Combined large-scale tropical and subtropical forcing on the severe 2019–2022 drought in South America

The study investigates the unprecedented 2019–2022 drought over central-east South America (CESA), set against a backdrop of a strengthening global water cycle under climate change. Rising temperatures increase lower-tropospheric water vapor, altering the evaporation–precipitation balance that controls soil moisture, groundwater recharge, and runoff. Feedbacks among soil moisture, atmospheric water budget, and temperature can amplify heat and drought via constraints on latent heat flux and enhanced sensible heat, and can also modulate convective precipitation. South America has warmed by ~0.1–0.4 °C per decade since 1981–2020, with complex and regionally varying precipitation changes. The Amazon and La Plata basins are key to regional hydroclimate via moisture recycling and transport, modulated by large-scale variability modes (ENSO, MJO, AMM/AZM) and systems like the ITCZ, South Atlantic Convergence Zone, South American Low-Level Jet (SALLJ), and mid-latitude synoptic features. Since mid-2018, severe drought spread across Pantanal, Paraguay, Bolivia, and northern Argentina, causing agricultural, hydropower, and navigation impacts and catastrophic wildfires in 2020. The research aims to provide a detailed spatiotemporal characterization of the 2019–2022 drought in CESA, assess its historical exceptionality in soil moisture anomalies, and elucidate the atmospheric mechanisms from daily to multiyear timescales, focusing on internal variability and predictability of climate extremes.

The paper situates the 2019–2022 CESA drought within literature on climate change–driven intensification of the water cycle and land–atmosphere feedbacks. Prior work shows: (1) South America’s warming trends and increased warm extremes (IPCC AR6; de Barros Soares et al.; Regoto et al.); (2) complex precipitation projections, with a dipole of drying in the Amazon and wetter conditions over La Plata Basin; (3) Amazon and La Plata basins’ critical roles in moisture recycling and cross-basin transport via trade winds and SALLJ; (4) modulation by ENSO, MJO, and Atlantic modes (AMM/AZM), influencing ITCZ, South Atlantic Convergence Zone, SALLJ, and transient systems; (5) strong soil moisture–temperature and soil moisture–precipitation feedbacks in CESA that can amplify heatwaves and droughts; (6) links between La Niña and SA precipitation anomalies via Walker circulation and Pacific–South American teleconnections; and (7) potential amplification of downstream drying through Amazon deforestation reducing moisture transport. The study builds on these to jointly assess temperature, precipitation, and soil moisture trends and to attribute the 2019–2022 drought to combined tropical and subtropical dynamical forcings.

Data and domains: Daily meteorological fields (precipitation, temperature, humidity, winds, geopotential, vertical velocity) from ERA5 (1959–2022) and 0–7 cm soil moisture from ERA5-Land (1951–2022). The CESA analysis region is defined by a box spanning the Pantanal, La Plata Basin sectors, and adjacent areas. Trend significance used the non-parametric Mann–Kendall test. Some time series were filtered using a 10-year low-pass Lanczos filter. Indices and diagnostics: (1) Vertically Integrated Water Vapor Transport (IVT) computed from vertically integrated zonal and meridional moisture fluxes; (2) Vertically Integrated Moisture Convergence (VIMC) obtained as a line integral of IVT across CESA borders (west, north, east, south) to quantify the net moisture convergence/divergence over the region; (3) Partition of precipitation into components due to moisture convergence versus local moisture recycling from evaporation; (4) R-index, adapted from event-ranking metrics, to rank daily drought severity by multiplying the area fraction within CESA with soil moisture anomalies below −2 standard deviations by the mean anomaly over those grid points; soil moisture anomalies were standardized relative to 1981–2010 seasonal climatology and smoothed with a 31-day running mean; (5) Rossby Wave Source (RWS) computed from the barotropic vorticity equation using divergent winds and absolute vorticity to identify source regions and wave forcing; (6) Bivariate Gaussian distributions used to characterize joint changes in mean annual temperature and precipitation between 1959–1989 and 1990–2022. Composites and regressions: Spatial anomaly composites were generated for drought years (2019–2022) and for days corresponding to the nine most extreme R-index peaks. Fields included soil moisture, 500 hPa geopotential height, 850 hPa temperature and divergent winds, vertically integrated moisture divergence, 200 hPa velocity potential and divergent winds, and 200 hPa meridional wind, along with RWS and jet stream diagnostics. Significance testing used Student’s two-tailed t test at 5%. Climate mode linkages: Monthly and annual correlations (with and without 10-year filtering) were computed between northern-border IVT and ENSO indices (SOI, ONI), PDO, and AZM to assess relationships and seasonality. Lead–lag behavior was examined qualitatively. Additional diagnostics assessed changes in subtropical highs and jet streams relative to climatology.

