High-Resolution Moisture Recycling Networks for Atmospheric Water Management in South America

Effective water resource governance requires a fundamental transition from isolated catchment-based approaches to a holistic perspective that integrates atmospheric moisture transport. Large-scale vegetation transitions in upwind regions fundamentally modify evapotranspiration fluxes, triggering "green water" feedbacks that influence hydroclimatological extremes in remote downwind territories. However, integrating these teleconnections into management strategies has been hindered by the limitations of data spatiotemporal resolution. Previous model versions (e.g., WAM-2layers v2) were typically constrained to coarse-resolution (1.5°) ERA-Interim input data, which limited the ability to resolve fine-scale moisture pathways and often conflated localized recycling with regional transport due to spatial averaging. In this study, we propose a framework for "Atmospheric Water Management" utilizing the WAM-2layers v3 model, driven directly by high-resolution (0.25°) ERA5 reanalysis data over the last 30 years. This represents a 6-fold increase in spatial resolution, allowing us to capture the full heterogeneity of anthropogenic landscapes. By utilizing this refined grid, we can distinguish specific moisture transport corridors and recycling loops that were previously obscured by the coarse discretization of earlier studies.

Building on this refined climatology, we apply complex network theory to construct a connection-graph of South America's "flying rivers." This approach enables us to pinpoint critical "moisture hubs"—specific geographic nodes where land-surface integrity exerts the strongest control over water security in downwind agricultural sinks, such as the La Plata Basin. We hypothesize that these hubs act as critical atmospheric infrastructure; their degradation via deforestation weakens the continental network's resilience, amplifying drought intensity downstream. Consequently, we argue that these identified hubs should be prioritized as conservation targets within national land-use planning. By quantifying these source-sink dependencies, this work aims to establish process-based thresholds for delimiting "precipitationsheds," supporting spatial planning where land conservation decisions are recognized as essential upstream water management strategies.