Deforestation associated with seasonal rainfall intensity, frequency, and clustering in the Amazon-Cerrado transition zone

Abstract:

Deforestation in the Amazon-Cerrado transition zone is associated with changing rainfall patterns. Previous empirical work has commonly focused on agriculturally-relevant rainfall metrics, and relied on the assumption that deforestation’s effects on rainfall exhibit no preferred spatial direction around impacted areas. Here, we used fixed effects regressions to relate rainfall intensity, frequency, and spatiotemporal clustering to forest cover across multiple spatial scales, directions, and seasons. We present three main findings. First, many of the examined rain metrics were significantly associated with forest cover in multiple seasons. Higher forest cover was associated with higher cumulative seasonal rainfall in all seasons. During the wet and transition seasons, higher forest cover was also associated with less extreme rainfall, fewer dry days, and less spatially coherent rainfall. Additionally, shorter dry spell lengths were associated with higher forest cover during the wet and dry-to-wet transition seasons. During the dry season, however, the sign of the association between forest cover and most rainfall metrics flipped such that higher forest cover was associated with more extreme rainfall, more frequent dry days, and an increase in spatial coherence. Second, most forest-rain associations were stable across several spatial scales and two rainfall datasets. Third, for most combinations of rain metrics and seasons examined, forest-rain associations did not depend on direction, though there were exceptions to this during the dry and dry-to-wet transition seasons. Our results are largely consistent with previous findings, but our approach adds data-driven evidence to the body of knowledge showing that deforestation is associated with changes in a broad suite of rainfall metrics that characterize hydroclimatic dynamics in the Amazon-Cerrado transition zone.