Quantifying decadal-scale erosion rates and their short-term variability on ecological sites in a semi-arid environment

- We measure erosion rates under simulated rainfall on six ecological sites.
- We determine decadal erosion on these sites using cesium-137 method.
- Year to year erosion rates on the same plots exhibit great variability.
- Short term variability is attributed to natural and management changes.
- Cesium-137 method showed significant differences between the states of STM.

Soil erosion rates on six semi-arid loamy upland rangeland sites located in southeastern Arizona were measured using a rainfall simulator and 137Cs fallout methods. Site characteristics that have the greatest effects on soil erosion and runoff were identified. Long term (50 years) soil erosion rates as estimated using 137Cs method varied between 5.1 and 11.0 Mg ha− 1 y− 1 and showed significant differences between Historic Climax Plant Community and Mesquite/Native states within the State and Transition Model. Erosion rates under simulated rainfall were measured between 0.9 and 17.2 g m− 2 min− 1 at 175 mm h− 1 precipitation across all sites and varied as much as 8-fold at the same location, depending on the time of the simulation. Temporal variability of erosion rates within a site was in some cases much greater than inter-site differences. This variability was attributed to natural or management driven changes in plant community and soil characteristics. Bare soil area, an aggregate indicator for all types of cover combined, was the main controlling factor of erosion process across ecological sites. For meaningful interpretation rainfall simulation, results must be placed in the context of the range of possible vegetation and surface conditions within a given ecological site.