Debris-flow timing and occurrence 1 to 3 years after wildfire provides insights into how peak and triggering rainfall intensity-duration thresholds differ and change with time since fire

Abstract:

Wildfire-induced hydrologic and geomorphic changes on steep slopes increase the likelihood of debris flows. Rainfall intensity-duration (ID) thresholds, which are derived from observations of rainfall and watershed response (i.e. no runoff, flood, debris flow), are often used by government agencies for issuing watches and warnings for post-wildfire floods and debris flows. In many cases, the precise timing of a debris flow within a rainstorm is unknown. As a result, empirical rainfall ID thresholds are sometimes based on the peak rainfall intensities reached during debris-flow producing storms rather than the potentially smaller rainfall intensity that was responsible for triggering the debris flow (i.e. the triggering intensity). Furthermore, rainfall ID thresholds are generally based on data from the first year following a fire and therefore have limited predictive capability after 1 year of recovery. In this study, we use post-fire debris flow monitoring data from six different sites throughout the southwestern U.S., including 3 sites where monitoring has been ongoing for 2-3 years, to (1) quantify differences between peak rainfall intensities during storms and rainfall intensities that trigger debris flows and (2) provide insight into how ID thresholds change as a function of time since burning. At the Pinal Fire (central Arizona, USA) and the Buzzard Fire (southwestern New Mexico, USA), which burned primarily in chaparral and ponderosa pine, respectively, we determined that the 15-minute rainfall ID threshold would be overestimated by roughly 10 mm h^-1 if derived using observations of peak rainfall intensity rather than the triggering intensity. In general, we also observed less extreme debris flow responses at all sites after 1+ year of recovery. For example, we observed only 1 debris flow during the second year of monitoring at the Pinal Fire despite exceeding the year 1 ID threshold eight times. Results from these and the other study sites will help guide agencies for selecting, and adjusting, warning thresholds with time since fire.