Forest ecosystems have been shown to effectively buffer climate extremes and mitigate climate change effects below the canopy [1, 2, 3, 4, 5] as they create a distinct microclimate, which is decoupled from their surrounding with substantial deviations between ambient air temperature and temperatures within and below the canopy [3, 4, 6, 7, 8, 9]. Globally, understory temperatures in forests were found to be on average 1.7 ± 0.3°C (mean) and 4.1 ± 0.5°C (maximum), cooler than macroclimate temperatures [4], which is of greater magnitude than the global warming of land and ocean temperatures (0.99°C) over the past century [10]. However, the magnitude of this cooling effect differs between forest stands. Next to edaphic factors [2, 3, 11], forest composition appears to play a major role for cooling [8, 12] as tree species differ substantially in morphological and physiological characteristics influencing the energy balance. The tree species influence is expected to be highest in the canopy and in mixed stands, this may result in a complex three-dimensional mosaic of distinct microclimatic situations [9, 13]. These create ecological temperature niches, and may thus influence diversity and composition of floristic and faunistic communities [5, 14, 15, 16, 17]. This complex ‘heatscape’ also feeds back on the physiological performance of trees themselves [18, 19], and thus directly or indirectly affects tree health [20], ecosystem functioning [20, 21] and therefore also ecosystem services provided by forest ecosystems [22, 23]. The capability of forest to maintain microclimate regulation is thus of increased importance when being faced with the consequences of climate change [4].
In general, trees cool their canopy microclimate by regulating the local energy budget via the interplay of two contrasting mechanisms, namely evapotranspiration, where water is converted to water vapor by extracting energy from the local environment [24] and also by absorbing and reflecting the incoming shortwave radiation at the top canopy which prevents the shaded canopy parts below from warming up [6]. While in hot and arid areas the cooling potential of trees can largely be attributed to shading effects [12, 25, 26], cooling by transpiration may become an important driver of canopy temperatures in the temperate zone [6, 12]. However, more frequent and longer lasting hot and dry periods, recent phenomena related to climate change [10, 27, 28, 29, 30], typically cause a reduction in tree transpiration rates [31]. Thus, for the temperate zone, the contribution of the latent heat flux to the local energy budget might be reduced by raising temperatures and less precipitation, which in turn would lead to an overall decrease in the cooling potential provided by trees. Davis et al. [3] even found evidence that some forests will lose their capacity to buffer climate extremes as sites become increasingly water limited.
Differences in the cooling effectiveness of trees species [9, 12, 32] were found to be caused by complex interactions of meteorological factors as well as leaf, canopy and hydraulic characteristics of trees [11, 33, 34]. The amount of incoming solar irradiance being reflected or absorbed by the canopy or transmitted through the canopy, depends on the total leaf surface area available, and thereby determines the degree of canopy warming [35]. Leaf temperature needs to be regulated via transpiration to avoid leaf damage [36]. Tree transpiration is highest at high solar irradiance, high temperatures and consequently high water pressure deficits (VPD), provided that water availability is not limited [37, 38, 39]. Under high evaporative demands and saturated soil water conditions, the total transpiration rate of a tree is strongly constrained by the hydraulic architecture. Significantly higher sap flux densities were reported for diffuse- than for ring-porous trees species [40, 41]. The total transpiration rate of a tree is additionally determined by morphological and physiological leaf characteristics such as the thickness of the cuticle, boundary layer resistance, stomatal density and stomatal control [37, 39, 42]. When being exposed to water stress and high ambient temperatures, trees species as e.g. English oak (Querucs robur L.) tend to close their stomata early on to maintain high leaf water potentials, whereas more water spending species as e.g. European ash (Fraxinus excelsior L.) continue to transpire to maintain their photosynthetic activity and to provide sufficient leaf cooling [40, 43, 44]. Low leaf water potentials and beginning leaf wilting, can cause reversible changes in leaf inclination, towards a more vertical position or leaf abscission if leaves are irreversible damaged, which leads to higher canopy transparency and in turn reduces the potential of cooling provided by shading [45]. The ability of a tree species to maintain canopy cooling during drought events might thus be strongly related to their water use strategies and their response dynamics which are likely to induce shifts in the species rank order of canopy temperatures as drought stress continues. However, little is known about the species signature and temporal dynamics of the influence of extended droughts on within-canopy temperatures.
