Seasonal dynamics and sun/shade heterogeneity of leaf gas exchange and VOC emissions inside a tall temperate forest canopy

22 Leaf gas exchange is the key driver of forest carbon uptake and directly determines forest 23 carbon sink activity. Additionall y, plants release a variety of biogenic volatile organic 24 compounds (VOCs) acting as stress signals of trees. However, continuous hourly resolved 25 measurements of leaf gas exchange and VOC emissions in tall tree canopies are challenging 26 and remain scarce. To this end, we developed a sophisticated in -situ leaf gas exchange 27 measurement system with 24 cuvettes deployed o n mature Fagus sylvatica (n=3) and 28 Pseudotsuga menziesii (n=3) individuals in a mixed temperate forest. We additionally 29 measured sap flux density ( Js), radial growth and tree water deficit ( TWD) to gain a holistic 30 picture of seasonal leaf and stem water and carbon flux dynamics during the summer of 2024. 31 During midsummer, w e found a gradual reduction of stomatal conductance ( gs) and VOC 32 emissions of sun, but not shade branchlets of P. menziesii in response to moderate 33 atmospheric and edaphic drying. Decreased gs led to a downregulation of transpiration (E), Js, 34 and carbon isotope discrimination accompanied by an increase in TWD and intrinsic water 35 used efficiency . Leaf gas exchange of shade branchlets remained unaffected due to 36 microclimatic buffering effects. Contrarily, sun leaves of F. sylvatica , profited from sunny 37 midsummer conditions and increased leaf gas exchange, wh ereas shade leaves benefitted 38 from more diffuse light during early summer exhibiting similar carbon assimilation, transpiration 39 and VOC emissions as sun leaves. For both species we found a clear time lag of four to five 40 hours between maximum leaf and stem water fluxes and a delay of up to 20 hours for the 41 recovery of TWD, highlighting the role of stem water reserves. 42 Pronounced seasonal and diurnal differences o f leaf gas exchange, stem water fluxes and 43 VOC emissions showed, that continuous data are essential to better understand variability of 44 ecosystem flux dynamics. 45 46

continuous leaf gas exchange measurement, leaf carbon and water fluxes, stem 47 water fluxes, temperate forest trees, Fagus sylvatica, Pseudotsuga menziesii, carbon isotope 48 discrimination 49 50 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 3

51 Forests represent a major terrestrial carbon sink, taking up approximately 20% of annual 52 anthropogenic CO2 emissions (Pan et al. 2024; Friedlingstein et al. 2025) . Yet, increasing 53 intensity of heat waves and droughts strongly compromise the carbon sink strength (Ciais et 54 al. 2005; Schuldt et al. 2020; Gharun et al. 2024; Haberstroh et al. 2025; Knutzen et al. 2025; 55 Werner et al. 2025). Leaf gas exchange is the key driver of carbon uptake and water loss in 56 forest ecosystems shaping plant productivity and stress responses. Stomata jointly regulate 57 carbon uptake and water release on the leaf level , which are both affected by microclimat ic 58 conditions (e.g. air temperature, solar irradiance, water availability, atmospheric humidity), and 59 concentration gradients of CO2 and H2O between the atmosphere, the leaf boundary layer and 60 the chloroplast (Martínez-Vilalta et al. 2014; Salmon et al. 2020). In response to drought stress, 61 stomata close to prevent excessive water loss which has a cascading effect on intrinsic water 62 use efficiency (WUEi) and leaf 13C discrimination (Δleaf). An increase in WUEi and a decrease 63 in Δleaf is expected in response to stomatal closure, yet the magnitude of change depends on 64 the species-specific stress tolerance (Cernusak et al. 2013; Werner et al. 2021) . Moreover, 65 leaf exchange rates of biogenic volatile organic compounds (VOCs) via the stomata and their 66 composition provide insights into stress status, defense mechanisms and communication 67 strategies of trees (Laothawornkitkul et al. 2009) . VOC emissions and compositions are 68 species-specific and de novo synthesis is mainly controlled by microclimate and availability of 69 fresh assimilates , yet se veral species including conifers also possess specialized storage 70 compartments emitting VOCs dependent on air temperature and partially decoupled from 71 synthesis (Lerdau et al. 1997; Niinemets et al. 2004a; Dindorf et al. 2006; Holzke et al. 2006; 72 Joó et al. 2011; Van Meeningen et al. 2016). Under heat stress, VOC emissions are generally 73 increased and s ome compounds, such as isoprene and monoterpenoids, are known to 74 increase membrane stability and thermotolerance (Peñuelas and Llusià 2003; Copolovici et al. 75 2005; Werner et al. 2020; Bourtsoukidis et al. 2024; Meischner et al. 2024). Moderate drought 76 stress induces an initial increase of VOC emissions, while under severe drought decreased 77 emission rates in accordance with reduced photosynthetic rates were observed (Bertin et al. 78 1997; Staudt et al. 2002; Wu et al. 2015; Haberstroh et al. 2018; Kreuzwieser et al. 2021; 79 Daber et al. 2025) . Beyond their biogenic importance, VOCs are highly reactive in the 80 atmosphere and have a strong impact on atmospheric chemistry and air quality (Atkinson and 81 Arey 2003; Folberth et al. 2006; Guenther et al. 2012). These examples illustrate the need for 82 continuous measurements of leaf gas exchange and VOC emissions in forested ecosystems, 83 which remain scarce. 84 Moreover, leaf-level fluxes strongly impact stem -level fluxes (Andrade et al. 1998; Köcher et 85 al. 2009), however, only few studies have assessed the seasonal and diurnal coordination 86 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 4 between the two levels so far (Peters et al. 2023; Peters et al. 2025) . V arious studies 87 demonstrated a time lag of up to two hours between sap flow rates at the stem top and at the 88 stem base, which they explained by the withdrawal of stem water reserves contributing 9-25% 89 to total daily water use (Phillips et al. 2003; Cermak et al. 2007; Köcher et al. 2013; Leuschner 90 et al. 2024). High stem hydraulic capacitance can significantly enhance drought -tolerance of 91 species and reduce daily fluctuations of water potentials (Betsch et al. 2011; Leuschner et al. 92 2024). Depletion of stored stem water due to soil water limitations results in decreased 93 predawn water potentials ultimately triggering stomatal closure (Peters et al. 2025). Moreover, 94 stomata sensitively respond to high vapor pressure deficit (VPD), which is predicted to increase 95 in the future (Novick et al. 2016; Grossiord 2020; Werner et al. 2025), but due to microclimatic 96 buffering inside the canopy, this response will differ between the sun and shade c rown 97 (Niinemets 2023). 98 Microclimatic gradients inside the tree crown are driven by vertical gradients of photosynthetic 99 photon flux density ( PPFD), resulting in decreased air temperature (Tair) and VPD in shaded 100 areas of the canopy compared to the surrounding atmosphere (Dai et al. 2004; Zellweger et 101 al. 2020; De Frenne et al. 2021) . Consequently, shade foliage responds less pronounced to 102 edaphic drying, since concurrent atmospheric water demand and the n eed for stomatal 103 regulation are lower (Niinemets et al. 2004 b; Peltoniemi et al. 2012; Richter et al. 2022) . 104 Contrarily, sunlit foliage is more prone to water limitations and therefore usually operates with 105 a higher water use efficiency and earlier stomatal closure than the shade canopy (Niinemets 106 et al. 2004 b; Valladares et al. 2016) . Vertical microclimatic buffering highly depends on air 107 mass exchange (Flerchinger et al. 2015), canopy density (De Frenne et al. 2021; Gillerot et al. 108 2021; Zhang et al. 2022) , tree species composition and their respective shade casting ability 109 (Ehbrecht et al. 2019; Zellweger et al. 2019; Wang et al. 2025). Microclimatic buffering capacity 110 within the canopy is assumed to mitigate negative impacts of global warming on forest 111 ecosystems (He et al. 2018; De Frenne et al. 2019) , and might even enhance the current 112 contribution of the shade canopy to gross primary productivity of 50% (He et al. 2018) to 70% 113 (Sprintsin et al. 2012). 114 Contributions of shade foliage to whole -canopy carbon gain is determined by PPFD 115 penetration into the canopy . High foliage clumping, as in coniferous species, reduces PPFD 116 penetration, but at the same time enhances penumbra of foliage, i.e. the partial diffuse shade 117 from the shadow cast of other leaves, which redistributes light more evenly within the crown 118 (Stenberg 1998; Miyashita et al. 2012; Way and Pearcy 20 12). Generally, augmenting the 119 diffuse fraction of incoming light, e.g. by aerosol scattering or cloud coverage, enhances whole-120 canopy carbon gain due to increased carbon assimilation of shade foliage (Dai et al. 2004; 121 Knohl and Baldocchi 2008; Zhou et al. 2021) , since d iffuse light better penetrates into the 122 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 5 canopy and reduces the total area of fully shaded leaves (Roderick et al. 2001; Gu et al. 2002; 123 Urban et al. 2007) . Higher light availability in the shade under overcast conditions can also 124 increase VOC emissions of forest stand s due to higher contributions of the shade foliage 125 (Laffineur et al. 2013). 126 Since both leaf gas exchange and VOC emissions are controlled by microclimatic gradients, 127 simultaneous in-situ measurements in various canopy layers are crucial to better understand 128 responses of forest canopies to meteorological conditions and validate results of current 129 modelling approaches (e.g. Niinemets et al. 2002b, Sprintsin et al. 2012, Guenther et al. 2012, 130 He et al. 2018, Chang et al. 2018) . Due to logistic challenges , there are only few studies 131 measuring leaf -level gas exchange and VO C emissions continuously in tall tree canopies 132 (Hakola et al. 2006; Kolari et al. 2007; Kolari et al. 2009; Aalto et al. 2014; Werner et al. 2021). 133 Most data on gas exchange and VOC emissions in mature stands were derived from 134 campaign-based measurements neglecting either vertical or seasonal and day -to-day 135 variations (e.g. Bertin et al. 1997, Kesselmeier et al. 1997, Staudt et al. 1997, Niinemets et al. 136 2002a, Pressley et al. 2004, Plaza et al. 2005, Holzke et al. 2006, Niinemets et al. 2010, 137 Šimpraga et al. 2011, Van Meeningen et al. 2016) . While diurnal and seasonal dynamics of 138 total ecosystem net carbon, water and VOC exchange can be estimated by eddy covariance 139 measurements (Laffineur et al. 2013; Haberstroh et al. 2022; Pohl et al. 2023; Scapucci et al. 140 2024), we lack continuous data on real-time regulation processes of VOC emissions, 141 corresponding leaf gas exchange and the coordination between stem - and leaf-level fluxes, 142 which can significantly enhance our understanding of processes driving seasonal dynamics 143 under varying meteorological conditions. 144 To this end, we developed and installed a continuous leaf gas exchange measurement system 145 in a tall tree canopy to assess seasonal dynamics of sun and shade foliage. We sampled VOC 146 emissions in regular campaigns from July to October and measured sap flux density, tree water 147 deficit and radial growth in the tree stems to gain a holistic picture of whole -tree carbon and 148 water fluxes. We selected a forest stand containing a deciduous ( Fagus sylvatica ) and a 149 coniferous species ( Pseudotsuga menziesii) which differed in canopy architecture, foliage 150 clumping (Stenberg 1998; Miyashita et al. 2012) , water use strategy (Schumann et al. 2024; 151 Paligi et al. 2025) and VOC storage capacity (Lerdau et al. 1995; Lerdau et al. 1997; Holzke 152 et al. 2006; Holzke et al. 2006) to answer the following research questions: (1) Are there 153 seasonal differences of leaf gas exchange and VOC emissions between both species? (2) 154 How do stomatal regulation and VOC emissions differ between the sun and shade canopy? 155 (3) How are stem-level and leaf-level water fluxes coordinated during the day? Thereby, we 156 aim to better understand seasonal and within-canopy dynamics of leaf gas exchange and VOC 157 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 6 emissions resulting of changing PPFD, VPD and Tair. We further wanted to gain new insights 158 into the relationship between diurnal stem and leaf-level water fluxes regulation. 159 160

