**5. Conclusions**

This study examined the effect of precipitation parameters on soil erosion through six years of high-resolution weekly monitoring in an Appalachian hillslope paying particular attention to seasonal effect. The long-term data provided an understanding of the seasonal pattern of soil erosion in a humid sub-tropical environment, which was not noticeable in other studies in the region using an annual dataset.

Different gully morphologies responded differently to long-term erosion. Channels were most active, showed a wide range of variability, and responded most dynamically, whereas the interfluves were least disturbed by erosion. Sidewalls were prone to erosion but were not as dynamic as channels. To explore the reason behind varied gully erosion patterns in the different geomorphic settings, further studies are recommended to evaluate how erosion fluctuates with soil cover thickness, soil types, moisture contents, slope aspect, and slope angle.

Precipitation duration was the most important factor in initiating and continuing erosion year-round, yet seasonality played a significant role in the severity of gully erosion. Erosion was most pronounced in winter months, followed by spring, indicating the influence of high-intensity precipitation from frontal systems and repeated freeze-thaw cycles. Erosion in summer was driven by high-intensity precipitation from convectional storms. Soils in the study area were least prone to erosion during the moderate months of autumn. In channels, precipitation duration was the dominant driver for erosion due to runoff-related erosion, while in sidewalls and interfluves, intensity parameters were equally important as duration, likely related to rain splash erosion. This research shows that soil erosion is seasonally variable and an understanding of the seasonal pattern of soil erosion with respect to precipitation-related drivers improves the potential to achieve strategic conservation measures.

**Author Contributions:** Conceptualization, I.L. and A.N.; methodology, I.L. and A.N.; data collection, I.L. and A.N.; formal analysis, I.L. and A.N.; funding acquisition, I.L. and A.N.; data curation, I.L.; writing—original draft preparation, I.L. and A.N.; writing—review and editing, I.L. and A.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received funding support from East Tennessee State University's Honors College Federal Work Study program for collection of field data.

**Acknowledgments:** The authors gratefully acknowledge the assistance in data collection provided by Tim Spiegel, Nicholas Barnes, Tim Land, Jamie Kincheloe, Nicholas McConnell, and Jennifer Grant. The authors are grateful for the valuable contribution of the anonymous reviewers.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

OLS Regression models of erosion are presented in Table A1. While model equations are useful for prediction when determination coefficients are high, even when they are relatively low, useful information can be revealed with respect to the importance of independent variables. Standardized coefficients can likewise provide information about the relative importance of independent variables within each model. For channels, Duration and TotAcc during the current and prior measurement periods were retained most often, and these variables had the highest standardized coefficients compared to the intensity parameters (AvgInt and MaxInt) (standardized coefficients are not shown in the table). For interfluves, AvgInt and MaxInt were also retained in the models, and for the IDep and IErosion models, standardized coefficients for all retained variables were of similar magnitudes. For sidewalls, a similar pattern was generally noted, with retention of the intensity variables. For the SErosion model, Duration and TotAcc parameters had the largest standardized coefficients.

**Table A1.** Regression equations for erosion variables (dependent variables) using lagged precipitation parameters (independent variables). Lagged variable names are appended with "LagN", where N indicates the number of measurement periods of antecedent lag. Duration\_Lag1 indicates precipitation duration in prior measurement period (Lag of 1 period).