As the effects of canopy structure, tree hydraulic characteristics and the environmental template on latent and sensible heat fluxes and thus on the energy budget are not uniform throughout the canopy, canopy temperatures not only differ between tree species, but also vertically in the tree crown [11, 42]. Canopy transparency, for instance, which is considered to have the greatest effect in temperature regulation as it affects both shading and transpirational cooling [12, 33, 36], exerts opposing effects on temperatures at the upper and lower canopy [42]. While the upper canopy surface temperature was found to increase with canopy density due to an increase in absorption of solar radiation [46], lower canopy temperatures decrease with increasing canopy density due to stronger radiative shielding [47]. Species-specific vertical gradients of leaf morphology, leaf angle and leaf area density were found to additionally modulate the transmission of radiation through the tree canopies [48, 49]. Also, the amount of cooling provided by transpiration varies along a vertical gradient as stomatal density per leaf area and stomatal conductance increases and the distribution of leaf area density varies along the height of a tree [50]. Differences in the magnitude and the form of those gradients affect canopy temperatures within tree crowns, but might also cause temperature differences between species at the same height layer. For instance, Richter et al. [9] found frequent and pronounced interspecific temperature differences of up to 4°C during midday for the top and the bottom canopy, but less distinct temperature differences for the middle position of the tree crown. However, due to the restricted access to forest canopies in the majority of studies, the biophysical and meteorological conditions that cause and shape those interspecific temperature variations along the height gradient, as well as their relative importance and their interactions, remain largely unknown.
High temperatures and low soil water availability additionally – as typical for recent hotter-drought events – modify temperature gradients within tree crowns due to their impact on physiological processes and structural tree characteristics. Especially, high midday temperatures force a strong stomatal regulation at the top canopy to prevent leaf water loss, which could shift the transpirational cooling to occur at lower canopy layers [19]. Gallego [39], for example, has shown that the relative contribution of different height layers on total transpiration rates varies with soil moisture availability. Peiffer et al. [51] demonstrated for beech, a water saving species, that the vertical gradient of stomatal conductance within the crown is strongly reduced under extreme soil water depletion, whereas a lower reduction in the vertical gradient of stomatal conductance can be expected for water spending tree species. Moreover, drought-related leaf damage or premature senescence can reduce both transpirational cooling and shading [52]. Whether the shift from latent heat fluxes – being under strong species-specific control – to sensible heat fluxes as stomata close during drought conditions results in an equalization of interspecific canopy temperatures is still debated. For example, Richter et al. [9] found a low interspecific variability when comparing remotely sensed canopy temperatures during a moderate drought event, whereas McGloin et al. [53] and Schwaab et al. [54] found an opposite tendency when comparing broad- and needle-leaved tree species. However, in how far direct and indirect effects of tree species characteristics and the environmental template on interspecific differences in canopy temperatures along the height gradient change as drought conditions become more severe remains currently unresolved.
To predict the consequences of a changing environment on the cooling potential of forest stands, it is therefore of high importance to capture the essential drivers and the existing direct and indirect effects between several processes in the soil-plant-atmosphere continuum that cause and shape the variations in the cooling efficiencies of contrasting tree species [11, 42]. In this study, we aim at disentangling direct and indirect effects of tree characteristics and the environmental template on the cooling potential on midday temperatures for two contrasting hydrological situations (‘moist’, ‘dry’) during the year 2018, the most severe and long-lasting summer drought and heat wave ever recorded in Central Europe [55]. We therefore measured air temperature profiles, sap flow and canopy cover of five mature co-occurring broadleaved tree species using a canopy crane facility which is located within a floodplain forest in Leipzig (Saxony, Germany). As both, the meteorological and biophysical impact on the cooling potential changes with increasing distance to the top canopy [11], we expected those effects to not only vary between tree species and spatially within the tree crown, but also with hydrological situations. We therefore aim at answering the following research questions:
Q1: Do vertical differences in canopy temperature of tree species change for two contrasting hydrological situations and do these potential changes result in a changed species rank order along the height gradient?
Q2: What are the direct and indirect effects of tree species characteristics and the environmental template such as incoming solar radiation, wind speed and ambient temperature on interspecific differences in canopy temperatures along the height gradient?
Q3: Do the direct and indirect effects of tree species characteristics and the environmental template on interspecific differences in canopy temperatures along the height gradient change as drought conditions become more severe and does the decrease in latent heat fluxes, approximated by sap flux rates, result in an equalization of tree species canopy temperatures?