Fig. 4: The impacts of forest cover change on recycled precipitation.

a Correlation between weighted forest cover and terrestrial recycled precipitation in the southern Amazon basin. Recycled precipitation is derived from the moisture-trakcing model driven by water balance-based evapotranspiration. Regression line (black line) is based on different spatial points, with each point representing local changes in recycled precipitation and weighted forest cover in the southern Amazon basin for the common period 1982–2016. Each point represents a 1° × 1° gridbox within the southern basin. The blue, red and green lines mark the changes in weighted forest cover in the past 35 years, SSP2-4.5 (primf) and SSP2-4.5 (primf + secdf) scenarios, respectively, and the corresponding reductions in terrestrial recycled precipitation. “Primf” represents primary forested land, while “secdf” represents secondary forested land including forest regrowth and climate mitigation strategies such as afforestation and reforestation (“Methods”). The shaded areas denote the 95% confidence intervals of changes in the southern basin. For illustration purposes, the horizontal and vertical zero lines are shown as grey dashed lines. b The same as (a), but instead, the level of changes in precipitation caused by altered land moisture recycling is derived from the difference between projections of the moisture-tracking model when driven directly by evapotranspiration estimates with and without deforestation. The evapotranspiration post-deforestation was based on a forest cover-evapotranspiration scaling approach. As (b) is direct process model output, we do not present this as a statistical finding (e.g. with p value), but we do fit a linear regression line (black line) to aid comparison with (a). Source data are provided with this paper.