161 Field site and experimental design 162 The field site is located in the Black Forest on a plateau at an elevation of 520 m a.s.l. close to 163 Ettenheim, Baden-Württemberg, Germany (Tesch et al. 2025) and part of the ECOSENSE 164 collaborative research centre (Werner et al. 2024). Soil in this area is classified as a carbonate-165 free Cambisol with a silty loam to loamy clay texture originating from sandstone (Werner et al. 166 2024). Long-term mean Tair and mean annual precipitation of the region was 11°C and 911 167 mm, respectively, from 1991 to 2020 (Tesch et al. 2025) . The stand is mainly composed of 168 European beech ( Fagus sylvatica L.), but also contains Norway spruce ( Picea abies (L.) H. 169 Karst.), Silver fir (Abies alba Mill.), European larch (Larix decidua Mill.), Sessile oak (Quercus 170 petraea (Matt.) Liebl.), and larger patches of Douglas fir ( Pseudotsuga menziesii (Mirbel) 171 Franco). The site is equipped with three canopy access towers, one in a patch dominated by 172 F. sylvatica, one in a patch dominated by P. menziesii and one in a more mixed patch, where 173 both species interact. Five individuals of F. sylvatica and five individuals of P. menziesii were 174 selected around a canopy-access tower which provided access to the sun (26 m) and shade 175 (24 m) crowns of three individuals of each species. Diameters at breast height of the se 176 individuals ranged between 50.1±11.2 cm and 37.4±5.5 cm for P. menziesii and F. sylvatica, 177 respectively. Tree height and age varied between 27-30 m and 45-50 years for P. menziesii 178 and 24-28 m and 55-110 years for F. sylvatica. 179 180 Environmental data 181 In February 2023, a meteorological station was installed on an open area 250 m southwest of 182 the field site. The station continuously measured shortwave radiation using a CNR4 sensor 183 (Kipp & Zonen, Delft, Netherlands), Tair and relative humidity (RH) with a HygroVUE5 sensor 184 (Campbell Scientific Ltd., Shepshed, UK) as well as precipitation with a Young 52202 Tipping 185 Bucket (R.M. Young Company, Traverse City, MI, USA). Additionally, in Mai 2024, a LI -190 186 quantum sensor (LI-COR Environmental, Lincoln, NE, USA) was installed above the canopy 187 (46 m height) to measure PPFD and two HygroVUE10 sensors ( Campbell Scientific Ltd., 188 Shepshed, UK) were in installed in 27 m and 18 m height to measure Tair and RH in the sun 189 and shade canopy, respectively. Data w ere recorded on CR1000 data loggers (Campbell 190 Scientific Ltd., Shepshed, UK) every minute. VPD was calculated from Tair and RH using the 191 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 7 guideline provided by the WMO (WMO 2018). PPFD at the open area was calculated from 192 shortwave radiation data using a linear regression model with the PPFD sensor above the 193 canopy. 194 At four l ocations four SMT100 soil moisture sensors (Truebner GmbH, Neustadt, Germany) 195 were installed at four depths (0.30 m, 0.50 m, 0.70 m, 0.90 m) . Additionally, 35 SMT100 196 sensors were distributed radially around the measurement trees at 0.05 m depth . Sensors 197 recorded volumetric soil water content ( VWC) every 15 minutes on a data logger (CR350, 198 Campbell Scientific Ltd., Shepshed, UK). Data quality of VWC was checked using the quality 199 control procedure by Dorigo et al. (2013) and only values with a ‘good’ quality flag were 200 selected for analysis. 201 Leaf water potentials (ψleaf) were measured once a month on small branchlets of the sun crown 202 of all individuals which could be reached by the canopy -access tower using a Scholander 203 pressure chamber (Soil Moisture Equipment, Santa Barbara, CA, USA , Scholander et al. 204 1965). Predawn water potential (ψPD) was measured in the two hours prior to sunrise and 205 midday water potential (ψMD) in the two hours around solar noon from May to October 2024. 206 207 Sap flow and dendrometer measurements 208 In autumn 2023, tree stems were equipped with heat pulse velocity sensors (SFM5, UGT, 209 Muencheberg, Germany) at breast height (1.3 m above the ground) to determine sap flux 210 density (Js) at two different xylem depths (10 and 20 mm < the cambium). Insulating foil was 211 wrapped around the sensor heads to protect them from UV -radiation and Tair fluctuations. 212 Measured data were recorded on a data logger (CR1000, Campbell Scientific Ltd., Shepshed, 213 UK) every 15 minutes and transmitted in real-time to the ECOSENSE data base (Tesch et al. 214 2025). 215 The Dual Meth od Approach (Forster 2019, 2020), combining the heat ratio method (HRM) 216 (Burgess et al. 2001; Marshall 1958) and the Tmax method (Cohen et al. 1981), was used to 217 calculate Js separately for the inner and outer xylem. Total Js of the active sapwood was then 218 calculated by scaling Js of inner and outer xylem to the respective proportion of the sapwood 219 area measured by the thermistors. For the determination of sapwood area, tree cores were 220 drilled in October 2024 at breast height and illuminated immediately. Translucent zones were 221 classified as functional sapwood with active water transport (Munster -Swendsen 1987; 222 Quiñonez-Piñón and Valeo 2018). Active sapwood area was 158.8 ± 35.3 cm2 in P. menziesii 223 and 117.6 ± 17.4 cm 2 in F. sylvatica. Needle probe misalignment was corrected by solving 224 HRM and Tmax equations for needle distances (Kinzinger et al. 2024; Dumberger et al. 2025). 225 During nights (PPFD < 10 µmol m -2 s-1 ) with VPD < 0.1 kPa, no sap flow was expected, 226 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 8 resulting in heat velocity values equal to zero in both equations. Corrected needle distances 227 between the 5% and 95% percentiles of each individual tree were then interpolated for all 228 nights with non -zero flow conditions. Two sensors in two individuals of F. sylvatica failed 229 temporarily leading to data gaps. Linear regression models with an adjacent tree, which 230 showed a similar trend of Js (Spearman´s rho: 0.81-0.91), were calculated and used to fill the 231 data gaps. 232 Custom-made point dendrometers as describ ed in Wang and Sammis (2008) were installed 233 on the stems at 1.3 m above the ground to record stem radius variation as in Dumberger et al. 234 (2025). A potentiometer (Model 9605 BEI, Duncan Electronics, Commerce, Texas, USA) 235 mounted to a metal angle bracket and two stainless -steel rods was attached to the trees 236 directly touching the stem (Wang and Sammis 2008). In all individuals of P. menziesii the upper 237 layers of bark were removed gently to ensure direct contact to water conducting wood tissue. 238 Sensors were connected to CR1000 data loggers (Campbell Scientific Ltd., Shepshed, UK) 239 and recorded data at an interval of 15 minutes. Raw data w ere further processed with the 240 “treenetproc” package which was used to remove outliers and calculate radial growth and tree 241 water deficit ( TWD) by the zero -growth approach (Zweifel et al. 2016, Knüsel et al. 2021). 242 During the whole vegetation period of 2024, values of one F. sylvatica individual had to be 243 removed from the data set due to sensor failure. 244 245 Leaf and needle cuvettes 246 For the measurements of leaf gas exchange and VOC emissions, two different types of leaf 247 enclosure systems were used. Single leaves of F. sylvatica were equipped with a novel, 248 lightweight and minimally invasive cuvette (“ECOvette”, Frey et al. 2025a , Figure 1, 249 Supplementary Material Figure S1 A & B). The ECOvette comprises a hemispherical capsule 250 which is attached to the leaf by two ring magnets with an ultra-soft silicone seal and covers a 251 leaf area of 6.16 cm2. ECOvettes were equipped with an Tair and RH sensor (SHT40, Sensirion 252 GmbH, Gerlingen, Germany) inside and outside of the cuvette and a thermocouple (P TF-50, 253 Type K, PeakTech Prüf - und Messtechnik GmbH, Ahrensburg, Germany) to measure leaf 254 temperature (Tleaf, Frey et al. 2025b) . Sensor values were sent every ten seconds via a Wifi 255 from a microcontroller (ESP8266 Wemos D1 Mini, Espressif Systems, Shanghai, China) to a 256 Raspberry Pi (Raspberry Pi Trading Ltd., Cambridge, UK) which logged the data on a central 257 database (Tesch et al. 2025) . A more detailed description and a proof of concept for gas 258 measurements with the ECOvette can be found in Frey et al. (2025a). 259 Small branchlets of P. menziesii were enclosed into larger cuvettes custom-built for conifer 260 branchlets (Werner et al. 2021) made of FEP film (PTFE -Spezialvertrieb GmbH, Stuhr, 261 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 9 Germany) and a supporting frame of PFA-tubing (1/4”, Wolf-Technik eK, Stuttgart, Germany). 262 Cuvettes were equipped with small fans (MF40101V2 -1000-UA99, Sunon, Kaohsiung City, 263 Taiwan) to ensure homogenous mixing of the air and sealed air -tight with teroson (Teroson 264 RBII, Henkel AG & CoKG, Düsseldorf, Germany) and rubber ban d (Supplementary Material 265 Figure S1 C). We additionally attached the environmental sensors of six ECOvettes to three 266 conifer cuvettes in the sun and shade canopy, respectively, to measure Tair, Tleaf and RH. At 267 the end of the measurement period or if cuvet tes were damaged after thunderstorms, 268 branchlets were harvested and projected needle area was determined in the laboratory using 269 a commercial scanner connected to the GSA Image Analyzer Software (GSA GmbH, Rostock, 270 Germany). Needles were detached from the branchlets and spread evenly on the scanner 271 surface with as minimum overlapping as possible. 272 Twelve ECOvettes and twelve conifer cuvettes were installed in the sun (n=6 per species) and 273 shade (n=5 per species) canopy of three individuals of F. sylvatica and P. menziesii , 274 respectively (Figure 1A). We measured PPFD at the cuvettes with a hand-held light meter (LI-275 250A, LI-COR Environmental, Lincoln, NE, USA) on several days throughout the entire study 276 period to ensure proper shading of the shade cuvettes. Shade foliage received on average 9% 277 (53.7 ± 24.7 µmol m-2 s-1) and 8% (33.8 ± 3.8 µmol m-2 s-1) of PPFD received by sun foliage of 278 F. sylvatica (555.9 ± 71.9 µmol m -2 s-1) and P. menziesii (544.9 ± 111.4 µmol m -2 s-1), 279 respectively. One cuvette of each type was left empty as a reference for the incoming air into 280 the cuvettes. A small piece of FEP film (PTFE -Spezialvertrieb GmbH, Stuhr, Germany) was 281 used to seal the empty ECOvette gas-tight. 282 Environmental sensors of the ECOvettes failed during several periods due to technical issues. 283 Data gaps were filled using linear regression models with adjacent sensors (R2 = 0.68-0.99), 284 if available, or with values measured at the canopy -access tower (R2 = 0.47-0.98) in 27 m 285 height (sun leaves) and 18 m height (shade leaves). For six conifer cuvettes, which were not 286 equipped with environmental sensing, Tair, Tleaf and RH values from adjacent sun and shade 287 cuvettes were used. Mounting of the thermocouples to the needles of P. menziesii was 288 challenging, and constant direct contact with needles could not be ensured. We thus used the 289 temperature inside the cuvette (Tcuv) for calculations of gas exchange and further analysis of 290 microclimatic conditions. 291 292 Automated air sampling systems 293 An automated air sampling system (Figure 1B) was established in the forest following the 294 approach already used in climate chambers (Fasbender et al. 2018; Werner et al. 2020) and 295 in an artificial tropical rain forest (Werner et al. 2021). 