A new study from researchers at the University of British Columbia suggests that the flood risk created by forest harvesting depends far less on the percentage of trees removed than on the landscape where logging occurs. Watersheds that are flatter, drier and warmer may experience sharper increases in flood peaks after forests are cut, while mountainous basins with substantial elevation variation can naturally spread runoff over time and reduce the intensity of flooding.
Published in Advances in Water Resources, the research examined data from nine paired watersheds in Oregon’s coastal mountain forests. Each harvested watershed was compared with a similar nearby forest that remained intact, allowing researchers to isolate the effects of logging from broader influences such as regional weather. The analysis showed that forest removal interacts with terrain, climate, road density and snowmelt timing to determine how strongly streamflows respond during storms.
The findings challenge a widely used method of estimating post-harvest flood risk: calculating the proportion of a watershed that has been logged. According to the researchers, that measure can be misleading because identical amounts of harvesting may produce very different hydrological effects in different locations. A clear-cut area in a steep, highly variable mountain basin may have a smaller effect on peakflow than a similarly sized harvest in a flatter watershed where snow melts and rainfall reaches streams more uniformly.
Elevation is one of the most important factors. In mountainous terrain, snow can accumulate and melt at different times across slopes and elevations. This staggered melt acts like a natural buffer, distributing water over a longer period rather than sending a single large pulse into the stream network. In flatter landscapes, however, snowpacks may experience similar temperatures and melt more synchronously. If forests have also been removed, more water can reach the ground and move rapidly into channels, increasing the likelihood of higher flood peaks.
Forest canopies influence this process in several ways. Trees intercept rainfall, slow the movement of water across the ground and take up moisture through their roots. Their presence can also alter snow accumulation by shading the ground and catching snowfall on branches. After harvesting, snow may accumulate differently and melt more quickly, while reduced vegetation and disturbed soils can increase the speed at which water travels toward streams. These changes become especially important during rain-on-snow events, when warm rainfall rapidly melts an existing snowpack.
The study identifies fog as another overlooked part of the equation. In humid coastal forests, tree canopies capture tiny droplets of water suspended in fog, a process known as fog interception. The captured moisture can drip to the forest floor and influence local water availability, while the canopy itself affects how snow is distributed and when it melts. Removing trees disrupts these interactions, potentially changing the amount and timing of runoff during storms. The researchers say this may help explain why flood responses after harvesting can be highly variable, even among watersheds with similar forest cover and climate.
The authors used a stochastic hydrology framework to examine how multiple sources of natural variation shape the relationship between harvesting and peakflow. Rather than treating floods as the result of a single fixed cause, this approach considers the probability and timing of different hydrological processes, including rainfall, snowmelt, evaporation and watershed storage. It helps distinguish the effects of forest disturbance from the background variability that naturally occurs between storms and across years.
“This is especially relevant in the province because the region’s coastal mountain landscapes are complex,” said Dr. Younes Alila, senior author and a professor at UBC’s faculty of forestry and environmental stewardship. The Pacific Northwest shares many of the conditions examined in the study, including steep coastal terrain, transient snowpacks and frequent rain-on-snow events. Those processes were implicated in the severe flooding that affected Sumas Prairie and southwestern British Columbia during the 2021 atmospheric river events, although the new study does not claim that logging alone caused that disaster.
The researchers say the results support a more targeted approach to watershed management in British Columbia and elsewhere. Instead of evaluating harvest blocks in isolation, planners could assess how terrain, elevation range, climate, roads, snow conditions and forest structure combine across an entire drainage basin. Such landscape-level analysis could help identify watersheds where small, frequent floods are likely to become more common, as well as locations where the largest flood events may be amplified after harvesting.
Nick Rong, the study’s lead author and a UBC master’s graduate, said the framework is intended to improve predictions of where forest harvesting is most likely to alter stream behaviour. Co-author Kyle Bishop, a PhD candidate in forest hydrology, added that every watershed responds differently to disturbance. By recognizing the natural features that regulate water movement, forest managers may be able to prioritize vulnerable basins, adjust harvesting patterns and reduce the risk of intensifying floods in regions already facing more extreme weather.
Subject of Research: Not applicable
Article Title: Stochastic Hydrology Reveals the Controls of Forest Harvesting – Peakflow Causal Relations in Snow-Transient Environment
Web References: https://www.sciencedirect.com/science/article/pii/S0309170826002277
References: Advances in Water Resources. DOI: 10.1016/j.advwatres.2026.105432
Keywords: Hydrology, watersheds, forest harvesting, flooding, peakflow, snowmelt, rain-on-snow events, fog interception, watershed management, British Columbia, Pacific Northwest
