RESEARCH ARTICLE The Duality of Reforestation Impacts on Surface and 10.1029/2019JG005543 Air Temperature Key Points: Kimberly A. Novick1 and Gabriel G. Katul2 • Evidence is mounting that reforestation cools the land surface 1O'Neill School of Public and Environmental Affairs, Indiana University, Bloomington, Bloomington, IN, USA, 2Nicholas in many places, but its been challenging to understand how School of the Environment, Duke University, Durham, NC, USA reforestation affects air temperature • A novel approach for detecting the fl in uence of land cover on multiple Abstract Evidence is mounting that temperate‐zone reforestation cools surface temperature (Tsurf), fl metrics of air temperature using ux mitigating deleterious effects of climate warming. While T drives many biophysical processes, air tower observations is presented surf • The analysis shows that temperature (Ta) is an equally important target for climate mitigation and adaptation. Whether reductions in reforestation in the southeastern Tsurf translate to reductions in Ta remains complex, fraught by several nonlinear and intertwined processes. United States cools the near‐surface ‐ – In particular, forest canopy structure strongly affects near surface temperature gradients, complicating air temperature by 1 3 °C during ‐ fl daytime but not nighttime cross site comparison. Here the in uence of reforestation on Ta is assessed by targeting temperature metrics that are less sensitive to local canopy effects. Specifically, we consider the aerodynamic temperature (Taero), estimated using a novel procedure that does not rely on the assumptions of Monin‐Obukhov similarity theory, as well as the extrapolated temperature into the surface layer (T ). The approach is tested with Correspondence to: extrap K. A. Novick, flux tower data from a grass field, pine plantation, and mature hardwood stand co‐located in the Duke Forest [email protected] (North Carolina, USA). During growing season daytime periods, Tsurf is 4–6 °C cooler, and Taero and near‐surface Textrap are 2–3 °C cooler, in the forests relative to the grassland. During the dormant season, Citation: daytime differences are smaller but still substantial. At night, differences in Taero are small, and near‐surface Novick, K. A., & Katul, G. G. (2020). Textrap is warmer over forests than grasslands during the growing season (by 0.5 to 1 °C). Finally, the The duality of reforestation impacts on influence of land cover on T at the interface between the surface and mixed layer is small. Overall, surface and air temperature. Journal of extrap Geophysical Research: Biogeosciences, reforestation appears to provide a meaningful opportunity for adaption to warmer daytime Ta in the 124. https://doi.org/10.1029/ southeastern United States, especially during the growing season. 2019JG005543 Plain Language Summary Reforestation—the process of reestablishing trees where they once Received 26 OCT 2019 dominated—has long been viewed as a strategy to remove CO2 from the atmosphere. Recently, attention Accepted 10 JAN 2020 has focused on understanding if reforestation also offers a direct temperature cooling benefit. By using more Accepted article online 13 MAR 2020 water (a cooling process) and increasing the transfer of heat energy away from the surface, forests may offer a meaningful opportunity for local climate mitigation and adaptation. Evidence is mounting that indeed, in the temperature and tropical zones, the surface of forests is cooler than grasslands and croplands. However, due to confounding effects of forest canopies on wind and temperature profiles near the surface, it has previously been hard to assess if forests also cool the air. Here we present a new approach that accounts for canopy effects, allowing for a more direct assessment of the potential for reforestation to cool near‐surface air temperature. Using a case study from the North Carolina Piedmont, we find that while the air cooling effect of forests is not a large as the surface cooling effect, it is still on the order of 2–3°C during summer daytime periods—times when the need for climate adaptation strategies are particularly pressing. 