296 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 10 Air was supplied to the c uvettes from the tower base by a compressor (AG OFCAS 36 Plus, 297 LNI Swissgas GmbH, Kamen, Germany) connected to a buffer tank with a volume of 100 L 298 (LNI Swissgas GmbH, Kamen, Germany), which sent compressed air to the tower platform (26 299 m height) through PFA -tubing (1/4”, Wolf -Technik eK, Stuttgart, Germany) with a constant 300 pressure of 2 bar. On the tower platform , the supplied air was distributed to 24 thermal mass 301 flow controllers (G -Series, Brooks Instrument GmbH, Dresden, Germany) by two custom 302 aluminum manifolds. Mass flow controllers regulated the air flow to the ECOvettes to 350 ml 303 min-1 and to the conifer cuvettes to 800 ml min-1 accounting for the differing air volume of the 304 cuvettes (5.8 cm3 & ~1000 cm3). Supplied air was delivered to the conifer cuvettes by 1/4” PFA 305 tubing and to the ECOvettes by 1/8” PFA -tubing. Sampling air was pumped with a vacuum 306 pump (KNF LABOPORT® N840G, Faust Lab Science GmbH, Klettgau, Germany) situated at 307 the tower base to a distribution unit at the tower platform consisting of two flow -through 308 multiposition valves (16 positions, 1/8”, Valco Instruments Co. Inc., Schenken, Switzerland) 309 which automatically switched between the 24 cuvettes. Sampled air from the measured cuvette 310 was sent through 1/4” PFA tubing at a rate of 200 ml min -1 controlled by a thermal mass flow 311 controller (G-Series, Brooks Instrument GmbH, Dresden, Germany) to an Isotope and Gas 312 Concentration Analyzer (G -2131-i, Picarro Inc., Santa Clara, USA) at the tower base 313 measuring 12CO2, 13CO2 and H2O concentrations every two seconds. Remaining cuvettes were 314 constantly flushed at a flow rate of 200 ml min -1 controlled by flowmeters (SHLLJ, Sorekarin, 315 Yueqing, China) to ensure continuous supply of fresh CO2 and avoid condensation of 316 transpired H2O. Switching between the cuvettes was realized every nine minutes, of which the 317 last 60 seconds were averaged. For data quality control, the standard deviation of the averaged 318 60 s interval was calculated and measurements exceeding a standard deviation of 2.5 ppm 319 (12CO2) and/or 0.05 % (H2O) were removed from the data, as the measurement was likely not 320 stable. Additionally, 16 full days and several single data points were removed d ue to 321 maintenance work at the cuvettes, damage of the cuvettes or failure of the whole measurement 322 system. After this rigid data cleaning process, 65% of the initial data (6452 of 9964 data points) 323 were further analyzed. 324 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 11 325 Figure 1: Schematic overview of the automated measurement system. The left panel (A) shows the 326 distribution of sun and shade cuvettes in the canopy of F. sylvatica (n=3) and P. menziesii (n=3). Empty 327 cuvettes represent the cuvettes used for measuring refer ence air. The right panel (B) shows the 328 connection between the air supply unit (black color), the measurement unit (blue color) and the constant 329 flushing unit (grey color) of the automated measurement system. Flow rates are indicated in green color. 330 331 Calculation of leaf gas exchange 332 For the calculation of leaf gas exchange, 12CO2, 13CO2 and H 2O concentration s of the air 333 entering the leaf cuvettes were assumed to be equal to the air measured inside the reference 334 air cuvettes. Reference air cuvettes were measured every 54 minutes and extended to the 335 entire data set by linear interpolation . Differences between 12CO2, 13CO2 and H 2O 336 concentrations of the interpolated reference air and the air exiting the leaf cuvettes were then 337 used to calculate transpiration (E, Supplementary Material Formula S1), net carbon 338 assimilation (Anet, Supplementary Material Formula S2), stomatal conductance for water vapor 339 (gs, Supplementary Material Formula S3), leaf discrimination for 13C during photosynthesis 340 (Δleaf, Supplementary Material Formula S4), water use efficiency ( WUE, Supplementary 341

Formula S5) and intrinsic water use efficiency ( WUEi, Supplementary Material 342 Formula S6) of enclosed leaves and b ranchlets based on von Caemmerer and Farquhar 343 (1981) and Evans et al. (1986). 344 345 Measurements of biogenic volatile compound emissions 346 For the measurements of VOC emissions, an additional 1/4” PFA tee connector and a short 347 piece of 1/4” PFA tubing were installed between the PFA tubing coming from the cuvettes and 348 the two multi -position valves (Figure 1B) . On thirteen da ys from July to October , VOC 349 emissions were collected from the cuvettes between 11 am to 2 pm on glass tubes filled with 350 Tenax (Sigma-Aldrich, Munich, Germany) at a rate of 5 0 ml min -1 for one hour with pumps 351 (Pocket Pump Touch, SKC, Dorset, UK). After sampling, tubes were stored in airtight vials 352 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 12 (Labco Limited, Lampeter Ceredigion, United Kingdom) in the fridge (4°C) until they were 353 measured with a Gas Chromatograph (GC, Mod el 7820A and 7890B, Agilent Technologies, 354 Santa Clara, USA), see Kreuzwieser et al. (2021) for a detailed description. Briefly, samples 355 were heated to 240°C in a thermal desorption unit (TDU, Gerstel, M ühlheim a.d. Ruhr, 356 Germany), thereafter cryotrapped at -70°C (KAS, Gerstel, Mühlheim a.d. Ruhr, Germany), 357 reheated to 240°C at a rate of 12°C s -1 and subsequently released into a separation column 358 (DB-5MS UI, Agilent Technologies, Santa Clara, USA). For the separation of the compounds 359 the column was heated for 54 minutes in three steps from 45°C to 280°C. Separated 360 compounds were then introduced into the mass spectrometer, operating at 70.35 eV, an ion 361 source temperature of 230°C and a quadrupole temperature of 150°C. Measured mass spectra 362 were analyzed using the MassHunter Software (Agilent Technologies, Santa Clara, USA). 363 Concentrations of the compounds were quantified using a standard mixture composed of eight 364 compounds (isoprene, ⍺-pinene, β -pinene, limo nene, sabinene, trans -β-ocimene, 365 caryophyllene and farnesene). VOC fluxes were calculated using Formula S7 (Supplementary 366 Material). 367 368 Data analysis 369 Data analysis was performed using the software R (version 4.5.1) and RStudio (version 370 2025.09.1, R Core Team 2025)). 371 For the analysis of seasonal dynamics and diurnal courses we classified the vegetation period 372 into three periods. We selected these three periods since Tair, VWC at 0.05 m depth and VPD 373 differed, indicative for a potential heat and/or drought stress. Meteorological conditions during 374 early summer (18/06/2024 to 19/07/2024) were warm and wet, during midsummer (24/07/2024 375 to 29/08/2024) we recorded the warmest period with highest VPD and lowest VWC, and during 376 late summer (09/09/2024 to 18/09/2024) cold conditions with low VWC were observed (Table 377 1). Furthermore, continuous data from our leaf gas exchange measurements system were 378 available during all three periods. 379 We computed three different linear mixed effect models (package “lme4”) using the three tree 380 individuals of each species as a random effect to account for repeated measurements (Bates 381 et al. 2015). We further analyzed the models comparing contrasts of estimated marginal means 382 (package “emmeans”) for the respective predictor variables (Lenth 2023). Model assumptions 383 were checked with the package “performance” and, if necessary, response variables were 384 transformed as specified below (Lüdecke et al. 2021) . The first model calculated d ifferences 385 of Js, radial growth and TWD between the two species on a monthly resolution . A cube root 386 transformation was applied to Js and a square root transformation to radial growth and TWD, 387 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 13 respectively, to ensure normality of residuals and homogeneity of variance. The second model 388 identified d ifferences between the measurement campaigns and between the species for 389 predawn and midday ψleaf. The third model analyzed d ifferences between log-transformed 390 VOC emission rates of the sun and shade foliage for each of the thirteen sampling dates. 391 Additionally, to the estimated marginal means a type III ANOVA was used as a post-hoc test 392 to analyze seasonal effects influencing VOC emissions. 393 For the analysis of diurnal courses, two different generalized additive models (package “mcgv”) 394 were fitted and estimated marginal means of the models were compared (Wood 2017; Lenth 395 2023). The first model was fitted for Tcuv, VPD, gs, Anet and E with species, sun/shade exposition 396 and the three predefined periods as predictors, a smoothing spline for the interaction of the 397 three predictors, a thin-plate regression spline for time of a day on a half-hourly resolution and 398 tree individual as a random effect. The second model was fitted for normalized values of Js, gs, 399 E, VPD and TWD with species and the three predefined periods as predictors, a smoothing 400 spline for the interaction of the two predictors, a thin -plate regression spline for time of a day 401 on a 15-minute resolution and tree individual as a random effect. The second model was fitted 402 separately for the sun and shade foliage. Normalized values were obtained for each tree 403 individually by dividing the respective variable by the 99% quantile of maximum values over 404 the whole season. 405 406 Table 1: Daily minimum soil water content ( VWC) in 0.05 m depth , daily maximum vapor 407 pressure deficit ( VPD) and daily maximum air temperature ( Tair) ± the standard error for the 408 respective variable averaged over the three analyzed periods. 409 early summer (18th Jun - 19th Jul) midsummer (24th Jul - 29th Aug) late summer (9th Sep – 18th Sep) VWC [Vol. %] 26.6 ± 0.3 17.1 ± 0.4 16.9 ± 0.2 VPD [kPa] 1.2 ± 0.1 1.5 ± 0.1 0.4 ± 0.1 Tair [°C] 23.8 ± 0.7 26.2 ± 0.6 14.1± 0.5 410 411 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 14

412 Meteorological conditions at the field site 413 The summer of 2024 was dominated by wet and warm conditions with no pronounced heat 414 waves, but two cold spells (beginning of July and mid-September) with a sudden drop of 10 K 415 in Tair (Figure 2 A&B). VWC remained high in all layers during early summer, while a slight dry-416 down period was recorded from mid-July to early September, when VWC in 0.05 m depth 417 dropped to 12.9 ± 1.4 % (Figure 2C). Js resembled meteorological conditions with pronounced 418 reductions during the colder periods (Figure 2D). Js of F. sylvatica tended to be higher than Js 419 of P. menziesii during July and August. During the other months , Js of P. menziesii slightly 420 exceeded Js of F. sylvatica, albeit this difference was only significant during March and April 421 (p < 0.01), when le aves of F. sylvatica were not yet fully developed. Radial growth started 422 earlier in coniferous P. menziesii, but daily growth rate was only significantly higher compared 423 to F. sylvatica during April (p < 0.05). There was no difference in the daily growth rate between 424 the species during May, but during June and July, F. sylvatica exceeded the daily growth rate 425 of P. menziesii (p = 0.05) which resulted in a slightly higher total radial increment for F. sylvatica 426 (3.6 ± 0.7 mm) than for P. menziesii (3.2 ± 0.9 mm, Figure 2E). A significant TWD was absent 427 for both species until the end of June; however, coinciding with decreasing VWC (Figure 2C), 428 TWDmin of P. menziesii increased in August to 35.5 ± 33.8 µm and stayed elevated until 429 October. For F. sylvatica, TWDmin only reached maximum of 12.1 ± 6.0 µm at the end of the 430 study period in October (Figure 2F). 431 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 15 432 Figure 2: Environmental conditions (A-C) as well as stem-level water and carbon fluxes (D-F) at the field 433 site during the vegetation period of 2024. Panels A-C show (A) daily mean air temperature ( Tair, dark 434 red line) and Tair on a 5-minute resolution (bright red line), daily mean vapor pressure deficit (VPD, dark 435 grey line) and VPD on a 5-minute resolution (bright grey line), (B) daily maximum photosynthetic photon 436 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 16 flux density above the canopy (PPFDmax, yellow points and line) and daily sum of precipitation (blue 437 bars), and (C) volumetric soil water content (VWC) in five depths (0.05, 0.30, 0.50. 0.70. 0.90 m). Panels 438 D-F show (D) daily sum of sap flux density (Js), (E) cumulative radial growth, and (F) daily minimum tree 439 water deficit ( TWDmin) of F. sylvatica (red line, n=5) and P. menziesii (blue line, n=5), respectively. 440 Shaded areas in panel C-F show the standard error, while lines show the mean va lues of four soil pits 441 (C) and five tree individuals (D-F), respectively. Due to sensor failure after heavy rainfalls during spring 442 and autumn, available data of VWC (C) were limited to the period between June and October. 