Deforestation intensifies daily temperature variability in the northern extratropics

Forests have experienced significant changes in recent decades due to human activities and natural disturbances. Deforestation profoundly impacts the carbon cycle and local/regional climate through biogeophysical processes. While deforestation increases surface albedo leading to cooling, it also reduces evapotranspiration resulting in warming due to decreased aerodynamic roughness, rooting depth, leaf area, and canopy conductance. The dominant effect (cooling vs. warming) varies geographically; albedo-driven cooling prevails in boreal regions while evapotranspiration-driven warming dominates the tropics. The effect in mid-latitudes is complex and uncertain due to variations in model agreement. Deforestation also creates a diurnal asymmetry, causing daytime warming and nighttime cooling, thus amplifying the diurnal temperature range. Its influence on temperature extremes, particularly hot extremes, remains uncertain with some studies reporting aggravation while others report alleviation. Although the effects of deforestation on mean, diurnal cycle, and extreme temperatures are widely studied, its influence on temperature variability remains largely unexplored. Daily temperature variability is crucial as it significantly impacts human and natural systems. Increased variability is linked to higher mortality risks from chronic and epidemic diseases, it affects ecosystem functions (crop yields, coral bleaching), and even threatens macroeconomic growth. Therefore, understanding deforestation's effect on temperature variability is vital. Historical observations show a decrease in daily temperature variability in northern mid- and high-latitude continents during winter, spring, and autumn, and increases in some areas during summer. These changes are attributed to various factors including greenhouse gas emissions, aerosols, urbanization, and internal climate variability, but the role of forest changes has been largely neglected. This study addresses this gap by investigating the biogeophysical effects of deforestation on daily temperature variability using both idealized scenarios and historical data, to improve our understanding of human contributions to changes in daily temperature variability and assess the potential impact of afforestation efforts.

Existing research extensively documents deforestation's impact on average and extreme temperatures, revealing a complex interplay between albedo and evapotranspiration effects. Studies have shown that in boreal regions, the albedo effect (increased reflection of sunlight) dominates leading to cooling, while in the tropics, reduced evapotranspiration (loss of water from plants) results in warming. However, mid-latitude responses are less consistent across models. The diurnal asymmetry of temperature responses, with daytime warming and nighttime cooling, has also been reported. The effect on temperature extremes is similarly uncertain, with studies showing both increased hot extremes in some regions, and alleviation in others. A significant gap in the literature is the lack of comprehensive research on the impact of deforestation on daily temperature variability. While studies have examined the influence of other factors on this variability (e.g., greenhouse gas emissions, aerosols, urbanization), the role of deforestation has received less attention. This research aims to address this gap by focusing specifically on the relationship between deforestation and daily temperature variability, bridging this knowledge gap.

This study employs a combination of idealized deforestation scenarios and analysis of historical deforestation patterns to investigate the biogeophysical effects on daily temperature variability. Idealized deforestation was simulated using multiple Earth System Models (ESMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6) and the Land Use Model Intercomparison Project (LUMIP). The idealized scenario involved removing 20 million km² of forest globally over 50 years, maintaining constant tree cover thereafter. The biogeophysical impact was isolated by comparing the results of this simulation ('deforest-globe') with a preindustrial control simulation ('piControl'). Five ESMs were used. Daily temperature variability was quantified using two indices: day-to-day temperature variation (DTDT) – the mean absolute difference between consecutive days' temperatures – and standard deviation of daily mean 2-meter temperature (SDT). Model performance was evaluated by comparing simulated DTDT with two independent reanalysis datasets: ERA5 and NCEP-DOE AMIP-II. The models' ability to reproduce the observed biogeophysical effects of deforestation on mean surface temperature was also assessed using a satellite-based dataset. This dataset uses a 'space-for-time' substitution method, comparing forest and openland pixels to estimate deforestation's impact. To examine the mechanisms driving changes in DTDT, the thermodynamic energy equation was used. Changes in DTDT were linked to changes in near-surface horizontal temperature advection (TADV) and daily variability in surface sensible heat flux (DTDSHF). To investigate the impact of historical deforestation, historical simulations ('historical') from CMIP6 and historical simulations without land use changes ('hist-noLu') from LUMIP were used. Nine ESMs were used. The difference between these simulations isolates the deforestation's effects. The analysis focused on North America where significant and consistent deforestation occurred. Finally, the potential effect of future afforestation was explored using simulations under the Shared Socioeconomic Pathway (SSP) scenarios (ssp370 and ssp370-ssp126Lu) from CMIP6 and LUMIP, comparing a high-emission scenario (ssp370) with a scenario including substantial afforestation (ssp370-ssp126Lu). Six ESMs were used. The Kolmogorov-Smirnov two-sample test was used to assess statistical significance, and paired forest and openland sites from FLUXNET and AmeriFlux datasets were used to provide observational context. The analysis includes investigation of the changes in near-surface wind speed and horizontal temperature gradient, the daily variability of surface upward shortwave radiation and albedo. Calculations used both DTDT and SDT.

