Patterns and rates of landsliding and fluvial sediment transfer in mountain catchments are determined by the strength and location of rain storms and earthquakes, and by the sequence in which they occur. To explore this notion, landslides caused by three tropical cyclones and a very large earthquake have been mapped in the Chenyoulan catchment in the Taiwan Central Range, where water and sediment discharges and rock strengths are well known. Prior to the M _W 7·6 Chi‐Chi earthquake in 1999, storm‐driven landslide rates were modest. Landslides occurred primarily low within the landscape in shallow slopes, reworking older colluvial material. The Chi‐Chi earthquake caused wide‐spread landsliding in the steepest bedrock slopes high within the catchment due to topographic focusing of incoming seismic waves. After the earthquake landslide rates remained elevated, landslide patterns closely tracking the distribution of coseismic landslides. These patterns have not been strongly affected by rock strength. Sediment loads of the Chenyoulan River have been limited by supply from hillslopes. Prior to the Chi‐Chi earthquake, the erosion budget was dominated by one exceptionally large flood, with anomalously high sediment concentrations, caused by typhoon Herb in 1996. Sediment concentrations were much higher than normal in intermediate size floods during the first 5 years after the earthquake, giving high sediment yields. In 2005, sediment concentrations had decreased to values prevalent before 1999. The hillslope response to the Chi‐Chi earthquake has been much stronger than the five‐fold increase of fluvial sediment loads and concentrations, but since the earthquake, hillslope sediment sources have become increasingly disconnected from the channel system, with 90 per cent of landslides not reaching into channels. Downslope advection of landslide debris associated with the Chi‐Chi earthquake is driven by the impact of tropical cyclones, but occurs on a time‐scale longer than this study. Copyright © 2008 John Wiley & Sons, Ltd.