- Long-term drying: CESA shows the steepest soil moisture decline in South America since the 1990s (trend ≈ −0.010 m³/m³ per decade, significant), alongside a warming shift and stronger negative temperature–precipitation correlation in 1990–2022 (R = −0.63) versus 1959–1989 (R = −0.43). - Unprecedented event: The 2019–2022 drought is unmatched in the historical record for intensity and duration. April 26, 2020 recorded the lowest R-index since 1951; over 30% of CESA had soil moisture anomalies below −2σ that day, and ~100,000 km² experienced record-breaking dryness at daily scales. For seasonal/annual scales (300–365 days), >700,000 km² (~20% of CESA) saw record-low soil moisture. - Moisture convergence control: Annual precipitation anomalies closely track VIMC (Spearman ρ = 0.92), while the moisture recycling contribution shows lower correlation (ρ = 0.62) and a post-1980s decline. Years 2019–2020 had the two lowest VIMC values; 2021–2022 had the two lowest recycling contributions. - Inflow/outflow patterns: During 2019–2022, moisture inflow across CESA’s north and west borders was below normal. Correlations between annual northern/western IVT and CESA VIMC are positive and significant (north R = 0.70; west R = 0.55), highlighting Amazon-sourced inflow as key. Eastern/southern outflows are not dominant for VIMC. Spatially, IVT anomalies exhibited a weakened NW–SE transport from Amazon to CESA and enhanced divergence over CESA at 850 hPa, yielding positive E−P (evaporation minus precipitation) anomalies over CESA and negative E−P to the northeast. - ENSO/PDO/AZM linkages: Northern-border IVT correlates significantly with ENSO indices outside summer and more strongly on decadal timescales (10-yr filtered annual correlations: SOI r = −0.73; ONI r = 0.67; AZM r = −0.19 to −0.36 depending on filtering; non-filtered annual: SOI −0.56; ONI 0.56; AZM −0.36 filtered). Dry multi-year periods align with positive SOI/negative ONI (La Niña). PDO modulates ENSO teleconnections, with negative PDO concurrent with La Niña enhancing SA rainfall anomalies. - Tropical circulation anomalies: Cold SST anomalies in central/eastern tropical Pacific coincided with an eastward-shifted Walker cell, with subsidence over NW South America and ascent over the equatorial Atlantic near NE SA. An amplified Hadley cell placed its descending branch over CESA, producing strong subsidence, moisture divergence, clear-sky conditions, and reduced SALLJ moisture transport from Amazon. - Synoptic reinforcement via Rossby waves: Composite analyses of the nine most extreme flash-drought peaks show positive 500 hPa height anomalies, warm 850 hPa temperatures, strong low-level divergence, and a Rossby wave train (wavenumber ~3) spanning west-central South Pacific to South Atlantic, linked to sources east/south of Australia and a poleward-shifted 200 hPa jet. Subtropical highs over South Atlantic and South Pacific expanded zonally, with increased ridging and a poleward jet shift reducing cyclones/frontal passages into southern CESA. - Impacts and spatial evolution: Flash droughts in 2020 repeatedly hit Pantanal, Bolivia, Paraguay, and northern Argentina. Deforestation may have intensified La Niña-related precipitation deficits by reducing Amazonian moisture, increasing downstream drying susceptibility (not quantified here).

The findings attribute the exceptional 2019–2022 CESA drought primarily to internal climate variability manifested through coupled tropical and subtropical dynamics. La Niña-driven SST anomalies reorganized the Walker circulation, reducing moisture transport via a weakened SALLJ and establishing an amplified Hadley cell with subsidence and moisture divergence over CESA, thereby suppressing precipitation and enhancing evaporation. Rossby wave trains emanating from the west-central South Pacific/east and south of Australia reinforced subsidence during flash drought peaks, fostering quasi-stationary anticyclonic conditions over CESA. Decadal modulation by PDO further strengthened ENSO teleconnections affecting South American hydroclimate, while AZM phases likely contributed to weakening the SALLJ and reducing moisture inflow. These mechanisms explain the observed alignment between VIMC and precipitation anomalies and the record-setting soil moisture deficits. The study underscores that accurate simulation of daily-to-multiyear dynamical processes—ENSO, MJO-related intraseasonal variability, Rossby wave propagation, and jet stream shifts—is essential for improving predictability of drought extremes in CESA. Land–atmosphere feedbacks in this coupling hotspot likely amplified heat and dryness once atmospheric circulation initiated the deficits, compounding impacts such as wildfires in Pantanal.

This work provides a comprehensive spatiotemporal characterization of the 2019–2022 drought in CESA and demonstrates that a coupled tropical–subtropical forcing—La Niña–induced shifts in Walker/Hadley circulations and synoptic Rossby wave trains—produced unprecedented soil desiccation. The study quantifies the dominance of moisture convergence (VIMC) in controlling precipitation anomalies, documents weakened Amazon-to-CESA moisture inflow, and introduces a soil-moisture-based R-index to diagnose and rank flash drought intensities and spatial extents. The results offer a conceptual framework for diagnosing past events and improving prediction of future droughts, stressing the need to better represent internal variability and dynamical processes across timescales in climate models. Future work should (i) quantify the roles of deforestation and AZM/SAM linkages; (ii) assess land–atmosphere feedback amplification over broader CESA domains; (iii) refine lead–lag relationships among climate modes; and (iv) reduce model biases in tropical Pacific dynamics and teleconnections to enhance regional hydrological projections.

- Reliance on reanalysis (ERA5/ERA5-Land) due to sparse in situ soil moisture observations in CESA introduces dataset-specific uncertainties in trends and anomalies. - ENSO lead–lag relationships with CESA moisture transport show weak temporal consistency across the full record, warranting further investigation. - The influence of Atlantic modes (AZM/AMM) and Southern Hemisphere modes (SAM) on CESA precipitation and jet variability was not fully quantified; correlations with SAM were limited. - Potential amplification by Amazon deforestation is discussed but not explicitly modeled or attributed in this study. - Climate model–observation discrepancies in representing Pacific dynamics (e.g., Walker circulation responses) limit confidence in projections; attribution to anthropogenic forcing beyond contextual warming is not performed. - The R-index and composites depend on a 31-day smoothing choice and regional box definition, which may affect detection of rapid-onset events.