Surface albedo regulates aerosol direct climate effect

Surface albedo (SA), the ratio of reflected to incoming short-wave radiation at the surface, varies substantially across space and time due to natural processes and human activities. Changes in SA alter radiative fluxes and thus climate, and SA is integral to weather and climate prediction. Atmospheric aerosols scatter and absorb solar radiation, affecting the Earth's energy budget via direct effects (ADRE) and by modifying clouds. ADRE can be cooling (negative) or warming (positive, termed AWE) depending on aerosol optical properties—AOD, single scattering albedo (SSA), asymmetry factor (ASY)—and SA. A critical SA marks where ADRE transitions from negative to positive. Prior work showed AWE depends on AOD, SSA, backscattering, and SA. Given recent SA changes from land use, snow/ice variability, and other drivers, the study asks: how do SA changes affect ADRE and AWE? The purpose is to map global spatio-temporal ADRE/AWE, quantify SA’s contribution to ADRE changes, and evaluate SA’s role in enabling AWE.

Foundational studies (Atwater; Chylek & Coakley; Haywood & Shine; Liao & Seinfeld) established how aerosol absorption/scattering and surface reflectance control planetary albedo and radiative forcing, with the possibility of warming over bright surfaces. AERONET-based analyses advanced Critical SA estimation, generally finding values above ~0.3 and strong dependence on SSA and aerosol type. Satellite studies indicated radiative flux sensitivity to SSA near Critical SA and a slight increase of Critical SA with increasing AOD. Recent analyses decomposed ADRE trends into contributions from AOD, SA, and insolation, highlighting the importance of SA changes, especially from sea ice loss in the Arctic.

Data: CERES SYNIdeg Level 3 (monthly, 1°×1°) provides TOA and surface short-wave fluxes and surface albedo under clear-sky and pristine-sky conditions for March 2001–February 2020 (used for ADRE/AWE distributions and monthly variations). AERONET Version 3 Level 1.5 (15-minute resolution at 1403 sites) provides aerosol optical properties (AOD, SSA, phase function) and derived radiative parameters; used to estimate Critical SA vs AOD and SSA. MACv2 aerosol climatology (monthly, 1°×1°, 550 nm) provides climatological AOD, SSA, and ASY to parameterize aerosol reflectivity and transmissivity. Computation of ADRE/AWE: ADRE at TOA is defined as the difference between upward flux in aerosol-free sky and with aerosols (Ft − Fa), using CERES clear-sky fluxes. Positive ADRE denotes warming (AWE). Critical SA estimation: Using AERONET, samples were divided into quartiles of AOD and SSA (equal sample counts). Within each AOD–SSA bin, ADRE is regressed linearly against SA; the Critical SA is where the regression crosses ADRE = 0. Significance required P < 0.01. Sensitivity to solar zenith angle was examined; comparisons to SBDART-based estimates showed consistency. Geographical Critical SA mapping: For each CERES grid, linear fits of ADRE vs SA were performed, retaining only grids with P < 0.05 and observed AWE; results emphasized 60–90°N due to data availability and significance. Quantifying SA-driven ADRE changes (ΔADRE due to ΔSA, denoted AADRE): Using a simplified two-stream framework, TOA albedo Rtoa is expressed via aerosol layer reflectivity (raer) and transmissivity (taer) and SA. The relation ADRE = F*(SA − Rtoa) connects ADRE and SA. Partial derivative ∂R/∂SA is computed using climatological aerosol properties (MACv2 SSA, ASY; CERES AOD) and multi-year mean SA and downward TOA flux Ftoa. The Theil–Sen median estimator is used for trends (ΔSA per decade), and Mann–Kendall test assesses significance (P < 0.05). Long-term annual and monthly analyses (2000/2001–2020) are performed; global averages of ΔSA and ΔADRE are computed only over grids with significant SA trends. Uncertainty: Gaussian error propagation quantifies σADRE from uncertainties in SA, raer, taer, and Ftoa. Monthly grid flux measurement uncertainties (O(1–10) W/m²) reduce to O(0.001–0.01) W/m² in multi-year global averages; propagated global σADRE ≈ 0.17 W/m² (0.20 NH, 0.15 SH).

