Close Window

Abstract ID: 95

Revisiting the Avissar Hypothesis: Climate modeling of extreme deforestation in Amazônia

The Avissar Hypothesis states that the relationship between rainfall and deforestation can depart significantly from a linear relationship, depending on the spatial heterogeneity and amount of disaggregation of forested areas. In some cases, rainfall may even increase with deforestation, which is contrary to some received wisdom -- and much popular opinion -- about the climatological effects of Amazonian land cover change. If the Avissar Hypothesis does indeed hold, then identifying a precipitation &ldquotipping point&rdquo becomes an important environmental question, one whose answer may be non-unique, given the dependency of atmospheric dynamics and land-atmosphere feedbacks on the spatial arrangement of forest cover. This presentation considers the Avissar Hypothesis in a policy relevant manner, and under the uncertainty of external forcings on regional climate. To do so, it presents results of simulations of a linked, land-climate system, deploying the RAMS model version 4.4 with several regionally specific modifications. The years 1997-2001 are used to provide a range of exogenous climate conditions, in order to explore the annual and seasonal characteristics of precipitation responses. For each year, annual precipitation across the basin is determined for individual, 20km pixels, and for four settings on Amazonian land-cover conditions, including a best-case scenario, with deforestation effectively stopping at the GLC/PRODES 2004 levels, a worst case scenario, with 100 percent deforestation, and two intermediate cases reflecting assumptions about the effectiveness of protected areas and the impact of road-paving on land cover change. The presentation in effect provides a new test on the Avissar Hypothesis, and tends to support it for domain-averaged values of annual precipitation, independently of external forcings. Rainfall over smaller regions is also used to explore the Avissar hypothesis in order to explore its scalar dimensions. It appears that for some range of basin-scale deforestation, the functional relationship between extent of cleared forest and precipitation is flat, and even increasing. Moreover, the tipping point at which basin-wide precipitation falls sharply appears at fairly high levels of deforestation. The presentation concludes by describing the mechanisms responsible for this result, and by considering it in light of predictions of future Amazonian climate produced by global climate models.

Session:  Feedbacks to Climate - Effects of deforestation on regional and global climate.

Presentation Type:  Oral

Close Window