Decadal scale influence of rainfall induced landslides on fluvial sediment transport following typhoon Morakot

Landslides are of wide interest because they shape topography and influence fluvial sediment transport. To investigate the duration of landsliding's influence on fluvial sediment discharge, we analyzed suspended sediment concentration measurements at 77 gauging stations across Taiwan over an 11-year period after Typhoon Morakot in 2009, the wettest typhoon on record in Taiwan which generated nearly 20,000 landslides. At each gauging station, we computed annual rating curves for suspended sediment concentration as a power-law function of the centered water discharge, which isolates the efficiency of suspended sediment discharge from temporal variations in water discharge. Among the 40 stations in basins that were strongly impacted by landsliding, the discharge-normalized rating curve coefficient ã increased by a mean factor of 4.89 within 1-2 years of Morakot, while the rating curve exponent b exhibited no systematic response to Morakot (mean factor of 1.21). Elevated values of ã declined exponentially at 26 of the 40 stations with a median characteristic timescale of 13.2 years (interquartile range: 7.6-20.7) and tended to respond faster in basins with more intense landsliding. In contrast, at the 37 stations that were not impacted by landsliding, neither ã nor b exhibited a systematic response to Morakot. To quantify the effect of landsliding on sediment discharge, we compared the measured sediment discharges after Morakot to the hypothetical sediment discharges that would have occurred if no change in sediment transport efficiency had occurred at the time of Morakot, calculated by applying the post-Morakot water discharge history to pre-Morakot rating curves. This analysis suggests that Morakot-induced landsliding increased sediment discharge by as much as >10-fold in some basins in southeast Taiwan in the 1-2 years after Morakot. Post-Morakot changes in ã were positively correlated with landslide intensity for approximately seven years after Morakot, and changes in b were negatively correlated from 2011 to 2014, indicating that the influence of landsliding on rating curves diminished within a few years. Together, these results suggest that Morakot-induced landsliding amplified fluvial sediment fluxes for a relatively short time (< 10 years). To the extent that these results are applicable to rivers in other landscapes, this suggests that rivers may more efficiently transport landslide-derived sediment for relatively short times.