A substantial amount of the tropical forests of South America and Africa is generated through moisture recycling (i.e., forest rainfall self-reliance). Thus, deforestation that reduces evaporation and dampens the water cycle can further increase the risk of water-stress-induced forest loss in downwind areas, particularly during water scarce periods. However, few studies have investigated dry period forest rainfall self-reliance over longer records and consistently compared the rainforest moisture recycling in both continents. Here, we analyze dry-season anomalies of moisture recycling for mean-years and dry-years, in the South American (Amazon) and African (Congo) rainforests over the years 1980-2013. We find that, in the dry seasons, the reliance of forest rainfall on their own moisture supply (ρfor) increases by 7% (from a mean annual value of 26% to 28%) in the Amazon and up to 30% (from 28% to 36%) in the Congo. Dry years further amplify dry season ρfor in both regions by 4-5%. In both the Amazon and Congo, dry season amplification of ρfor is strongest in regions with a high mean annual ρfor. In the Amazon, forest rainfall self-reliance has declined over time. At the country scale, dry season ρfor can differ drastically from mean annual ρfor. In for example Bolivia and Gabon, mean annual ρfor is ~30% while dry season ρfor is ~50%. The dry period amplification of forest rainfall self-reliance further highlights the role of forests for sustaining their own resilience, and for maintaining downwind rainfall at both regional and national scales.

As oceanic moisture evaporates, it leaves a signature on sea surface salinity. Roughly 10% of the moisture that evaporates over the ocean is transported over land, allowing the salinity fields to be a predictor of terrestrial precipitation. This research is among the first in published literature to assess the role of sea surface salinity for improved predictions on low-skill summertime subseasonal timescales for terrestrial precipitation predictions. Neural networks are trained with the CESM2 Large Ensemble using North Atlantic salinity anomalies to quantify predictability of U.S. Midwest summertime heavy rainfall events at 0 to 56-day leads. Using explainable artificial intelligence, salinity anomalies in the Caribbean Sea and Gulf of Mexico are found to provide skill for subseasonal forecasts of opportunity, e.g. confident and correct predictions. Further, a moisture-tracking algorithm applied to reanalysis data demonstrates that the regions of evaporation identified by neural networks directly provide moisture that precipitates in the Midwest.