Imprints of evaporation and vegetation type in diurnal 1 temperature variations 2

Annu Panwar, Maik Renner, Axel Kleidon 3 4 Biospheric Theory and Modeling group, Max Planck Institute for Biogeochemistry, Jena, 07745, 5 Germany 6 7 Correspondence to: Annu Panwar (apanwar@bgc-jena.mpg.de) 8 9 Abstract. Diurnal temperature variations are strongly shaped by the absorption of solar radiation, but 10 evaporation, or the latent heat flux, also plays an important role. Generally, evaporation cools. Its 11 relation to diurnal temperature variations, however, is unclear. This study investigates the diurnal 12 response of surface and air temperatures to evaporation for different vegetation types. We used the 13 warming rate of temperature to absorbed solar radiation in the morning under clear-sky conditions and 14 evaluated how the warming rates change for different evaporative fractions. Results for 51 FLUXNET 15 sites show that the diurnal variation of air temperature carries very weak imprints of evaporation across 16 all vegetation types. However, surface temperature warming rates of short vegetation decrease 17 significantly by ~23 x 10 K/W m from dry to wet conditions. Contrarily, warming rates of surface 18 and air temperatures are similar at forest sites and carry literally no imprints of evaporation. We 19 explain these contrasting patterns with a surface energy balance model. The model reveals a strong 20 sensitivity of the warming rates to evaporative fraction and aerodynamic conductance. However, for 21 forests the sensitivity to evaporative fraction is strongly reduced by 74 % due to their large 22 aerodynamic conductance. The remaining imprint is reduced further by ~ 50% through their enhanced 23 aerodynamic conductance under dry conditions. Our model then compares the individual contributions 24 of solar radiation, evaporation and vegetation types in shaping the diurnal temperature range. These 25 findings have implications for the interpretation of land-atmosphere interactions and the influences of 26 water limitation and vegetation on diurnal temperatures, which is of key importance for ecological 27 functioning. We conclude that diurnal temperature variations may be useful to predict evaporation for 28 short vegetation. In forests, however, the diurnal variations in temperatures are mainly governed by 29 their aerodynamic properties resulting in no imprint of evaporation in diurnal temperature variations. 30 31

aerodynamic conductance. The remaining imprint is reduced further by ~ 50% through their enhanced 23 aerodynamic conductance under dry conditions. Our model then compares the individual contributions 24 of solar radiation, evaporation and vegetation types in shaping the diurnal temperature range. These 25 findings have implications for the interpretation of land-atmosphere interactions and the influences of 26 water limitation and vegetation on diurnal temperatures, which is of key importance for ecological 27 functioning. We conclude that diurnal temperature variations may be useful to predict evaporation for 28 short vegetation. In forests, however, the diurnal variations in temperatures are mainly governed by 29 their aerodynamic properties resulting in no imprint of evaporation in diurnal temperature variations. 30 31 1 Introduction 32 Temperature is one of the most widely monitored variables in meteorology. Besides being important 33 for our day-to-day activities, temperature serves as a primary attribute for the understanding of Earth 34 system processes. The diurnal variation of temperature is considered informative in climate science, as 35 described by the diurnal temperature range (DTR), which is basically the difference between daily 36 maximum and minimum temperatures. Information on the diurnal temperature range has facilitated a 37 broad spectrum of research including agriculture, health welfare, climate change and ecological 38 studies. 39 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License.
accounted for by what we refer to as the warming rate, the increase in temperature due to a unit 78 increase in the absorbed solar radiation, expressed as the derivative dT a /dR s for air temperature and 79 dT s /dR s for surface temperature with units of K/W m -2 . One can approximate the warming rate by the 80 ratio of DTR to maximum solar radiation, so that the warming rate can be seen as an efficient 81 characteristic that captures effects on DTR that are not caused by solar radiation. In this study, we use 82 linear regressions of observed data from the morning to noon to calculate warming rates. Certainly, it is intriguing to find out how evaporation alters this coupling. In our earlier work (Panwar 100 et al., 2019) we looked at the temperature warming rate for a cropland site in the Southern Great Plains. 101 We observed that the warming rate of surface temperature decreases from dry (less-evaporative) to wet 102 (evaporative) conditions but the warming rate of air temperature remained unaffected by evaporation. 103 Combining the boundary layer information and heat budget expression we explained that the diurnal 104 variation of air temperature does not contain the imprints of evaporation due to the compensating role 105 of boundary layer development. If this is a general finding, then the surface temperature warming rate 106 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License.
can be used for estimating evaporation of short vegetation. However, it is also interesting to see how 107 the evaporative cooling effect competes with the cooling effect by a higher aerodynamic conductance. 108

