Stochastic framework reveals the controls of forest treatment – peakflow causal relations in rain environment

Abstract

Decades of literature on forest treatment – peakflow relations have generated considerable disagreements and turned the topic into one regarded as enigmatic. Factors affecting peakflows are multiple and chancy and, hence, can only be investigated via a probabilistic approach. We analyze peakflow data using the peakflow frequency distribution framework in two pairs of control-treatment watersheds in the rain environment of an experimental forest in North Carolina. We demonstrate how a range of forest treatments can change the magnitude and frequency of all peakflows on record and how such effects can increase with increasing event size as a consequence of changes to the peakflow frequency distribution. Changes to the distribution's mean (−0.2 to +47.4 %) and variance (−30.9 to +162.1 %) resulted in a range of effects from no significant impacts on peakflows to making larger peakflows becoming even larger (up to 105 % increase) and more frequent (up to 18 times more frequent). Changes to peakflows are attributed to treatment-induced suppression of evapotranspiration and changes to non-vegetative factors, which can alter the soil storage capacity, moisture available for runoff, and the efficiency of runoff arriving to the outlet. The dominant topographical aspect of the watershed, seasonal differences in storm types, extent to which the storm events are in-phase or out-of-phase with high antecedent soil moisture, and lagged runoff responses with watershed memory emerged as key indicators of the sensitivity of peakflows to forest treatment. The application of the stochastic framework in forest hydrology can help fully understand forest treatment – peakflows relations.