1. Introduction Reforestation has long been viewed as an instrument for mitigating the pace of climate change, particularly in the temperate and tropical zone where much of the historical forest cover was lost to harvest within the last 200–300 years (Williams, 1989). This view is largely linked to the carbon sequestration potential of these forests. Observations of the net ecosystem exchange of CO2 from regional and global networks of flux towers, as well as forest inventory data, reveal that forests in the temperate regions are indeed strong carbon sinks (Jung et al., 2011; Pan et al., 2011) and that even maturing temperate forests are capable of assimilating sub- stantially more CO2 than expected from conventional ecological theory (Novick et al., 2015; Stoy et al., 2008). ©2020. American Geophysical Union. However, the fate of the future forest carbon sink is less certain (Friedlingstein et al., 2014). As atmospheric All Rights Reserved. CO2 continues to rise, forest carbon uptake potential may saturate due to a number of limitations. Some are NOVICK AND KATUL 1of15 Journal of Geophysical Research: Biogeosciences 10.1029/2019JG005543 imposed by nutrient and energy availability (Baldocchi & Penuelas, 2019; Oren et al., 2001), while others are intrinsic to leaf‐level photosynthesis and its saturating behavior with increased CO2. Moreover, expected increases in air temperature, drought, and insect and fire regimes will likely decrease the magnitude, and increase the variability, of forest carbon uptake across much of the world (Frank et al., 2015; McDowell & Allen, 2015; Wear & Coulston, 2015). While this uncertainty about future forest CO2 uptake continues to motivate research, substantial atten- tion is now focused on the potential for forests to mitigate temperature conditions through alterations to the ecosystem energy balance. Theoretical links between land cover and temperature have long been recognized and incorporated into models (Avissar & Werth, 2005; Foley et al., 2003; Pielke et al., 1998; Raupach, 1991). Modeling work has shown that in the tropics, relatively large evapotranspiration causes forests to be cooler than nonforested ecosystems (Costa, 2005). On the other hand, in boreal cli- mates, relatively low forest albedo likely causes forests to be warmer than nonforested ecosystems (Lee et al., 2011; Swann et al., 2010). In the temperate zone, the overall impact of temperate reforestation on surface temperature was, for a long time, not clear (Bonan, 2008; South et al., 2011). Evaporative cool- ing and emitted longwave radiation from the surface both act to reduce surface temperature, whereas net shortwave radiative load acts to warm the surface. The sensible heat flux, whose efficiency varies with the mean wind and turbulence conditions overlying the surface, plays a dual role and may contri- bute to warming or cooling (Huang et al., 2015). With rapid advancements and proliferation of remote sensing products, observational evidence has emerged to suggest that the combined influence of increased sensible and latent heat in temperate forest ecosystems has an overall surface cooling effect that outweighs albedo‐driving warming by a magnitude of 1–2 °C, annually averaged, across a wide range of temperate ecosystems (Bright et al., 2017; Burakowski et al., 2018; Juang, Katul, et al., 2007; Wickham et al., 2012; Zhang et al., 2020). This work is encouraging, as it suggests a direct and substantial climate mitigation benefit of reforestation in the temperate zone that is mechanistically quite different from the carbon sequestration benefit. However, thus far, observation‐driven studies of land cover effects on microclimate have largely been focused on the response of surface temperature, and not the response of air temperature (e.g. Bright et al., 2017; Juang, Katul, et al., 2007; Luyssaert et al., 2014; Wickham et al., 2012; but see Baldocchi & Ma, 2013). This focus is not surprising for three reasons: (i) Unlike air temperature, surface temperature does not vary with height making it a more logical reference to compare land cover temperature patterns; (ii) likewise, surface tem- perature, representing the integrated radiometric temperature of all canopy elements, is closely coupled to the temperature experienced by foliage in dense canopies, or by microbes near the soil surface of sparse canopies, and is thus biologically relevant; and (iii) operationally, surface temperature is convenient to esti- mate from meteorological towers that report patterns of radiation, albedo, and energy fluxes necessary to attribute variations in surface temperature between ecosystems to specific mechanisms (Juang, Katul, et al., 2007; Lee et al., 2011; Luyssaert et al., 2014). However, when considering direct impacts of reforestation and other land cover changes on climate, air tem- perature (hereafter Ta), as opposed to surface temperature (Tsurf), is an equally important metric. This rele- vance is certainly true for boundary layer dynamics and rainfall initiation (Juang, Katul, et al., 2007, Juang, Porporato, et al., 2007; Manoli et al., 2016; Siqueira et al., 2009) as well as a plethora of associated “hand‐shakes” between the climate system and the land surface (Baldocchi & Ma, 2013; Luyssaert et al., 2014). Ultimately, climate change is driven by long‐term increases in the temperature of the air (or kinetic temperature) due to increases in