A 2009 study revealed serious flaws in a standard technique used by hydrological researchers to understand how changes in watershed land use affect stream flow behaviors, such as peak flows. The study caused academics and government agencies alike to rethink decades of watershed research and prompted Kuraś et al. to reinvestigate a number of long-standing assumptions in watershed research using a complex and well-validated computer model that accounts for a range of internal watershed dynamics and hydrologic processes. For the test site at 241 Creek in British Columbia, Canada, the authors found not only that deforestation increased the severity of floods but also that it had a scaling influence on both the magnitudes and frequencies of the floods. The model showed that the larger the flood, the more its magnitude was amplified by deforestation, with 10-to 100-year-return-period floods increasing in size by 9%–25%. Following a simulated removal of half of the watershed's trees, the authors found that 10-year-return-period floods occurred twice as often, while 100-year-return-period events became 5–6.7 times more frequent. This proportional relationship between the increase in flood magnitudes and frequencies following deforestation and the size of the flood runs counter to the prevailing wisdom in hydrological science.