443 444 Seasonal leaf gas exchange 445 Leaf gas exchange differed substantially not only between the two investigated species, 446 between sun and shade foliage and between predefined periods (Figure 3). On average, daily 447 maximum Anet and E measured at single leaves of F. sylvatica were higher than rates 448 measured at small branchlets of P. menziesii (Figure 3 C -F) and showed more pronounce d 449 fluctuations in response to changing weather conditions. However, comparisons between the 450 two species need to be interpreted carefully, since penumbral effect s in measured shoots of 451 P. menziesii cannot be excluded. 452 Differences between the predefined periods were observed in the comparison of shade and 453 sun foliage (mean daily maxima in brackets): during early summer, Anet of shade leaves of F. 454 sylvatica (9.8 ± 0.7 µmol m-2 s-1) nearly reached the same rates as in sun leaves (11.1 ± 0.5 455 µmol m-2 s-1), whereas during midsummer Anet of sun exposed leaves (15.5 ± 0.5 µmol m-2 s-1) 456 clearly exceeded shade exposed leaves (6.7 ± 0.6 µmol m-2 s-1) and the same pattern was also 457 observed in E (Figure 3C&E). A pronounced decrease of Anet and E in the sun foliage of F. 458 sylvatica was visible during the two periods with low Tair and low PPFD at the beginning of July 459 and in mid -September, while rates of the shade foliage remained relatively stable over the 460 season and were less affected by changes in PPFD and Tcuv. 461 Anet of the sun branchlets of P. menziesii exceeded rates of the shade branchlets during most 462 of early summer (sun: 3.5 ± 0.2 µmol m-2 s-1 and shade: 1.6 ± 0.1 µmol m-2 s-1) and midsummer 463 (sun: 3.1 ± 0.1 µmol m-2 s-1 and shade: 1.4 ± 0.1 µmol m-2 s-1), which was also resembled in E. 464 Both, Anet and E, were generally more influenced by changing meteorological conditions in the 465 sun than in the shade. At the beginning of early summer and especially during the cold spell 466 in September shade foliage of P. menziesii reached Anet and E rates comparable to the sun 467 foliage. 468 Δleaf, representing the in-situ discrimination of the leaf against 13C during photosynthesis, 469 fluctuated in response to changing meteorological conditions and showed high diurnal 470 variability (Supplementary Material Figure S 2). Δleaf of b oth species showed a pronounced 471 decline by 3-6 ‰ in response to the cold spell in mid-September concurrent with an increase 472 of WUEi (Supplementary Material Figure S3). During midsummer, daily mean Δleaf was lower 473 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 17 in the sun leaves than in the shade leaves of F. sylvatica (17.7 ± 0.4 ‰ and 19.8 ± 0.5 ‰, 474 respectively) and P. menziesii (21.1 ± 0.4 ‰ and 22.6 ± 0.5 ‰, respectively). A pronounced 475 reduction of Δleaf was observed in the second half of midsummer, particularly in P. menziesii, 476 which reduced Δleaf in the sun leaves by 6 ‰ and in the shade leaves by 7‰. 477 During this period, we detected a gradual dry-down, when VWC at 0.05 m depth decreased to 478 12.9 ± 1.4 % and VPD increased to 3.5 ± 0.1 kPa and 3.4 ± 0.2 kPa in the sun foliage of F. 479 sylvatica and P. menziesii, respectively (Figure 4 A-C). Concurrently, ψPD (-1.0 ± 0.1 MPa) and 480 ψMD (-2.2 ± 0.2 MPa) of P. menziesii decreased significantly (p < 0.05) compared to the values 481 measured in early summer (ψPD: -0.3 ± 0.1 MPa , ψMD: -1.2 ± 0.1 MPa ) and both were also 482 significantly lower than those of F. sylvatica (p < 0.01, Figure 4A). ψPD of F. sylvatica remained 483 unaffected, while ψMD also decreased significantly to -1.7 ± 0.1 MPa compared to early summer 484 (p < 0.001). Sun foliage of P. menziesii responded with a gradual decline of gs during 485 midsummer, reducing maximum gs by 45% (from 195.2 ± 22.8 to 86.1 ± 12.2 mmol m -2 s-1), 486 whereas maximum gs of sun leaves of F. sylvatica (188.4 ± 6.5 mmol m-2 s-1) was unaffected 487 by drier conditions (Figure 4 D &E). Reduced gs in the sun branchlets of P. menziesii was 488 accompanied by a twofold increase of WUEi during midsummer compared to early summer, 489 which was not visible in the shade branchlets (Supplementary Material Figure S3). Shade 490 foliage of P. menziesii maintained similar gs throughout the whole study period and reached 491 highest values (88.7 ± 11.3 mmol m -2 s-1) during midsummer. During midsummer, s hade 492 foliage of F. sylvatica also slightly reduced gs by 20% to 85.1 ± 6.2 mmol m-2 s-1, but recovered 493 again to 101.5 10.9 mmol m-2 s-1 at the end of the period after a heavy precipitation event. 494 495 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 18 496 Figure 3: Continuous measurements of carbon and water fluxes on the leaf -level from June to 497 September 2024. Lines and shaded areas display the mean and standard error of all measured cuvettes 498 (n=5-6) on an hourly resolution, respectively. Shown are (A & B) the temperature inside the cuvette 499 (Tcuv) and photosynthetic photon flux density measured at the top of the tower in 46 m height (PPFD, 500 yellow line), (C & D) ne t carbon assimilation rate ( Anet), and (E & F) the transpiration rate ( E) of single 501 leaves of F. sylvatica (A, C, E, n=3) in the sun (orange line) and shade (red lines) canopy and of small 502 branchlets of P. menziesii (B, D, F, n=3) in the sun (bright blue line) and shade (dark blue line) canopy. 503 Large gaps were caused by technical failure of the measurement system due to heavy wind, rain or 504 heat. Due to low data availability, transpiration of shade leaves of F. sylvatica in panel E is displayed as 505 single points. Note the different y-axis scales between panel C and D and between panel E and F. 506 507 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 19 508 509 Figure 4: Soil water content in 0.05 m depth ( VWC, A , brown line ), leaf water potentials ( ψleaf, A) 510 measured on sunlit twigs of F. sylvatica (n=3, red points) and P. menziesii (n=3, blue triangles) before 511 sunrise (predawn, filled symbols ) and around noon (midday , empty symbols ), vapor pressure deficit 512 (VPD) at the leaf enclosures (B & C) and stomatal conductance (gs) of enclosed leaves of F. sylvatica 513 (D, n=3) and enclosed branchlets of P. menziesii (E, n=3) in the sun (bright color) and shade (dark color) 514 canopy. Shaded areas and error bars show the standard error of the respective variable, while lines and 515 points show the mean of all individuals (n=3 per species) and soil pits ( n=4). Grey shaded area in the 516

highlights the period of midsummer. 517 518 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 20 Seasonal VOC emissions 519 For both species, a declining seasonal trend of monoterpene emissions was detected (p < 520 0.001), which differed between the species (p < 0.001). In sun foliage of F. sylvatica, we found 521 a gradual decline of emissions towards late summer . We could not detect any monoterpene 522 emissions by sun leaves of F. sylvatica on the last two measurement dates which was probably 523 related to early senescence in the sun canopy. In the sun foliage of P. menziesii , two 524 pronounced declines of emissions from early summer to midsummer and from midsummer to 525 late summer were observed (Figure 5 A&B). The first decline occurred simultaneously with 526 decreasing gs and ψleaf during the dry -down period (Figure 4E) and the second during the 527 sudden cold spell (Figure 3B). Monoterpene emissions of the sun foliage of both species were 528 usually higher than those of the shade foliage, albeit this difference was only significant on four 529 days for F. sylvatica due to the high variability of emissions between the tree individuals (Figure 530 5 A&C). On cloudy days, differences between emissions of sun and shade foliage diminished 531 in both species, particularly in F. sylvatica, leading to similar monoterpene emissions in both 532 canopy layers. 533 Overall, 32 compounds could be identified from measured samples of which only the seven 534 monoterpenes were further analyzed. On average, sun leaves of F. sylvatica emitted mainly 535 sabinene (89.5 ± 3.3%), but also Δ3-carene (9.5 ± 3.4%) and limonene (1.3 ± 0.6%), while in 536 the shade, monoterpenoid emissions were more equally composed of sabinene ( 52.8 ± 537 11.4%), limonene (28.3 ± 12.4%) and Δ3-carene (23.6 ± 8.9%). P. menziesii emitted a more 538 diverse blend of monoterpenoids which was mainly composed of ⍺-pinene (sun: 55.5 ± 5.5%, 539 shade: 42.8 ± 7.1%), sabinene (sun: 20.1 ± 3.8%, shade: 31.7 ± 5.9%) and ꞵ-pinene (sun: 540 20.1 ± 3.8%, shade: 12.7 ± 1.6%), but also to a lower extent of camphene ( sun: 8.5 ±1.3%, 541 shade: 12.1 ± 4.4%) and Δ3-carene (sun: 2.7 ± 0.9%, shade: 5.5 ± 3.4%). Additionally, during 542 September and October low emissions of 1,8-cineol (0.6-17.8 nmol m-2 s-1) were detected in 543 the sun and shade foliage of P. menziesii. 544 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 21 545 Figure 5: Monoterpene emissions in the sun ( A & B) and shade canopy ( C & D) of single leaves of F. 546 sylvatica (A & C, n=3) and small branchlets of P. menziesii (B & D, n=3) measured during 13 campaigns 547 from July to October 2024. Error bars show the standard error calculated with error propagation . 548 Asterisks indicate a significant difference between the monoterpene emission sum of sun and shade 549 leaves (* p < 0.05). Note the different y -axis scale between panels A&B and C&D due to the different 550 emissions rates of the two species. 551 552 Diurnal courses of leaf gas exchange, sap flux density and tree water deficit 553 In both species, w e found seasonal and within-canopy differences of diurnal courses of leaf 554 gas exchange. During midsummer, microclimatic differences between sun and shade leaves 555 of both species were highest with maximum Tcuv differences of 2.8 K (p < 0.05) and 3.8 K (p 0.05) and 0.8 kPa (p < 0.05) in F. 557 sylvatica and P. menziesii, respectively. Differences of Tcuv (1.3°C and 0.7°C, respectively) and 558 VPD (0.1 kPa and 0.1 kPa , respectively) were lower during early summer and during late 559 summer, but due to low data availability no model could be fitted for this period (Figure 6 A-F). 560 Strong difference of VPD and Tcuv in midsummer resulted in 1.5 -fold higher gs and E and 561 twofold higher Anet of sun leaves of F. sylvatica compared to shade leaves (p < 0.001), while 562 no significant difference was found during early summer (Figure 6G-O). The deviation of sun 563 and shade leaves of F. sylvatica during midsummer was mainly driven by an increase of gs, 564 Anet and E in the sun compared to early summer (p < 0.001), whereas leaf gas exchange in 565 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 22 the shade remained similar. Contrarily, the difference of leaf gas exchange between sun and 566 shade branchlets of P. menziesii remained constant during midsummer compared to early 567 summer (p > 0.05), since no increase was observed for gs, E and Anet despite higher Tcuv and 568 VPD. Stem water dynamics of both species resembled water fluxes of sun leaves with a 1.5-569 fold increase of maximum Js (p < 0.001) and no significant increase of TWD in F. sylvatica 570 from early summer to midsummer , whereas only a non-significant increase of Js and a 571 significant twofold increase of TWD (p < 0.001) of P. menziesii was observed (Figure 7). 572 A substantial time lag was detected between stem - and leaf-level water f luxes, particularly 573 during midsummer (Figure 7): in sun leaves of F. sylvatica, maximum Anet and gs were reached 574 in the morning (Anet: 10:00 am, gs: 10:30 am), followed one hour later by maximum E and 2.5 575 hours later by maximum PPFD and another 2 hours later by maximum VPD. In the shade 576 foliage of F. sylvatica, maximum gs (11 am) was followed by a simultaneous peak of Anet and 577 PPFD (12.30 pm), 1.5 hours later by maximum E and another hour later by maximum VPD. Js 578 and TWD peaked 1 and 1.5 hours after maximum VPD and 5 and 5.5 hours after maximum gs 579 was reached in the canopy, respectively. In P. menziesii, maximum gs occurred first in in shade 580 (10:30 am) and then in sun branchlets (11 am ), followed by maximum Anet (sun: 11:30 am, 581 shade: 12 pm) and PPFD (12:30 am), and a simultaneous peak in E (1 pm). Maximum VPD 582 was reached 3.5 (sun) to 5.5 hours (shade) after the peak of gs was recorded and another 2-583 3.5 hours later, already in the evening, maximum Js (6 pm) and TWD (6:30 pm) were observed. 