Interacting contributions of climate and land use change to sediment yield in an agricultural geographically isolated wetland from 1940 to 2022

Soil loss due to erosion is a widespread problem in agricultural landscapes. However, a scarcity of long-term datasets and lack of reliable models challenge our understanding of the mechanisms that cause erosion. We developed a novel method for merging sediment core analysis with event-based sediment yield modeling using the Modified Universal Soil Loss Equation. The method was applied to a geographically isolated wetland in southwestern Georgia for a period from 1940 to 2022 using the Soil and Water Assessment Tool. This period included transitions from small-scale farm operations to plantation forestry to industrial, irrigated agriculture. Our results show a distinct spike in sedimentation rate in the 1970s with a peak of 872 mg per cm² of wetland area per year, or 3.6 metric tons of erosion per ha of catchment area. This spike occurred during a transitional period in land use that was also characterized by several large winter rain events. Both sediment core and modeled data indicate that these episodic sedimentation events dominate the long-term sediment record. Disagreement between the temporal pattern of sedimentation predicted by the model and those calculated from the sediment core was evident in certain parts of the sediment record. Our comparison suggested that sediment yield models require a unique calibration to accurately represent historical agricultural practices. The analysis also provides evidence that autochthonous sedimentation rates may be increasing in the most recent decade of the record due to anthropogenic changes.