• Global ADRE pattern: From March 2000 to February 2020, mean clear-sky ADRE is predominantly negative globally (cooling). Strongest cooling (>15 W/m²) occurs over high-AOD regions such as Eastern China and Northern India, with adjacent oceans also showing strong cooling due to transported aerosols.
• AWE distribution and seasonality: AWE is concentrated in mid-to-high latitudes (polar regions with persistent snow/ice) and regions with absorbing aerosols (Arctic, Northern Africa deserts, Siberia, Northeast Asia, Central North America). Polar regions, especially Greenland and parts of Antarctica (Weddell Sea, SW Antarctica), have the highest AWE occurrence. AERONET’s higher temporal resolution reveals more widespread land AWE than monthly CERES, e.g., frequent AWE in the SW U.S. and Chile (linked to declining SSA trends). Monthly, Arctic AWE peaks >5 W/m² in May–June, near zero during polar night; Antarctic AWE is near zero in June–August.
• SA effect on ADRE and Critical SA: ADRE increases with SA; higher SA weakens aerosol-induced cooling and increases AWE likelihood. Critical SA ranges from 0.18 to 0.96 across AOD–SSA bins and increases with both AOD and SSA. Thinner and/or more absorbing aerosols (lower AOD, lower SSA) yield smaller Critical SA and more readily cause AWE. For very thick, highly scattering aerosols (AOD > 0.26, SSA > 0.97), AWE does not occur; theoretical Critical SA can exceed 1 (e.g., 1.09, 2.42). Critical SA is more sensitive to SSA than AOD. Arctic mapping shows Critical SA > 0.4 around the Arctic Ocean and southern Greenland.
• SA trends and ADRE response (2000–2020): Global annual mean SA change across all regions is +5.58×10⁻³/decade (P < 0.05), implying AADRE of −0.134 W/m²/decade. In regions with significant SA trends, SA decreased on average −0.012/decade, producing AADRE of −0.2 ± 0.17 W/m²/decade. The Arctic shows the largest SA decline (−0.039/decade) and corresponding AADRE ≈ −0.46 W/m²/decade (more than twice the global average). Regions with significant SA decreases include India, Central/Northern Africa, and Southeast Asia; increases occur in Eastern China and Eastern South America.
• Sensitivity and seasonality: In the Northern Hemisphere, every 0.1 change in SA leads to >1 W/m² change in ADRE on average, larger than global and Southern Hemisphere sensitivities. Monthly, the most pronounced AADRE declines occur in June–July: −0.37 W/m²/decade (global) and −1.31 W/m²/decade (NH), coinciding with peak AOD. In the SH, peak average ΔSA (~+0.003/decade) and higher AOD are associated with a minimum AADRE near −0.16 W/m²/decade.
• Implication: Declining SA enhances aerosol short-wave cooling (more negative ADRE) and reduces AWE, partially counteracting the increased surface absorption from SA reduction, especially in the Arctic and during NH summer.

The study demonstrates that aerosols generally cool the Earth-atmosphere system, but over bright surfaces with SA above a Critical SA, aerosols can produce a net warming (AWE) due to enhanced absorption of increased surface-reflected radiation. By quantifying Critical SA across aerosol conditions and mapping its distribution, the work clarifies where and when AWE arises—primarily over high-latitude, high-albedo regions and in areas dominated by absorbing aerosols. The documented widespread decline in SA since 2000 intensifies aerosol-induced cooling (more negative ADRE) and diminishes AWE, particularly notable in the Arctic and NH summer, thereby partially offsetting the radiative warming associated with reduced SA (e.g., ice-albedo feedback). These findings highlight SA as a key regulator of aerosol radiative effects and underscore the necessity of accounting for SA variability in climate assessments. The enhanced aerosol cooling may modestly slow snow and ice melt, but the extent of such moderation is limited and requires further quantification.

This work integrates satellite (CERES), ground-based (AERONET), and reanalysis (MACv2) data to: (1) map global spatio-temporal patterns of clear-sky ADRE and AWE; (2) quantify how SA modulates ADRE and identify Critical SA values (0.18–0.96) as functions of AOD and SSA; and (3) attribute recent ADRE changes to observed SA trends, revealing that declining SA enhances aerosol cooling globally, especially in the Arctic and NH summer. Thick, highly scattering aerosols rarely produce AWE even over bright surfaces, while thinner or more absorbing aerosols do so more readily. The results show that SA changes significantly regulate aerosol radiative impacts and should be explicitly represented in climate models and mitigation strategies. Future research should better quantify how aerosol-enhanced cooling modulates cryospheric melt rates, improve long-term SSA datasets, and assess all-sky effects and feedbacks under evolving aerosol emissions and surface conditions.

• Limited availability of long-term, high-quality satellite/reanalysis SSA: MACv2 provides climatological SSA; temporal SSA variability was not fully incorporated in Critical SA mapping with CERES.
• Use of AERONET Version 3 Level 1.5 data: greater SSA uncertainties at low AOD compared to Level 2, which is AOD-limited (>0.4).
• Methodological simplifications: single-layer aerosol representation; assumption that aerosol-free atmosphere is mostly transparent in shortwave; reliance on climatological aerosol properties (AOD, SSA, ASY) for derivative estimates of ∂R/∂SA.
• Spatial constraints: significant Critical SA mapping was primarily feasible over high-latitude NH where AWE and statistical significance criteria were met.
• Differences in ADRE trend magnitudes relative to other studies may arise from varying datasets, spatial-temporal coverage, and methods.