584 The same order was observed for both species during early summer, albeit time lags were 585 slightly shorter (Supplementary Material Figure S4). Strong species differences were observed 586 for the time lag until trees replenished their stem water reserves during midsummer: refilling in 587 F. sylvatica was significantly faster (10:30 pm, 6 hours after maximum Js) than in P. menziesii 588 (06:45 am, 13 hours after max Js). 589 590 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 23 591 Figure 6: Mean diurnal courses of cuvette temperature (Tcuv, A-C), cuvette vapor pressure deficit (VPD, 592 D - F), stomatal conductance (gs, G-I), net carbon assimilation (Anet, J-L) and water vapor transpiration 593 (E, M-O) in the sun (solid lines, filled symbols) and shade canopy (dashed lines, empty symbols) of 594 single leaves of F. sylvatica (n=3) and small branchlets of P. menziesii (n=3) during the three pre-defined 595 periods. Points represent the hourly mean of the tree in dividuals during the respective period. Lines 596 show a fitted generalized additive model, while shaded areas show the standard error of the model. 597 Model fit of shade leaves of F. sylvatica in late summer is missing due to scarce data availability. 598 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 24 599 Figure 7: Diurnal courses of normalized sap flux density ( Js), stomatal conductance ( gs), transpiration 600 (E) and tree water deficit ( TWD) of sun (A) and shade ( B) leaves of F. sylvatica (n=3) and of sun ( C) 601 and shade (D) branchlets of P. menziesii (n=3) during midsummer (see Table 1) . Normalized values 602 were calculated using the 99% quantiles of maximum values of the respective variable measured during 603 the vegetation period of 2024. Shaded areas in the background show the standard error, while lines 604 show the fitted curve of a generalized additive model. 605 606 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 25

607 In 2024, we established a sophisticated measurement system enabling continuous 608 measurements of leaf gas exchange and campaign -based sampling of monoterpene 609 emissions inside the tall tree canopy of a mixed temperate forest and combined it with stem-610 level water flux measurements. In P. menziesii, high VPD during midsummer was buffered in 611 the shade and thus , only the sun canopy reduced stomatal conductance and monoterpene 612 emissions in response to a period of gradual edaphic and atmospheric drying. In F. sylvatica, 613 we found that shade foliage contributes significantly to whole-canopy photosynthetic carbon 614 uptake and monoterpene emissions under overcast conditions, since net carbon assimilation, 615 transpiration and emissions did not differ between the sun and shade canopy during overcast 616 periods with low Tair. We found substantial time lags between water fluxes in the stem and the 617 canopy indicating a significant contribution of stored stem water reserves to leaf -level 618 transpiration. In both species, our continuous measurement system revealed a high seasonal 619 and day-to-day variability of VOC emissions, leaf gas exchange and stem-level water fluxes, 620 which can provide a way forward in understanding coordination of whole-tree water and carbon 621 fluxes. 622 Microclimatic gradients shape within-canopy response to soil drying 623 In response to a dry-down period in midsummer, sun foliage of P. menziesii gradually 624 decreased gs and increased WUEi, while the shade foliage was unaffected. Stomatal closure 625 of P. menziesii resulted in a regulation of Js, E and Anet, which did all not increase from early 626 summer to midsummer despite more favorable weather conditions, a decline of ψPD, ψMD and 627 Δleaf, and an increase in TWD, which could not be fully replenished overnight (Figure 3, 4 & 7). 628 Hence, the increasing water potential gradients in response to high atmospheric demand in 629 the sun triggered stomatal closure to prevent excess loss of water and potentially protect stem 630 water reserves and ensure sufficient cell turgor (Dumberger et al. 2025; Peters et al. 2025; 631 Salomón and Camarero 2025). This is in line with findings from Schumann et al. (2024) and 632 Paligi et al. (2025) who found a more isohydric response for P. menziesii than for F. sylvatica 633 under drought, with early closure of stomata. Moreover, s tomata of sun foliage generally 634 respond faster to more water-limiting conditions in the soil, since they are exposed to higher 635 levels of PPFD, Tair and VPD (Niinemets et al. 2004b). Therefore, gs of sun foliage responds 636 more sensitive to reductions in soil moisture, while shade foliage is more decoupled from soil 637 water availability leading to dampened carbon uptake in the sun canopy under mild water 638 stress (Wozniak et al. 2020). Microclimatic buffering of Tair and VPD in the shade is assumed 639 to reduce negative impacts of severe heat and drought stress on whole-canopy carbon uptake 640 (He et al. 2018) . Our findings show, that even under moderate soil drying , lower VPD in the 641 shade of P. menziesii prevented shade foliage from the gradual, twofold reduction of gs 642 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 26 observed in sun foliage and the subsequent regulation of Anet and E despite warmer and 643 sunnier conditions. However, absence of stomatal regulation also led to lower WUEi in shade 644 foliage of P. menziesii and it is still uncertain how lower WUE of shade leaves will limit their 645 contributions to canopy photosynthesis under severe long-term stress (Way and Pearcy 2012; 646 Valladares et al. 2016). Lower WUEi in the shade is a result of microclimatic buffering and high 647 residual gs under low PPFD ensuring high intercellular CO 2 concentrations and thus, a quick 648 activation of RuBisCo to efficiently exploit dynamic light environments (Way and Pearcy 2012; 649 Campany et al. 2016). Future research with continuous leaf gas exchange data as presented 650 in this study will be needed to better understand within -canopy dynamics of gs under severe 651 stress conditions . However, from our data , we can show that under moderate soil and 652 atmospheric drying lower WUEi of shade foliage has no negative impact on Anet or E. 653 Simultaneous to decreasing gs, ψleaf and VWC during midsummer, we also found decreased 654 VOC emission rates of sun foliage in P. menziesii. Mild drought stress has been found to 655 slightly increase monoterpene emissions, while severe drought stress drastically reduces 656 emissions rates (Staudt et al. 2002; Wu et al. 2015; Haberstroh et al. 2018). We did not observe 657 severe drought stress, but declining soil moisture did induce a reduction of gs and thereby a 658 down-regulation of E and Js, thus we conclude that at least partially low emissions of P. 659 menziesii during midsummer were elicited by decreasing soil moisture availability. P. menziesii 660 comprises of specialized storage compartments, from which VOCs can be emitted under 661 sufficient air temperature even if de novo synthesis rates are low (Lerdau et al. 1995; Joó et 662 al. 2011). However, we also observed a clear seasonal decline in monoterpene emissions from 663 July to September in F. sylvatica, which has also been observed by other studies (Pressley et 664 al. 2004; Hakola et al. 2006; Holzke et al. 2006; Aalto et al. 2014) . From our data, we cannot 665 disentangle whether low monoterpene emissions of P. menziesii during midsummer resulted 666 from decreasing VWC, decreased de novo synthesis , or from a seasonal decline. For future 667 studies, the analysis of ẟ13C values of emitted compounds could improve our understanding of 668 driving factors of seasonal VOC emissions (Ghirardo et al. 2010; Haberstroh et al. 2019; Daber 669 et al. 2025). 670 We did not find a downregulation of gs in F. sylvatica during midsummer, but differences of 671 VPD and Tcuv between sun and shade leaves were also lower (Figure 6 B & E) and no strong 672 reduction of ψPD could be detected. Lower difference in Tcuv and VPD between sun and shade 673 foliage of F. sylvatica compared to P. menziesii were most likely related to the increase in E 674 and Js and the subsequent transpirational cooling effect. F. sylvatica is known to respond with 675 a more anisohydric, water -spending strategy to drying, maintaining transpiration even under 676 slight water stress (Schumann et al. 2024; Paligi et al. 2025). However, during midsummer we 677 observed that gs and E of sun leaves reached their peak values around 1.5 hours earlier than 678 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 27 shade leaves, and three hours before highest VPD was recorded (Figure 7). This time lag was 679 probably related to the higher radiation load in sun leaves and the subsequent faster increase 680 of VPD during the morning hours. Thus, we conclude that, even though gs of more anisohydric 681 F. sylvatica was not reduced during midsummer, gs still responded to rising VPD under high 682 PPFD, yet not as drastically as observed in more isohydric P. menziesii. 683 Shade leaves of F. sylvatica benefit from overcast skies 684 During early and late summer, we found similar gs, Anet and E in the sun and shade leaves of 685 F. sylvatica under overcast skies with frequent precipitation events, whereas sun leaves clearly 686 exceeded shade leaves in gs, Anet and E during midsummer (Figure 3 & 4). Higher maximum 687 photosynthetic and electron transport capacity of sun leaves optimize them to efficiently utilize 688 direct sunlight, while shade leaves are b etter adapted to exploit diffuse light environments 689 (Niinemets 2023). Increasing the fraction of diffuse light to direct light in modelling simulations 690 resulted in 10 -50% higher canopy assimilation rates , which were mainly related to higher 691 contributions of the shade foliage (Meir et al. 2002; Dai et al. 2004) . Diffuse light can better 692 penetrate the canopy and increase the radiation load on otherwise light -limited shade leaves 693 and has been found to contribute around 64% to global gross primary productivity despite 694 contributing only 54% to total available light (Knohl and Baldocchi 2008; Zhou et al. 2021) . 695 Global modelling showed that on average shade canopy contributes to ~48% of gross primary 696 productivity in broadleaved deciduous forests (He et al. 2018) and that big -leaf models 697 underestimate canopy productivity between 20% and 70% by neglecting differing behavior of 698 shade leaves (Meir et al. 2002; Dai et al. 2004; Sprintsin et al. 2012) . In F. sylvatica, only 699 around 20% of total leaf area index is comprised of sunlit leaves (Scartazza et al. 2016), which 700 would imply that in our study around 80% of whole-canopy assimilation was achieved by shade 701 foliage during early summer and late summer. 702 Furthermore, we found that monoterpene emissions of the sun leaves of F. sylvatica were 703 usually higher than those of the shade leaves on sunny days, but that this difference diminished 704 on cloudy days due to higher emissions of the shade leaves than under sunny conditions 705 (Figure 5). Using an eddy covariance system in a mixed forest ecos ystem, Laffineur et al. 706 (2013) also found higher VOC emissions under cloudy than under clear-sky conditions, which 707 was probably related to the higher contribution of shade leaves to emissions on these days . 708 Since F. sylvatica does not comprise of specializ ed VOC storage structures, emissions are 709 mainly dependent on PPFD, Tair and substrate availability from photosynthesis (Niinemets et 710 al. 2004 a; Dindorf et al. 2006; Van Meeningen et al. 2016; Bourtsoukidis et al. 2024) . 711 Therefore, we assume that better penetration of diffuse light into the canopy and resulting 712 higher photosynthetic activity fostered monoterpene emissions of the shade leaves. 