Biophysical impacts of northern vegetation changes on seasonal warming patterns

Northern Hemisphere vegetation has been greening in recent decades due to longer growing seasons and enhanced photosynthesis, which can alter surface energy exchanges and near-surface air temperature. The study investigates how seasonal changes in leaf area index (LAI) influence the seasonality of NH land air temperature, quantifying both contemporaneous (intra-seasonal) and time-lagged (inter-seasonal) biophysical feedbacks. The core question is whether, and by how much, vegetation greening in spring, summer, and autumn modifies seasonal warming patterns via evapotranspiration, albedo, atmospheric emissivity, and turbulence, and how these effects propagate to subsequent seasons through soil moisture and atmospheric circulation. Understanding these processes is important for interpreting observed seasonal temperature trends, projecting future climate, and informing land-based climate mitigation and adaptation strategies.

Prior work documents widespread NH greening and phenological shifts affecting carbon uptake and climate (e.g., Keenan 2014; Piao 2019; Zhu 2016; Piao 2020). Biophysical vegetation effects include albedo-driven warming in snow-covered high latitudes (Betts 2000; Loranty 2014) and evaporative cooling (Lee 2011; Zeng 2017; Alkama & Cescatti 2016). Studies also report historical decreases in the annual temperature range (Mann & Park 1996; Stine 2009; Xia 2014), and link vegetation changes to soil moisture and drought impacts (Lian 2020; Wolf 2016). Teleconnections can transmit land biophysical anomalies nonlocally (Winckler 2019; Devaraju 2018). Despite this context, the seasonal dependence and carry-over of vegetation biophysical feedbacks on air temperature remain underexplored, motivating this work.

The authors used the IPSL-CM coupled land–atmosphere GCM (LMDZ + ORCHIDEE) with satellite-prescribed LAI to isolate vegetation biophysical effects on air temperature (T_a) for 1982–2014. Four transient experiments with a 30-member initial-condition ensemble were run: (1) SCE forced by annually varying monthly GIMMS LAI3g; and three seasonal control experiments with fixed climatological LAI in MAM, JJA, or SON (LAI_MAM, LAI_JJA, LAI_SON), respectively. Differences (SCE − seasonal control) quantify LAI-induced intra-seasonal and inter-seasonal ΔT_a. All runs used observed SST, sea-ice (AMIP), and atmospheric CO2, enabling indirect vegetation–atmosphere–ocean and CO2 effects to be embedded. LAI3g (8 km, 15-day) was aggregated to monthly and 0.5° grids and distributed across plant functional types from a static Olson land cover map. A second model, CAM6–CLM5 (1.9° × 2.5°), provided 100-year equilibrium experiments contrasting LAI_2010 against seasonal experiments with 1982–1986 LAI in the focused season to test robustness. Observation-based near-surface air temperature from the Princeton Global Forcing (PGF) dataset (1°, 3-hourly) was used to evaluate simulated seasonality. Process attribution employed an analytical decomposition of the surface energy balance to express LAI-induced surface radiative forcing contributions from latent heat flux (ΔE), albedo (α), downward shortwave radiation (S_dn), air emissivity (ε), and aerodynamic resistance (r_a). This framework separated radiative and non-radiative components and allowed estimation of local ΔT_a attributable to ET-driven cooling versus radiative warming (albedo and water vapor) and residual advection/circulation (ΔT_adv). Statistical significance was assessed via ensemble means and trend analyses; spatial significance used stippling at p<0.05. Vegetated NH areas with climatological LAI ≥ 0.1 were analyzed.

