In 2005, the FAO and CIFOR published a report, Forests and floods: Drowning in fiction or thriving on facts? The report asserted that scientific evidence did not support the conventional wisdom linking deforestation to large-scale flooding. The report re-ignited a long-simmering controversy, and continues to provoke emotional responses from both sides of the debate.

Now a new study published in the journal Global Change Biology offers empirical evidence that flood risk is indeed correlated with loss of natural forest cover. Corey Bradshaw and his colleagues assembled national-level data from 56 developing countries on changes in forest cover and flood events over the period 1990 to 2000. Their models — which control for country size, differences in rainfall, steepness of terrain, and area of degraded landscape — succeed in explaining 65% of the variation in flood frequency across countries, with 14% explained by forest-related variables.

Were FAO and CIFOR wrong?

In an essay in Nature, William Laurance points out that Bradshaw and his co-authors have not included in their analysis extreme flooding events caused by major storms such as cyclones and typhoons. His observation provides at least a partial reconciliation of the two apparently contradictory reports.

The FAO & CIFOR report stated that “contrary to popular belief, forests have only a limited influence on major downstream flooding, especially large-scale events” [emphasis added]. A key policy message of the report was that governments should not use prevention of catastrophic floods to justify restricting the activities of forest-dependent upland farmers.

The Bradshaw study not only excludes some large-scale events, but does not include data from China due to the number of extreme outliers it produced. The results suggest that forests play an important role in moderating the incidence and severity of the type of floods included in the study. Their analysis suggests that a decrease in natural forest cover of 10% could lead to an increase of 4 to 28% in flood frequency, with the associated economic loss and human misery.

The debate isn’t over yet: Bruijnzeel and his colleagues have already challenged the Bradshaw study, asserting that it was a mistake not to include a variable for population size or density. According to their alternative interpretation (which has not yet appeared in a peer reviewed journal), it is what happens to landuse following forest cover loss that explains changes in flooding.

Policymakers want to know: does deforestation cause flooding? These debates demonstrate that simple ’yes’ or ’no’ answers are both wrong. Blaming upland farmers for massive downstream flooding can cause unnecessary suffering through inappropriate, repressive policies. At the same time, failing to protect forest cover is associated with significant costs to human welfare through increased incidence of floods caused by heavy rainfall. Our greatest challenge is to convey a nuanced message on this matter across the science-policy interface, even as the scientific debate continues.