SENSITIVITY OF FLUVIAL SOURCE CONTRIBUTIONS IN A MIXED LAND-USE ENVIRONMENT: OTTER CREEK, FORT KNOX, KENTUCKY

In mixed land-use watersheds where excess sediment is a problem, there is inherent difficulty in predicting source contributions. To optimize land-use management strategies, it is important to identify the main sources of sediment. Over a 56-week period, we examined the spatial and temporal variability in sediment-source contributions to Otter Creek, part of the Lower Otter Creek drainage area (103km²), which drains Fort Knox’s tracked-vehicle training areas and adjoining agricultural areas in Hardin and Meade counties, Kentucky (USA). The main objectives of this research were to 1) identify and differentiate characteristics of civilian and military source soils to Otter Creek that are unique and 2) apportion source contributions of fine-grained sediment to each of the five source categories. We hypothesized that the primary sources of fine-grained sediment to Otter Creek were derived from areas of average and extreme erosion which, were Fort Knox’s tracked-vehicle training areas within the watershed. Source-soil samples were classed in five categories: civilian riverbanks; military riverbanks; military forest (defined as tree canopy present); military average erosion (defined as surface soils with vegetation present but soil easily visible); and military extreme erosion (defined as surface soils with little to no vegetation present and well-defined rills and gullies). Samples were analyzed for 18 tracers and utilizing multivariate statistics we identified the optimal tracer fingerprint composite; TOC, Cu, Zn, Sr, Rb, Co, Ni Al, Na, Si, Ca, Mg and Fe. Results from the unmixing model indicated that at baseflow the highest percentage of fine-grained sediment was eroded from military forest (60%), followed by military riverbanks (16%), civilian riverbanks (10%), military average erosion (14%) and military extreme erosion (1%). During stormflow conditions erosion was from military forest (47%), followed by civilian riverbanks (26%), military riverbanks (12%), military average erosion (14%) and military extreme erosion (1%). Karst terrain within the study area could be a potential sink for sediment in areas of high to extreme erosion and further study would be needed. The results of this study demonstrate the potential of sediment fingerprinting to distinguish among multiple source contributions.