713 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 28 We did not find the same pattern in the shade foliage of P. menziesii, where gs, Anet and E were 714 lower than in the sun foliage throughout the entire study period (Figure 3). Most likely, different 715 canopy architecture of our coniferous and deciduous species led to the observed differences, 716 since foliage clumping is higher in P. menziesii than in F. sylvatica which distributes light more 717 evenly within the canopy (Stenberg 1998) . Such penumbral effect s need to be considered 718 regarding our gas exchange measurements, since whole branchlets of P. menziesii were 719 enclosed into the cuvettes and calculations were done with projected leaf area. For better 720 comparability, development of a needle cuvette system as deployed for F. sylvatica would be 721 useful (Frey et al. 2025a), but remains challenging due to the needle arrangement around the 722 branchlets and small leaf area of a single needle compared to a single leaf of F. sylvatica. 723 Seasonal and spatial variability of monoterpene fluxes 724 Both VOC emissions and leaf gas exchange were in the range of previous studies 725 investigating our two species in campaign -wise measurements (Lerdau et al. 1995; Pressley 726 et al. 2004; Dindorf et al. 2006; Holzke et al. 2006; Joó et al. 2011; Šimpraga et al. 2011; 727 Laffineur et al. 2013; Scartazza et al. 2016; Durand et al. 2022) , but clearly lack diurnal and 728 day-to-day variability and/or seasonal effects, which we were able to detect with our continuous 729 measurement system. Our data can crucially enhance parametrization of current upscaling 730 approaches and help to improve model uncertainties elicited by microclimatic variability 731 (Wozniak et al. 2020), foliage clumping effects (Chen et al. 2012), seasonal dynamics (Keenan 732 et al. 2009; Guenther et al. 2012; Grote et al. 2013; Chang et al. 2018) , and plant hydraulic 733 constraints (Peltoniemi et al. 2012; Niinemets 2023). 734 Stomatal aperture is tightly coordinated between stem-level water supply and leaf-level water 735 demand 736 We observed substantial time lags between diurnal stem - and leaf -level water fluxes, 737 particularly during midsummer (Figure 7). In P. menziesii, refilling of stem water reserves lasted 738 the entire night which was most likely related to reduced soil water supply and lower hydraulic 739 conductance of coniferous xylem tracheids compared to ring -porous vessels of F. sylvatica 740 (Berdanier et al. 2016). Differences in hydraulic conductance between both species were also 741 visible in the time lag between maximum canopy VPD and maximum Js in the stem, which was 742 only one hour in F. sylvatica but three hours in P. menziesii. Moreover, since in both species 743 maximum E occurred earlier than maximum Js and simultaneously we observed an increase 744 in TWD, water demand during peak E was partially supplied by stored stem water. This is in 745 line with previous studies which observed that stored stem water contributed 9 -16% in F. 746 sylvatica (Köcher et al. 2013; Leuschner et al. 2024) and 20-25% in P. menziesii (Phillips et al. 747 2003; Cermak et al. 2007) to total water consumption. This also indicates that during the 748 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 29 afternoon, when peak Js occurred, transported water was partially used to refill depleted stem 749 water reserves along the stem, since demand by the crown was already reduced during this 750 time of the day. Water withdrawal from the stem during morning hours to meet transpirational 751 demand by th e crown and refilling of depleted tissue by Js during the afternoon was also 752 observed by (Köcher et al. 2013) for five broadleaved temperate species. 753 Generally, in both species , we observed a temporal relationship between peak values of Js 754 and VPD, E and PPFD, and gs and Anet. Interestingly, in both species, maximum gs peaked 755 close to maximum Anet and occurred four to six hours before maximum VPD was reached 756 indicating a tight stomatal control in response to VPD. We see three potential explanations for 757 stomatal regulation: (1) stomatal aperture is optimized for maximum carbon gain and maximum 758 Anet happened already in the morning (Henry et al. 2019; Deans et al. 2020; Franklin et al. 759 2023), (2) reduced VWC, delayed stem water refilling and resulting decline in ψPD led to early 760 stomatal closure (Peters et al. 2025) , (3) or a species-specific threshold of VPD (VPD > 0.5 761 kPa) elicits stomatal regulation independent of maximum VPD (Oren et al. 1999; Novick et al. 762 2016; Grossiord et al. 2020), as it has also been observed for the onset of radial growth (Zweifel 763 et al. 2021). However, we showed that maximum gs occurred earlier during midsummer than 764 during early summer and also earlier in the sun than in the shade foliage of both species . 765 Therefore, we conclude that stomatal regulation represents a fine-tuned coordination between 766 ψleaf, VPD and Anet at the leaf level and TWDmin and VWC at the stem base an d cannot be 767 explained by one factor alone. 768 769 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 30

770 Continuous measurements of leaf gas exchange in the sun and shade canopy of F. sylvatica 771 and P. menziesii revealed strong seasonal dynamics of stomatal regulation as well as of VOC 772 emissions. More isohydric P. menziesii responded sensitive to increasing VPD with a reduction 773 of stomatal conductance and VOC emissions in the sun branchlets and a downregulation of 774 sap flux density, while shade branchlets did not respond due to microclimatic buffering of VPD. 775 Contrarily, sun leaves of more anisohydric F. sylvatica profited from higher air temperature and 776 VPD in midsummer, while shade leaves were limited by low light availability. Under overcast 777 skies, shade leaves of F. sylvatica benefitted from better penetration of diffuse light into the 778 canopy and therefore showed similar carbon assimilation and VOC emissions to sun leaves. 779 Time lags between stem - and leaf -level water fluxes were most pronounced during 780 midsummer, when air temperature and VPD were highest, and up to 20 hours were needed to 781 replenish stem water reserves. Pronounced seasonal courses of stomatal regulation 782 processes and VOC emission patterns, which varied between the sun and shade canopy, 783 demonstrated that continuous leaf gas exchange data in different canopy layers combined with 784 stem-level flux measurements are indispensable for a better understanding of whole 785 ecosystem water and carbon fluxes under stressed and non-stressed conditions. 786 787 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 31 Funding 788 Financial support w as obtained by the German Research Foundation (DFG) through the 789 ECOSENSE collaborative research center (Project-ID: 459819582-SFB 1537). 790

791 We would like to thank Lennart Nettler and Phyllis Lua -Mellmann for assistance with sensor 792 installations and maintenance and L. Erik Daber and Tim Ehrhardt for the support during the 793 construction and installation of the measurement system. We thank Mirjam Meischner for the 794 fruitful discussions of the on-site development of the measurement system and on calculations 795 of leaf gas exchange data. We also gratefully acknowledge permission from the city of 796 Ettenheim to set-up our field site in their city forest. 797 Author Contributions 798 Leading author: Stefanie Dumberger: experimental setup, data collection, analysis, and 799 interpretation, manuscript writing 800 Yasmina Frey: sensor development, data collection, manuscript revision 801 Clara Stock: experimental setup, data collection and interpretation, manuscript revision 802 Sophie Wehlings-Schmitz: data collection and analysis of VOC samples 803 Delon Wagner: electrical setup of the measurement equipment, software programming 804 Kathrin Kühnhammer: electrical setup and maintenance of the measurement equipment 805 Lea Dedden and Markus Weiler: data acquisition of soil moisture sensors, manuscript revision 806 Markus Sulzer and Andreas Christen: data acquisition of meteorological data, manuscript 807 revision 808 Jürgen Kreuzwieser: analysis and interpretation of VOC emissions, manuscript revision 809 Ulrike Wallrabe: study conception, manuscript revision, supervision 810 Christiane Werner: study conception, data interpretation, manuscript revision, supervision 811 Simon Haberstroh: study conception, data analysis and interpretation, manuscript revision 812 813 Open Research : The da ta that support the findings of this study are available from the 814 corresponding author upon reasonable request. 815 816 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 32

817 Aalto J et al. 2014. New foliage growth is a significant, unaccounted source for volatiles in 818 boreal evergreen forests. Biogeosciences. 11(5):1331–1344. https://doi.org/10.5194/bg-11-819 1331-2014 820 Andrade JL et al. 1998. Regulation of water flux through trunks, branches, and leaves in trees 821 of a lowland tropical forest. Oecologia. 115(4):463–471. 822 https://doi.org/10.1007/s004420050542 823 Atkinson R, Arey J. 2003. Gas-phase tropospheric chemistry of biogenic volatile organic 824 compounds: a review. Atmos Environ. 37:197–219 (The 1997 Southern California Ozone 825 Study (SCOS97-NARSTO). Dedicated to the Memory of Dr. Glen Cass (1947-2001)). 826 https://doi.org/10.1016/S1352-2310(03)00391-1 827 Bates D, Mächler M, Bolker B, Walker S. 2015. Fitting Linear Mixed-Effects Models Using 828 lme4. J Stat Softw. 67(1):1–48. https://doi.org/10.18637/jss.v067.i01 829 Berdanier AB, Miniat CF, Clark JS. 2016. Predictive models for radial sap flux variation in 830 coniferous, diffuse-porous and ring-porous temperate trees Phillips N, editor. Tree Physiol. 831 36(8):932–941. https://doi.org/10.1093/treephys/tpw027 832 Bertin N et al. 1997. Diurnal and seasonal course of monoterpene emissions from Quercus 833 ilex (L.) under natural conditions application of light and temperature algorithms. Atmos 834 Environ. 31:135–144 (BEMA: A European Commision Project on Biogenic Emmissions in 835 the Mediterranean Area). https://doi.org/10.1016/S1352-2310(97)00080-0 836 Betsch P et al. 2011. Drought effects on water relations in beech: The contribution of 837 exchangeable water reservoirs. Agric For Meteorol. 151(5):531–543. 838 https://doi.org/10.1016/j.agrformet.2010.12.008 839 Bourtsoukidis E et al. 2024. High temperature sensitivity of monoterpene emissions from 840 global vegetation. Commun Earth Environ. 5(1):23. https://doi.org/10.1038/s43247-023-841 01175-9 842 Campany CE, Tjoelker MG, von Caemmerer S, Duursma RA. 2016. Coupled response of 843 stomatal and mesophyll conductance to light enhances photosynthesis of shade leaves under 844 sunflecks. Plant Cell Environ. 39(12):2762–2773. https://doi.org/10.1111/pce.12841 845 Cermak J et al. 2007. Tree water storage and its diurnal dynamics related to sap flow and 846 changes in stem volume in old-growth Douglas-fir trees. Tree Physiol. 27(2):181–198. 847 https://doi.org/10.1093/treephys/27.2.181 848 Cernusak LA et al. 2013. Environmental and physiological determinants of carbon isotope 849 discrimination in terrestrial plants. New Phytol. 200(4):950–965. 850 https://doi.org/10.1111/nph.12423 851 Chang K-Y, Paw U KT, Chen S-H. 2018. Canopy profile sensitivity on surface layer 852 simulations evaluated by a multiple canopy layer higher order closure land surface model. 853 Agric For Meteorol. 252:192–207. https://doi.org/10.1016/j.agrformet.2018.01.027 854 Chen JM et al. 2012. Effects of foliage clumping on the estimation of global terrestrial gross 855 primary productivity. Glob Biogeochem Cycles. 26(1) [accessed 2025 Nov 14]. 856 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 33 https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2010GB003996. 857 https://doi.org/10.1029/2010GB003996 858 Ciais P et al. 2005. Europe-wide reduction in primary productivity caused by the heat and 859 drought in 2003. Nature. 437(7058):529–533 860 Cohen Y, Fuchs M, Green G. 1981. Improvement of the heat pulse method for determining 861 sap flow in trees. Plant Cell Environ. 4(5):391–397 862 Copolovici LO et al. 2005. The Capacity for Thermal Protection of Photosynthetic Electron 863 Transport Varies for Different Monoterpenes in Quercus ilex. Plant Physiol. 139(1):485–496. 864 https://doi.org/10.1104/pp.105.065995 865 Daber LE et al. 2025. Position-specific isotope labelling gives new insights into chiral 866 monoterpene synthesis of Norway spruce (Picea abies L.). Environ Exp Bot. 238:106238. 867 https://doi.org/10.1016/j.envexpbot.2025.106238 868 Dai Y, Dickinson RE, Wang Y-P. 2004. A Two-Big-Leaf Model for Canopy Temperature, 869 Photosynthesis, and Stomatal Conductance. J Clim. 17(12):2281–2299. 