Justino et al. (2025) Atmospheric rivers as mediators between climate teleconnections and burned area variability in North America

Identification

• Journal: Communications Earth & Environment
• Year: 2025
• Date: 2025-12-30
• Authors: Flávio Justino, David H. Bromwich, Carlos Gurjão
• DOI: 10.1038/s43247-025-03124-0

Research Groups

• Departamento de Engenharia Agrícola, Universidade Federal de Viçosa, Viçosa, MG, Brazil
• Byrd Polar and Climate Research Center, The Ohio State University, Columbus, OH, USA

Short Summary

This study identifies atmospheric rivers (ARs) as key mediators linking large-scale climate teleconnections (ENSO, PNA, AO) to variations in vegetation activity (NDVI) and burned area (BA) across North America. The findings highlight the central role of ARs in shaping regional fire regimes and improving prospects for seasonal fire prediction.

Objective

• To identify the spatial-temporal patterns and persistence of atmospheric rivers (ARs) and their inter-seasonal, inter-decadal, and annual trends.
• To evaluate how ARs respond to changes in the El Niño-Southern Oscillation (ENSO), Pacific-North American pattern (PNA), and Arctic Oscillation (AO).
• To assess the association of ARs with burned area (BA), Normalized Difference Vegetation Index (NDVI), and 2 meter temperature (T2m) across North America, particularly in 12 regional fire hotspots.

Study Configuration

• Spatial Scale: North America (170°W–50°W, 20°N–70°N), with a focus on 12 regional burned area hotspots.
• Temporal Scale:
  • Primary analysis period for burned area, NDVI, and T2m: 2001–2019.
  • Analysis of AR temporal and spatial characteristics: 1979–2020.
  • Climate indices (ENSO, PNA, AO): 1979–2022.

Methodology and Data

• Models used:
  • Tracking Atmospheric Rivers Globally as Elongated Targets (tARget) algorithm Version 3.0 for AR identification (IVT intensity > 85th percentile or 100 kg m⁻¹ s⁻¹, length > 2000 km, length-to-width ratio > 2).
  • K-means clustering for identifying 12 burned area hotspots.
  • Multiple linear regression residuals for orthogonalization of climate indices.
  • Principal Component Analysis (PCA) for combined climate indices.
  • Spearman rank correlation coefficient for lagged relationships.
  • Distributed Lag Nonlinear Model (DLNM) for cumulative, nonlinear effects of ARs, NDVI, and T2m on burned area.
• Data sources:
  • Burned Area (BA): MODIS Collection 6 NRT fire products (MCD14DL, 1 km) and MCD64A1 Version 6 BA data product (0.5° × 0.5°).
  • Normalized Difference Vegetation Index (NDVI): National Centers for Environmental Information (NCEI) from NOAA-14, NOAA-16, NOAA-18, NOAA-19, NOAA-20, and Suomi-NPP satellites.
  • Atmospheric River (AR) frequency and precipitation: UCLA website (based on Guan & Waliser, 2019).
  • 2 meter temperature (T2m), total column vertically-integrated divergence of geopotential flux (VIGD), and total column vertically-integrated moisture divergence flux (VIMDF): ERA5 reanalysis.
  • Climate Indices (ENSO, AO, PNA): Monthly standardized indices from publicly available sources.

Main Results

• Atmospheric rivers (ARs) act as key mediators linking large-scale climate teleconnections (ENSO, PNA, AO) to North American vegetation activity and burned area variability.
• ENSO generally reduces AR frequency and precipitation north of 40°N while increasing it across eastern North America, with lags ranging from 1 to 6 months.
• The PNA enhances AR frequency over central and eastern North America and Alaska (2–4 month lags), whereas the AO primarily influences southern and western regions (0–2 month lags).
• Vegetation responses (NDVI) are heterogeneous and lag-dependent: ENSO promotes greening in northwestern Canada (4–5 month lags) but browning in Alaska (1–2 month lags) and northeastern Canada (6 month lags).
• PNA reduces NDVI (drier vegetation) across the eastern/central United States and central Canada (5–8 month lags), while promoting greening in Alaska (2–3 months).
• The AO often counters ENSO effects, driving vegetation drying in the southern United States and southwestern Canada (1–2 month lags), and in central Canada and Alaska (5–6 month lags).
• ARs exert strong control over burned area, particularly across northern Canada and Alaska. When AR variability is explicitly included, much of the fire enhancement previously attributed to teleconnection phases is reversed, indicating that AR-teleconnection interactions are pivotal.
• DLNM analysis reveals a dual influence of AR precipitation: an immediate reduction in burned area (typically 0–3 month lags) by approximately 25–40% per standard deviation increase, followed by a delayed enhancement (3–6 month lags) by approximately 15–50% per standard deviation increase, likely due to fuel buildup.

Contributions

• Identifies atmospheric rivers as critical intermediate variables linking major climate teleconnections (ENSO, PNA, AO) to North American vegetation activity and burned area variability.
• Provides a comprehensive, spatially and temporally resolved analysis of AR-teleconnection interactions and their cascading impacts on terrestrial ecosystems.
• Demonstrates the dual, time-dependent influence of ARs on fire activity, showing immediate fire suppression due to moisture and delayed fire enhancement due to fuel accumulation.
• Highlights the importance of considering both individual and combined teleconnection influences, alongside ARs, for understanding complex regional fire regimes and improving seasonal fire prediction.
• Utilizes advanced statistical methods (orthogonalization, PCA, DLNM) to disentangle complex, lagged, and nonlinear relationships in climate-fire interactions.

Funding

• Simon Foundation (for visiting the International Centre for Theoretical Physics in Trieste, Italy)
• CNPq funding: 441744/2024-9 and 303882/2020
• BPCRC Polar Meteorology Development Fund

Citation

@article{Justino2025Atmospheric,
  author = {Justino, Flávio and Bromwich, David H. and Gurjão, Carlos},
  title = {Atmospheric rivers as mediators between climate teleconnections and burned area variability in North America},
  journal = {Communications Earth & Environment},
  year = {2025},
  doi = {10.1038/s43247-025-03124-0},
  url = {https://doi.org/10.1038/s43247-025-03124-0}
}