• Seasonal intra-seasonal responses (IPSL-CM): MAM warming was non-significant (0.005 ± 0.018 °C decade−1; p>0.1), JJA exhibited significant cooling (−0.044 ± 0.008 °C decade−1; p<0.05), and SON cooling was non-significant (−0.014 ± 0.016 °C decade−1; p>0.1). Over 1982–2014, these correspond to JJA cooling of −0.15 ± 0.03 °C, MAM +0.02 ± 0.06 °C, and SON −0.05 ± 0.05 °C.
• Summer greening offset ~12.5% of the modeled near-surface warming; JJA LAI and ΔT_a anomalies were strongly negatively correlated (r = −0.64, p<0.05), whereas no significant correlation was found in MAM (r = −0.14) or SON (r = 0.07).
• Spatially, strong greening in JJA aligned with cooling hotspots (Europe, Russia, central U.S.), while browning aligned with warming (southern U.S., Alaska, semi-arid Asia). In MAM and SON, despite significant greening over 42.0% and 40.3% of NH lands, respectively, <5% showed significant temperature trends.
• Inter-seasonal (carry-over) responses to previous-season greening (NH means): ΔT_a trends of −0.057 ± 0.029 °C decade−1 (p=0.06; MAM), −0.016 ± 0.013 °C decade−1 (p=0.21; JJA), and −0.029 ± 0.019 °C decade−1 (p=0.13; SON), corresponding to cumulative cooling of 0.19 ± 0.10 °C, 0.05 ± 0.04 °C, and 0.10 ± 0.06 °C (1982–2014). In JJA, this amplified contemporaneous cooling by ~36%; in MAM and SON, inter-seasonal signals exceeded intra-seasonal ones.
• Seasonal amplitude: Greening caused net winter (DJF) warming and growing-season cooling, reducing the annual temperature range (July minus January T_a) by −0.14 ± 0.07 °C decade−1 (p<0.1). Observations and SCE runs showed a leveling-off or non-significant recent increase in the annual range; greening diminished the recent increase in seasonal temperature amplitude.
• Process attribution: Across seasons, ET-related surface radiative forcing provided the largest cooling: −0.23 ± 0.05 (MAM), −0.64 ± 0.09 (JJA), and −0.15 ± 0.03 W m−2 decade−1 (SON; all p<0.05). Radiative warming (combined α, S_dn, ε) increased surface forcing by +0.25 (MAM), +0.18 (JJA), and +0.05 W m−2 decade−1 (SON; all p<0.05). In MAM, radiative warming (notably increased ε from higher water vapor and decreased α over snow) overrode ET cooling, yielding net warming; in JJA, radiative warming offset ~26% of ET cooling (net cooling); in SON, it offset ~35% (cooling became weak/insignificant).
• Mechanistic support: Higher transpiration fraction (T/ET) in JJA (~50% greater than in MAM/SON) strengthened ET sensitivity to LAI and cooling; 33 Earth system models corroborated stronger summer evaporative cooling capacity.
• Inter-seasonal mechanisms: Carry-over involves soil moisture memory (e.g., MAM greening enhanced ET, inducing JJA soil moisture deficits and regional warming via sensible heat increases) and atmospheric circulation/teleconnections (Δz_500 patterns). DJF showed an anomalous high over land across growing-season forcings, producing widespread winter warming.
• Robustness: CAM6–CLM5 equilibrium experiments reproduced strong JJA cooling (−0.18 °C; p<0.05) and much weaker MAM (−0.02 °C) and SON (−0.05 °C) responses.

The results demonstrate that vegetation greening exerts seasonally contrasting biophysical controls on NH near-surface air temperature. In summer, increased LAI enhances evapotranspiration and shifts surface energy partitioning toward latent heat, yielding robust cooling that offsets a meaningful fraction of greenhouse-driven warming. In spring and autumn, evaporative cooling is weaker and increasingly counterbalanced by radiative warming from decreased albedo (especially over snow) and increased atmospheric water vapor emissivity, resulting in small, statistically weak temperature changes. The net effect is a reduction in the seasonal amplitude of temperature, with cooler summers and warmer winters. The study also reveals that vegetation-induced surface energy perturbations propagate to subsequent seasons through soil moisture memory and large-scale atmospheric circulation, creating time-lagged temperature responses, including consistent winter warming. These findings clarify how biophysical feedbacks modulate seasonal warming patterns beyond carbon-cycle effects, with implications for interpreting observed seasonal trends, improving climate model projections of seasonality, and informing land-use strategies (e.g., afforestation/reforestation) to manage seasonal climate risks.

Seasonal Northern Hemisphere greening modifies near-surface air temperature through a trade-off between evaporative cooling and radiative warming. Summer greening produces strong, widespread cooling via enhanced evapotranspiration, while spring and autumn responses are weak because radiative warming offsets cooling. Greening also warms winters and reduces the annual temperature range, thereby weakening temperature seasonality. Inter-seasonal impacts arise through soil moisture memory and atmospheric teleconnections, extending greening effects beyond the forcing season. The study provides a mechanistic, seasonally resolved quantification of vegetation–climate biophysical feedbacks, supported by two independent models and multi-model evidence for stronger summer evaporative control. Future work should reduce uncertainties in inter-seasonal signals by strengthening observational constraints on atmospheric circulation responses, improving model representation of vegetation–snow–moisture interactions, exploring sensitivity to land-cover dynamics and LAI datasets, and conducting coordinated multi-model experiments to refine feedback magnitudes and regional expressions.

The analysis is model-based and subject to uncertainties in simulating vegetation–snow–moisture interactions and atmospheric circulation responses, especially during cooler seasons. Inter-seasonal (carry-over) temperature signals, while consistent in sign, were not always statistically significant at the hemispheric scale. Plant functional types were held static (Olson map), potentially biasing regions with strong land-use changes. The choice of LAI dataset (GIMMS LAI3g) carries trend uncertainties relative to other AVHRR-based products. Attribution of teleconnections relies on model dynamics with limited direct observational constraints, and residual advection/circulation effects are inferred rather than directly observed.