The study's key findings demonstrate a significant biogeophysical effect of deforestation on daily temperature variability. Idealized large-scale deforestation simulations across multiple ESMs consistently showed an increase in daily temperature variability in the northern extratropics, particularly in winter (up to a 20% increase or 0.7°C). This increase was stronger in winter, followed by spring and autumn, and less pronounced in summer. This increase in variability was robust across models, with all five models agreeing on the sign of the change. This intensification of daily temperature variability resulted in more frequent rapid extreme warming and cooling events. The primary driver for the increased variability was identified as the enhancement of near-surface horizontal temperature advection (TADV), partly offset by a reduction in the daily variability of surface sensible heat flux (DTDSHF). The analysis of historical deforestation in North America showed a detectable increase in wintertime daily temperature variability since 1850, with seven out of eight models showing increased DTDT (0.2-0.5°C) and consistent increases in TADV. This effect of historical deforestation even counteracted the overall decreasing trend in DTDT attributed to other anthropogenic forcings (GHGs and aerosols) in North America. In contrast, simulations of future afforestation projected a decrease in daily temperature variability, especially in eastern America (0.06-0.1°C reduction in winter, spring, and autumn). Analysis of changes in other surface energy balance components indicated that deforestation increased the daily variability in surface upward shortwave radiation and albedo, particularly in spring. The observational data from paired forest and openland sites partially corroborated the simulations, showing higher DTDT in openland sites during winter. However, inconsistencies emerged during spring and summer, possibly due to the influence of horizontal advection mitigating the TADV differences between paired sites. The increased frequency of extreme events was also demonstrated. For example, in North America, the frequency for wintertime δT within −3°C to 3°C decreased by 5.3%, whereas the frequency for wintertime δT below −3°C and above 3°C increased by 2.8% and 2.5%, respectively. The percentage change in the frequency for the tail regions was considerably amplified with higher magnitudes of θT.

The results of this study highlight the significant, yet often overlooked, impact of deforestation on daily temperature variability. The findings directly address the research question by demonstrating that large-scale deforestation, both idealized and historically observed, significantly increases daily temperature variability in the northern extratropics, especially during winter. The mechanism involving increased TADV and decreased DTDSHF provides a mechanistic explanation for the observed changes. The substantial impact of historical deforestation in North America, counteracting the effects of other anthropogenic forcings, emphasizes the importance of considering biogeophysical effects when assessing the overall human impact on temperature variability. The projected reduction in daily temperature variability with future afforestation offers valuable insight for policymakers, particularly regarding large-scale afforestation programs in northern extratropical countries. These results underscore the need to consider the multifaceted impacts of afforestation beyond carbon sequestration and average temperature changes, focusing on the potential benefits of mitigating increased daily temperature variability.

This study reveals a significant and previously underappreciated impact of deforestation on daily temperature variability in the northern extratropics. The increase in variability is primarily driven by enhanced near-surface horizontal temperature advection, which is partly offset by reduced variability in surface sensible heat flux. The findings emphasize the importance of considering deforestation's effects on temperature variability when evaluating human influence on climate. Conversely, future afforestation efforts are projected to reduce this variability. Future research could explore the impacts of small-scale deforestation using higher-resolution models, investigate regional variations in detail, and examine how this variability affects human and ecological systems.

The study primarily relies on model simulations, although observational data from paired sites are used to support the findings. The idealized deforestation scenario, while helpful for isolating biogeophysical effects, may not fully represent the complexities of real-world deforestation patterns. The reliance on a limited number of models for some aspects of the analysis (e.g., TADV calculation) introduces some uncertainty. The 'space-for-time' substitution method used in the observational data analysis does not fully capture the nonlocal effects of deforestation. Finally, the models used are coarse resolution, and higher resolution simulations may reveal additional complexities.