870 https://doi.org/10.1175/1520-0442(2004)017%253C2281:ATMFCT%253E2.0.CO;2 871 De Frenne P et al. 2019. Global buffering of temperatures under forest canopies. Nat Ecol 872 Evol. 3(5):744–749. https://doi.org/10.1038/s41559-019-0842-1 873 De Frenne P et al. 2021. Forest microclimates and climate change: Importance, drivers and 874 future research agenda. Glob Change Biol. 27(11):2279–2297. 875 https://doi.org/10.1111/gcb.15569 876 Deans RM, Brodribb TJ, Busch FA, Farquhar GD. 2020. Optimization can provide the 877 fundamental link between leaf photosynthesis, gas exchange and water relations. Nat Plants. 878 6(9):1116–1125 879 Dindorf T et al. 2006. Significant light and temperature dependent monoterpene emissions 880 from European beech (Fagus sylvatica L.) and their potential impact on the European volatile 881 organic compound budget. J Geophys Res Atmospheres. 111(D16) [accessed 2025 Dec 17]. 882 https://onlinelibrary.wiley.com/doi/abs/10.1029/2005JD006751. 883 https://doi.org/10.1029/2005JD006751 884 Dorigo W a. et al. 2013. Global Automated Quality Control of In Situ Soil Moisture Data 885 from the International Soil Moisture Network. Vadose Zone J. 12(3):vzj2012.0097. 886 https://doi.org/10.2136/vzj2012.0097 887 Dumberger S et al. 2025. Dynamic Shifts in Radial Sap Flow of Two Temperate Tree Species 888 in Response to the Dry Summer 2022. Ecohydrology. 18(4):e70054. 889 https://doi.org/10.1002/eco.70054 890 Durand M et al. 2022. Sunflecks in the upper canopy: dynamics of light-use efficiency in sun 891 and shade leaves of Fagus sylvatica. New Phytol. 235(4):1365–1378. 892 https://doi.org/10.1111/nph.18222 893 Ehbrecht M et al. 2019. Effects of structural heterogeneity on the diurnal temperature range in 894 temperate forest ecosystems. For Ecol Manag. 432:860–867. 895 https://doi.org/10.1016/j.foreco.2018.10.008 896 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 34 Flerchinger GN, Reba ML, Link TE, Marks D. 2015. Modeling temperature and humidity 897 profiles within forest canopies. Agric For Meteorol. 213:251–262. 898 https://doi.org/10.1016/j.agrformet.2015.07.007 899 Folberth GA, Hauglustaine DA, Lathière J, Brocheton F. 2006. Interactive chemistry in the 900 Laboratoire de Météorologie Dynamique general circulation model: model description and 901 impact analysis of biogenic hydrocarbons on tropospheric chemistry. Atmospheric Chem 902 Phys. 6(8):2273–2319 903 Franklin O et al. 2023. Optimal balancing of xylem efficiency and safety explains plant 904 vulnerability to drought. Ecol Lett. 26(9):1485–1496. https://doi.org/10.1111/ele.14270 905 Frey Y, Dumberger S, et al. 2025a. The ECOvette: Unveiling Plant Dynamics Through Direct 906 Leaf Emission Analysis. ACS EST Eng. 5(11):3096–3106. 907 https://doi.org/10.1021/acsestengg.5c00462 908 Frey Y, Simon L, et al. 2025b. Sensor evaluation for leaf temperature within a minimally 909 invasive leaf cuvette. Sens Actuators Phys. 396:117045. 910 https://doi.org/10.1016/j.sna.2025.117045 911 Friedlingstein P et al. 2025. Global Carbon Budget 2025. Earth Syst Sci Data Discuss. 1–139. 912 https://doi.org/10.5194/essd-2025-659 913 Gharun M et al. 2024. Effect of the 2022 summer drought across forest types in Europe. 914 Biogeosciences. 21(23):5481–5494. https://doi.org/10.5194/bg-21-5481-2024 915 Ghirardo A et al. 2010. Determination of de novo and pool emissions of terpenes from four 916 common boreal/alpine trees by 13CO2 labelling and PTR-MS analysis. Plant Cell Environ. 917 33(5):781–792. https://doi.org/10.1111/j.1365-3040.2009.02104.x 918 Gillerot L et al. 2021. Tree Neighbourhood Diversity Has Negligible Effects on Drought 919 Resilience of European Beech, Silver Fir and Norway Spruce. Ecosystems. 24(1):20–36. 920 https://doi.org/10.1007/s10021-020-00501-y 921 Grossiord C. 2020. Having the right neighbors: how tree species diversity modulates drought 922 impacts on forests. New Phytol. 228(1):42–49. https://doi.org/10.1111/nph.15667 923 Grossiord C et al. 2020. Plant responses to rising vapor pressure deficit. New Phytol. 924 226(6):1550–1566 925 Grote R, Monson RK, Niinemets Ü. 2013. Leaf-Level Models of Constitutive and Stress-926 Driven Volatile Organic Compound Emissions. In: Niinemets Ü, Monson RK, editors. 927 Biology, Controls and Models of Tree Volatile Organic Compound Emissions. Springer 928 Netherlands; p 315–355 [accessed 2025 Dec 17]. https://doi.org/10.1007/978-94-007-6606-929 8_12. https://doi.org/10.1007/978-94-007-6606-8_12 930 Gu L et al. 2002. Advantages of diffuse radiation for terrestrial ecosystem productivity. J 931 Geophys Res Atmospheres. 107(D6):ACL 2-1. https://doi.org/10.1029/2001JD001242 932 Guenther AB et al. 2012. The Model of Emissions of Gases and Aerosols from Nature version 933 2.1 (MEGAN2.1): an extended and updated framework for modeling biogenic emissions. 934 Geosci Model Dev. 5(6):1471–1492. https://doi.org/10.5194/gmd-5-1471-2012 935 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 35 Haberstroh S et al. 2018. Terpenoid Emissions of Two Mediterranean Woody Species in 936 Response to Drought Stress. Front Plant Sci. 9 [accessed 2026 Jan 8]. 937 https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2018.01071/full. 938 https://doi.org/10.3389/fpls.2018.01071 939 Haberstroh S et al. 2019. Natural Carbon Isotope Composition Distinguishes Compound 940 Groups of Biogenic Volatile Organic Compounds (BVOC) in Two Mediterranean Woody 941 Species. Front For Glob Change. 2 [accessed 2026 Jan 8]. 942 https://www.frontiersin.org/journals/forests-and-global-943 change/articles/10.3389/ffgc.2019.00055/full. https://doi.org/10.3389/ffgc.2019.00055 944 Haberstroh S et al. 2022. Central European 2018 hot drought shifts scots pine forest to its 945 tipping point. Plant Biol. 24(7):1186–1197. https://doi.org/10.1111/plb.13455 946 Haberstroh S et al. 2025. Recurrent hot droughts cause persistent legacy effects in a temperate 947 Scots Pine forest. Plant Biol. plb.70066. https://doi.org/10.1111/plb.70066 948 Hakola H et al. 2006. Seasonal variation of mono- and sesquiterpene emission rates of Scots 949 pine. Biogeosciences. 3(1):93–101. https://doi.org/10.5194/bg-3-93-2006 950 He L et al. 2018. Changes in the Shadow: The Shifting Role of Shaded Leaves in Global 951 Carbon and Water Cycles Under Climate Change. Geophys Res Lett. 45(10):5052–5061. 952 https://doi.org/10.1029/2018GL077560 953 Henry C et al. 2019. A stomatal safety-efficiency trade-off constrains responses to leaf 954 dehydration. Nat Commun. 10(1):3398. https://doi.org/10.1038/s41467-019-11006-1 955 Holzke C et al. 2006. Terpene emissions from European beech (shape Fagus sylvatica~L.): 956 Pattern and Emission Behaviour Over two Vegetation Periods. J Atmospheric Chem. 957 55(1):81–102. https://doi.org/10.1007/s10874-006-9027-9 958 Joó É et al. 2011. Constitutive versus heat and biotic stress induced BVOC emissions in 959 Pseudotsuga menziesii. Atmos Environ. 45(22):3655–3662. 960 https://doi.org/10.1016/j.atmosenv.2011.04.048 961 Keenan T et al. 2009. Seasonality of monoterpene emission potentials in Quercus ilex and 962 Pinus pinea: Implications for regional VOC emissions modeling. J Geophys Res 963 Atmospheres. 114(D22) [accessed 2025 Dec 22]. 964 https://onlinelibrary.wiley.com/doi/abs/10.1029/2009JD011904. 965 https://doi.org/10.1029/2009JD011904 966 Kesselmeier J et al. 1997. Emission of short chained organic acids, aldehydes and 967 monoterpenes from Quercus ilex L. and Pinus pinea L. in relation to physiological activities, 968 carbon budget and emission algorithms. Atmos Environ. 31:119–133 (BEMA: A European 969 Commision Project on Biogenic Emmissions in the Mediterranean Area). 970 https://doi.org/10.1016/S1352-2310(97)00079-4 971 Kinzinger L et al. 2024. Interaction between beech and spruce trees in temperate forests 972 affects water use, root water uptake pattern and canopy structure Meinzer F, editor. Tree 973 Physiol. 44(1):tpad144. https://doi.org/10.1093/treephys/tpad144 974 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 36 Knohl A, Baldocchi DD. 2008. Effects of diffuse radiation on canopy gas exchange processes 975 in a forest ecosystem. J Geophys Res Biogeosciences. 113(G2) [accessed 2025 Dec 17]. 976 https://onlinelibrary.wiley.com/doi/abs/10.1029/2007JG000663. 977 https://doi.org/10.1029/2007JG000663 978 Knutzen F et al. 2025. Impacts on and damage to European forests from the 2018–2022 heat 979 and drought events. Nat Hazards Earth Syst Sci. 25(1):77–117. https://doi.org/10.5194/nhess-980 25-77-2025 981 Köcher P, Gebauer T, Horna V, Leuschner C. 2009. Leaf water status and stem xylem flux in 982 relation to soil drought in five temperate broad-leaved tree species with contrasting water use 983 strategies. Ann For Sci. 66(1):101–101. https://doi.org/10.1051/forest/2008076 984 Köcher P, Horna V, Leuschner C. 2013. Stem water storage in five coexisting temperate 985 broad-leaved tree species: significance, temporal dynamics and dependence on tree functional 986 traits. Tree Physiol. 33(8):817–832. https://doi.org/10.1093/treephys/tpt055 987 Kolari P et al. 2009. CO2 exchange and component CO2 fluxes of a boreal Scots pine forest. 988 Boreal Environ Res. 14(4):761 989 Kolari P, Lappalainen HK, Hänninen H, Hari P. 2007. Relationship between temperature and 990 the seasonal course of photosynthesis in Scots pine at northern timberline and in southern 991 boreal zone. Tellus B. 59(3):542–552. https://doi.org/10.1111/j.1600-0889.2007.00262.x 992 Kreuzwieser J et al. 2021a. Drought affects carbon partitioning into volatile organic 993 compound biosynthesis in Scots pine needles. New Phytol. 232(5):1930–1943. 994 https://doi.org/10.1111/nph.17736 995 Kreuzwieser J et al. 2021b. Drought affects carbon partitioning into volatile organic 996 compound biosynthesis in Scots pine needles. New Phytol. 232(5):1930–1943. 997 https://doi.org/10.1111/nph.17736 998 Laffineur Q et al. 2013. Impact of diffuse light on isoprene and monoterpene emissions from 999 a mixed temperate forest. Atmos Environ. 74:385–392. 1000 https://doi.org/10.1016/j.atmosenv.2013.04.025 1001 Laothawornkitkul J, Taylor JE, Paul ND, Hewitt CN. 2009. Biogenic volatile organic 1002 compounds in the Earth system. New Phytol. 183(1):27–51. https://doi.org/10.1111/j.1469-1003 8137.2009.02859.x 1004 Lenth RV. 2023. emmeans: Estimated Marginal Means, aka Least-Squares Means. 1005 https://CRAN.R-project.org/package=emmeans 1006 Lerdau M, Litvak M, Palmer P, Monson R. 1997. Controls over monoterpene emissions from 1007 boreal forest conifers. Tree Physiol. 17(8–9):563–569. https://doi.org/10.1093/treephys/17.8-1008 9.563 1009 Lerdau M, Matson P, Fall R, Monson R. 1995. Ecological Controls over Monoterpene 1010 Emissions from Douglas-Fir (Pseudotsuga Menziesii). Ecology. 76(8):2640–2647. 1011 https://doi.org/10.2307/2265834 1012 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 37 Leuschner C et al. 2024. A multi-criteria drought resistance assessment of temperate Acer, 1013 Carpinus, Fraxinus, Quercus, and Tilia species. Perspect Plant Ecol Evol Syst. 62:125777. 1014 https://doi.org/10.1016/j.ppees.2023.125777 1015 Lüdecke D et al. 2021. performance: An R Package for Assessment, Comparison and Testing 1016 of Statistical Models. J Open Source Softw. 6(60):3139. https://doi.org/10.21105/joss.03139 1017 Martínez-Vilalta J et al. 2014. A new look at water transport regulation in plants. New Phytol. 1018 204(1):105–115. https://doi.org/10.1111/nph.12912 1019 Meir P et al. 2002. Acclimation of photosynthetic capacity to irradiance in tree canopies in 1020 relation to leaf nitrogen concentration and leaf mass per unit area. Plant Cell Environ. 1021 25(3):343–357. https://doi.org/10.1046/j.0016-8025.2001.00811.x 1022 Meischner M et al. 2024. Jasmonic acid and heat stress induce high volatile organic 1023 compound emissions in Picea abies from needles, but not from roots. Tree Physiol. tpae059. 1024 https://doi.org/10.1093/treephys/tpae059 1025 Miyashita A et al. 2012. Long-term, short-interval measurements of the frequency 1026 distributions of the photosynthetically active photon flux density and net assimilation rate of 1027 leaves in a cool-temperate forest. Agric For Meteorol. 152:1–10. 1028 https://doi.org/10.1016/j.agrformet.2011.08.001 1029 Niinemets Ü, Hauff K, et al. 2002a. Monoterpene emissions in relation to foliar 1030 photosynthetic and structural variables in Mediterranean evergreen Quercus species. New 1031 Phytol. 