|Title||Meteoric beryllium-10 as a tracer of erosion due to postsettlement land use in west-central Minnesota, USA|
|Publication Type||Journal Article|
|Year of Publication||2019|
|Authors||Jelinski N.A, Campforts B., Willenbring J.K, Schumacher T.E, Li S., Lobb D.A, Papiernik S.K, Yoo K.|
|Type of Article||Article|
|Keywords||agriculture; Beryllium-10; cesium-137; cs-137; denudation rates; erosion; Geology; landscape; radionuclide; river-basin; sediment transport; soil-erosion; tillage; water erosion; weathered soils|
Meteoric beryllium-10 (Be-10(m), t(1/2)=1.4Myr) is a cosmogenic radionuclide that remains largely underutilized for deriving hillslope-scale estimates of erosion on uplands under conditions of land use change. We applied two different models for estimating erosion rates from observed Be-10(m) concentrations (a one-dimensional model predicting vertical profiles of Be-10(m) within hillslope soils [loss only, diffusion only, LODO] and a two-dimensional model predicting the concurrent evolution of hillslope topography and Be-10(m) distributions via bioturbation, chemical mobility, and surface erosion [Be2D]). Both models were used to derive pre-European and post-European settlement erosion rates (E-nat and E-post, respectively) across paired cultivated and uncultivated hillslopes in west-central Minnesota, USA. E-post estimates from Be-10(m) were compared to E-post estimates derived from Cs-137 inventories and the process-based Water and Tillage Erosion Model (WaTEM). The results from these models suggest that erosion rates from upper positions on the cultivated hillslope have increased from an average of 0.047mm/year under natural conditions to E-post values of 3.09mm/year. The Be2D and LODO models, on average, produced E-post estimates that were similar in magnitude to WaTEM and Cs-137 conversion models. This numerical convergence does not imply absolute Be-10(m) model accuracy, particularly when considering the uncertainties inherent in each approach, but it does suggest that the orders of magnitude increase in estimated erosion rates from E-nat to E-post is robust. Additionally, the pattern of E-post estimates produced using Be-10(m) conversion models is supported by the distribution of soil inorganic carbon at the study site. Our results demonstrate that Be-10(m) can provide reasonable estimates of both predisturbance and postdisturbance erosion rates in landscapes that have undergone extensive human modification. Plain Language Summary Agricultural practices have substantially changed soil erosion rates in the Midwestern United States. Although much work has been devoted to understanding the changes in soil erosion rates with land cover change, the ability to quantify those changes at discrete locations on the landscape over long periods of time has been limited. We use a set of tracers and models to estimate presettlement and postsettlement erosion rates on a hillslope in west-central Minnesota, USA, and show that soil erosion has increased by approximately 1 to 2 orders of magnitude over a period of approximately 110years. This has implications for how we view our current agricultural landscapes and how we think about soil sustainability in the future.