soils on this hillslope, with associated ecosystem feedbacks.