153(2):243–256. https://doi.org/10.1046/j.0028-646X.2001.00323.x 1032 Niinemets Ü. 2023. Variation in leaf photosynthetic capacity within plant canopies: 1033 optimization, structural, and physiological constraints and inefficiencies. Photosynth Res. 1034 158(2):131–149. https://doi.org/10.1007/s11120-023-01043-9 1035 Niinemets Ü, Copolovici L, Hüve K. 2010. High within-canopy variation in isoprene 1036 emission potentials in temperate trees: Implications for predicting canopy-scale isoprene 1037 fluxes. J Geophys Res Biogeosciences. 115(G4) [accessed 2025 Dec 23]. 1038 https://onlinelibrary.wiley.com/doi/abs/10.1029/2010JG001436. 1039 https://doi.org/10.1029/2010JG001436 1040 Niinemets Ülo, Loreto F, Reichstein M. 2004a. Physiological and physicochemical controls 1041 on foliar volatile organic compound emissions. Trends Plant Sci. 9(4):180–186. 1042 https://doi.org/10.1016/j.tplants.2004.02.006 1043 Niinemets Ü, Seufert G, Steinbrecher R, Tenhunen JD. 2002b. A model coupling foliar 1044 monoterpene emissions to leaf photosynthetic characteristics in Mediterranean evergreen 1045 Quercus species. New Phytol. 153(2):257–275. https://doi.org/10.1046/j.0028-1046 646X.2001.00324.x 1047 Niinemets Ü., Sonninen E, Tobias M. 2004b. Canopy gradients in leaf intercellular CO2 mole 1048 fractions revisited: interactions between leaf irradiance and water stress need consideration. 1049 Plant Cell Environ. 27(5):569–583. https://doi.org/10.1111/j.1365-3040.2003.01168.x 1050 Novick K et al. 2016. The increasing importance of atmospheric demand for ecosystem water 1051 and carbon fluxes. Nat Clim Change. 6. https://doi.org/10.1038/nclimate3114 1052 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 38 Oren R et al. 1999. Survey and synthesis of intra- and interspecific variation in stomatal 1053 sensitivity to vapour pressure deficit. Plant Cell Environ. 22(12):1515–1526. 1054 https://doi.org/10.1046/j.1365-3040.1999.00513.x 1055 Paligi SS et al. 2025. Water consumption of beech, spruce and Douglas fir in pure and mixed 1056 stands in a wet and a dry year – Testing predictions of the iso/anisohydry concept. Sci Total 1057 Environ. 970:178948. https://doi.org/10.1016/j.scitotenv.2025.178948 1058 Pan Y et al. 2024. The enduring world forest carbon sink. Nature. 631(8021):563–569. 1059 https://doi.org/10.1038/s41586-024-07602-x 1060 Peltoniemi MS, Duursma RA, Medlyn BE. 2012. Co-optimal distribution of leaf nitrogen and 1061 hydraulic conductance in plant canopies. Tree Physiol. 32(5):510–519. 1062 https://doi.org/10.1093/treephys/tps023 1063 Peñuelas J, Llusià J. 2003. BVOCs: plant defense against climate warming? Trends Plant Sci. 1064 8(3):105–109. https://doi.org/10.1016/S1360-1385(03)00008-6 1065 Peters RL et al. 2023. Daytime stomatal regulation in mature temperate trees prioritizes stem 1066 rehydration at night. New Phytol. 239(2):533–546. https://doi.org/10.1111/nph.18964 1067 Peters RL et al. 2025. Uniform regulation of stomatal closure across temperate tree species to 1068 sustain nocturnal turgor and growth. Nat Plants. 11(4):725–730. 1069 https://doi.org/10.1038/s41477-025-01957-3 1070 Phillips NG et al. 2003. Reliance on stored water increases with tree size in three species in 1071 the Pacific Northwest. Tree Physiol. 23(4):237–245. 1072 https://doi.org/10.1093/treephys/23.4.237 1073 Plaza J et al. 2005. Field monoterpene emission of Mediterranean oak (Quercus ilex) in the 1074 central Iberian Peninsula measured by enclosure and micrometeorological techniques: 1075 Observation of drought stress effect. J Geophys Res Atmospheres. 110(D3) [accessed 2025 1076 Dec 23]. https://onlinelibrary.wiley.com/doi/abs/10.1029/2004JD005168. 1077 https://doi.org/10.1029/2004JD005168 1078 Pohl F et al. 2023. Observational evidence of legacy effects of the 2018 drought on a mixed 1079 deciduous forest in Germany. Sci Rep. 13(1):10863. https://doi.org/10.1038/s41598-023-1080 38087-9 1081 Pressley S et al. 2004. Monoterpene emissions from a Pacific Northwest Old-Growth Forest 1082 and impact on regional biogenic VOC emission estimates. Atmos Environ. 38(19):3089–1083 3098. https://doi.org/10.1016/j.atmosenv.2004.01.045 1084 R Core Team. 2025. R: A Language and Environment for Statistical Computing. R 1085 Foundation for Statistical Computing. https://www.R-project.org/ 1086 Richter R et al. 2022. Tree species matter for forest microclimate regulation during the 1087 drought year 2018: disentangling environmental drivers and biotic drivers. Sci Rep. 1088 12(1):17559. https://doi.org/10.1038/s41598-022-22582-6 1089 Roderick ML, Farquhar GD, Berry SL, Noble IR. 2001. On the direct effect of clouds and 1090 atmospheric particles on the productivity and structure of vegetation. Oecologia. 129(1):21–1091 30. https://doi.org/10.1007/s004420100760 1092 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 39 Salmon Y et al. 2020. Leaf carbon and water status control stomatal and nonstomatal 1093

of photosynthesis in trees. New Phytol. 226(3):690–703. 1094 https://doi.org/10.1111/nph.16436 1095 Salomón RL, Camarero JJ. 2025. Stem Growth and Dehydration Responses of Mediterranean 1096 Tree Species to Atmospheric and Soil Drought. Plant Cell Environ. 48(1):866–881. 1097 https://doi.org/10.1111/pce.15177 1098 Scapucci L et al. 2024. Compound soil and atmospheric drought (CSAD) events and CO2 1099 fluxes of a mixed deciduous forest: the occurrence, impact, and temporal contribution of main 1100 drivers. Biogeosciences. 21(15):3571–3592. https://doi.org/10.5194/bg-21-3571-2024 1101 Scartazza A et al. 2016. Investigating the European beech (Fagus sylvatica L.) leaf 1102 characteristics along the vertical canopy profile: leaf structure, photosynthetic capacity, light 1103 energy dissipation and photoprotection mechanisms. Tree Physiol. 36(9):1060–1076. 1104 https://doi.org/10.1093/treephys/tpw038 1105 Scholander PF, Bradstreet ED, Hemmingsen EA, Hammel HT. 1965. Sap Pressure in 1106 Vascular Plants. Science. 148(3668):339–346. https://doi.org/10.1126/science.148.3668.339 1107 Schuldt B et al. 2020. A first assessment of the impact of the extreme 2018 summer drought 1108 on Central European forests. Basic Appl Ecol. 45. https://doi.org/10.1016/j.baae.2020.04.003 1109 Schumann K et al. 2024. Xylem safety in relation to the stringency of plant water potential 1110 regulation of European beech, Norway spruce, and Douglas-fir trees during severe drought. 1111 Trees. 38(3):607–623. https://doi.org/10.1007/s00468-024-02499-5 1112 Šimpraga M et al. 2011. Comparing monoterpenoid emissions and net photosynthesis of 1113 beech (Fagus sylvatica L.) in controlled and natural conditions. Atmos Environ. 45(17):2922–1114 2928. https://doi.org/10.1016/j.atmosenv.2011.01.047 1115 Sprintsin M, Chen JM, Desai A, Gough CM. 2012. Evaluation of leaf‐to‐canopy upscaling 1116 methodologies against carbon flux data in North America. J Geophys Res Biogeosciences. 1117 117(G1):2010JG001407. https://doi.org/10.1029/2010JG001407 1118 Staudt M et al. 1997. Seasonal and diurnal patterns of monoterpene emissions from Pinus 1119 pinea (L.) under field conditions. Atmos Environ. 31:145–156 (BEMA: A European 1120 Commision Project on Biogenic Emmissions in the Mediterranean Area). 1121 https://doi.org/10.1016/S1352-2310(97)00081-2 1122 Staudt M, Rambal S, Joffre R, Kesselmeier J. 2002. Impact of drought on seasonal 1123 monoterpene emissions from Quercus ilex in southern France. J Geophys Res Atmospheres. 1124 107(D21):ACH 15-1-ACH 15-9. https://doi.org/10.1029/2001JD002043 1125 Stenberg P. 1998. Implications of shoot structure on the rate of photosynthesis at different 1126 levels in a coniferous canopy using a model incorporating grouping and penumbra. Funct 1127 Ecol. 12(1):82–91. https://doi.org/10.1046/j.1365-2435.1998.00169.x 1128 Tesch J et al. 2025. The ECOSENSE forest: A distributed sensor and data management 1129 system for real-time monitoring of ecosystem processes and stresses. EGUsphere. 1–29. 1130 https://doi.org/10.5194/egusphere-2025-4979 1131 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 40 Urban O et al. 2007. Ecophysiological controls over the net ecosystem exchange of mountain 1132 spruce stand. Comparison of the response in direct vs. diffuse solar radiation. Glob Change 1133 Biol. 13(1):157–168. https://doi.org/10.1111/j.1365-2486.2006.01265.x 1134 Valladares F, Laanisto L, Niinemets Ü, Zavala MA. 2016. Shedding light on shade: 1135 ecological perspectives of understorey plant life. Plant Ecol Divers. 9(3):237–251. 1136 https://doi.org/10.1080/17550874.2016.1210262 1137 Van Meeningen Y, Schurgers G, Rinnan R, Holst T. 2016. BVOC emissions from English 1138 oak ( Quercus robur ) and European beech ( Fagus sylvatica ) along a latitudinalgradient. 1139 Biogeosciences. 13(21):6067–6080. https://doi.org/10.5194/bg-13-6067-2016 1140 Wang J, Sammis TW. 2008. New inexpensive dendrometers for monitoring crop tree growth. 1141 Wang M et al. 2025. Influence of forest canopy structure on temperature buffering in young 1142 planted forests with varied tree species compositions revealed by terrestrial laser scanning. 1143 Agric For Meteorol. 371:110640. https://doi.org/10.1016/j.agrformet.2025.110640 1144 Way DA, Pearcy RW. 2012. Sunflecks in trees and forests: from photosynthetic physiology to 1145 global change biology. Tree Physiol. 32(9):1066–1081. 1146 https://doi.org/10.1093/treephys/tps064 1147 Werner C et al. 2020. Heat Waves Change Plant Carbon Allocation Among Primary and 1148 Secondary Metabolism Altering CO2 Assimilation, Respiration, and VOC Emissions. Front 1149 Plant Sci. 11 [accessed 2026 Jan 20]. https://www.frontiersin.org/journals/plant-1150 science/articles/10.3389/fpls.2020.01242/full. https://doi.org/10.3389/fpls.2020.01242 1151 Werner C et al. 2021. Ecosystem fluxes during drought and recovery in an experimental 1152 forest. Science. 374(6574):1514–1518. https://doi.org/10.1126/science.abj6789 1153 Werner C et al. 2024. ECOSENSE - Multi-scale quantification and modelling of spatio-1154 temporal dynamics of ecosystem processes by smart autonomous sensor networks. Res Ideas 1155 Outcomes. 10:e129357. https://doi.org/10.3897/rio.10.e129357 1156 Werner C et al. 2025. Impact of emerging compound droughts on forests: A water supply and 1157 demand perspective. Plant Biol. n/a(n/a) [accessed 2025 Dec 21]. 1158 https://onlinelibrary.wiley.com/doi/abs/10.1111/plb.70080. https://doi.org/10.1111/plb.70080 1159 WMO. 2018. Guide to Instruments and Methods of Observation. 8th edn World 1160 Meteorological Organization. 1161 Wood SN. 2017. Generalized Additive Models: An Introduction with R, Second Edition. 2nd 1162 edn Chapman and Hall/CRC. https://doi.org/10.1201/9781315370279 1163 Wozniak MC, Bonan GB, Keppel-Aleks G, Steiner AL. 2020. Influence of Vertical 1164 Heterogeneities in the Canopy Microenvironment on Interannual Variability of Carbon 1165 Uptake in Temperate Deciduous Forests. J Geophys Res Biogeosciences. 1166 125(8):e2020JG005658. https://doi.org/10.1029/2020JG005658 1167 Wu C et al. 2015. Impacts of soil moisture on de novo monoterpene emissions from European 1168 beech, Holm oak, Scots pine, and Norway spruce. Biogeosciences. 12(1):177–191. 1169 https://doi.org/10.5194/bg-12-177-2015 1170 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint 41 Zellweger F et al. 2019. Seasonal drivers of understorey temperature buffering in temperate 1171 deciduous forests across Europe. Glob Ecol Biogeogr. 28(12):1774–1786. 1172 https://doi.org/10.1111/geb.12991 1173 Zellweger F et al. 2020. Forest microclimate dynamics drive plant responses to warming. 1174 Science. 368(6492):772–775. https://doi.org/10.1126/science.aba6880 1175 Zhang S, Landuyt D, Verheyen K, De Frenne P. 2022. Tree species mixing can amplify 1176 microclimate offsets in young forest plantations. J Appl Ecol. 59(6):1428–1439. 1177 https://doi.org/10.1111/1365-2664.14158 1178 Zhou H et al. 2021. Large Contributions of Diffuse Radiation to Global Gross Primary 1179 Productivity During 1981–2015. Glob Biogeochem Cycles. 35(8):e2021GB006957. 1180 https://doi.org/10.1029/2021GB006957 1181 Zweifel R et al. 2021. Why trees grow at night. New Phytol. 231(6):2174–2185. 1182 https://doi.org/10.1111/nph.17552 1183 1184 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted January 23, 2026. ; https://doi.org/10.64898/2026.01.23.701264doi: bioRxiv preprint