Effectiveness of drainage improvements in reducing sediment production rates from an unpaved road

Abstract

Increased sediment yields associated with land disturbances pose a serious threat to Caribbean coral reefs due to their sensitivity to terrigenous sediment inputs. On St. John in the US Virgin Islands, unpaved roads are responsible for increasing sediment delivery rates into coastal waters by up to nine times above background (undisturbed watershed conditions). Hence, there is a need to implement and test cost-effective erosion mitigation strategies to protect these ecosystems. The objectives of this study were to (1) design and implement a simple unpaved road erosion control strategy, (2) collect empirical data to test the effectiveness of this mitigation strategy in reducing erosion, and (3) compare the empirical data to predictions based on a locally derived empirical model and the US Forest Service's X-Drain model. The erosion-control design consisted of five cemented cross drains that directed runoff into a paved ditch and onto a rip-rap protected outlet. Sediment production rates were measured with a sediment trap during the pretreatment (July of 1998 to November of 1999) and posttreatment (September of 2003 to September of 2005) periods. The mean posttreatment sediment production rate was 7.4 kg mm⁻¹ (414 lb in⁻¹) of rainfall, or 30% of the mean pretreatment rate (24.3 kg mm⁻¹ [1,360 lb in⁻¹]). Reductions in sediment yield rates were not influenced by differences in precipitation patterns, as the lower posttreatment rates were associated with slightly higher rainfall intensities than during the pretreatment period. This study showed that the establishment of five cemented cross drains and a paved ditch resulted in a 70% reduction of sediment yields rates, and this is attributed to three factors: (1) the protective cover provided by the new paved surfaces, (2) a decrease in runoff erosive energy over the road surface, and (3) the development of a coarse-textured, and thus more resistant, road tread surface. The locally developed road erosion model and the X-Drain model underestimated mean erosion during the pretreatment period by 40% and 19%, respectively. In contrast, observed posttreatment rates were within 3% to 6% of those predicted by both models. The results of this study have stimulated the application of similar erosion control methods in other parts of St. John and should encourage widespread road mitigation throughout the rest of the US Virgin Islands and the Caribbean wherever the application of more costly alternatives are not economically feasible or desirable.