109
In this study, we approach two major questions to advance our understanding of diurnal temperature 110 variations: a) Do the diurnal variations of surface and air temperature respond to evaporation? and b) 111 What is the role of the different aerodynamic conductance of vegetation in altering these responses? 112 Our previous work (Panwar et al., 2019) already shows the stronger imprints of evaporation in diurnal 113 surface temperature variations. Here we examine the generality of this finding in short vegetation. 114 Additionally, to understand the role of aerodynamic conductance in modifying these imprints we 115 analyze data from the taller and more complex vegetation like savanna and forests. 116

117
We first present a model based on the surface energy balance to provide an expression for the surface 118 temperature warming rate and its response to evaporation and aerodynamic conductance (all variables 119 used are summarized in Table A1). To evaluate our model, we used observations from 51 FLUXNET 120 sites that include short vegetation, savanna and forests. Surface and air temperature warming rates, 121 aerodynamic conductances and their response to evaporation are quantified for each site. We then use 122 these findings with our model to explain and reproduce observed temperature warming rates and their 123 response to evaporation. The cooling effect of evaporation and its relation to aerodynamic conductance 124 is quantified for each vegetation type. Combining the warming rates with the information on solar 125 radiation, we conclude the study by demonstrating the contribution of solar radiation, evaporation and 126 aerodynamic conductance in shaping the DTR using our observational analysis and model. 127 128 2 Modeling temperature-warming rate 129 Surface and air temperatures possess a strong diurnal variation that is driven by the absorbtion of solar 130 radiation. The amplitude of this variation is also affected by other components of surface energy 131 balance, among which the partitioning of turbulent heat fluxes into latent and sensible heat is 132 important. Generally, the surface energy balance is written as 133 Here, ! is the absorbed solar radiation at the surface, !,!"# is the net longwave radiation, is the 137 latent heat flux (with L being the latent heat of vaporization and E the evaporation rate), is the 138 sensible heat flux and is the ground heat flux. For simplification of the surface energy balance we 139 linearize !,!"# using the first order terms, such that !,!"# = ! + ! ( ! − !"# ). Here, ! is the net 140 radiation at a reference temperature !"# . The second term, ! = 4 !"# ! is the linearization constant. 141 Incorporating this simplification of !,!"# in Eq. (1), the surface energy balance can be rearranged to 142 yield an expression for ! , 143 144 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License.
The warming rate of surface temperature is obtained by taking the derivative of Eq. 2 with respect to 146 absorbed solar radiation, R s . The warming rate of surface temperature is given by 147 Since, ! and !"# are assumed to be constants and do not vary diurnally with ! , they disappear in 150 Eq. (3). Additionally it is assumed that the diurnal change in G in response to ! is negligible 151 ( / ! ~0) compared to other components of surface energy balance. This assumption is valid since 152 we are considering vegetated sites for our study, although we are aware that for non-vegetated surfaces 153 G can represent a noticeable share of absorbed solar radiation (Clothier et al., 1986;Kustas and 154 Daughtry, 1990). 155

156
We describe the evaporative conditions by the evaporative fraction ( ! ), the ratio of the latent heat 157 flux ( ) to the total turbulent heat fluxes ( + ). Given this, the term + in Eq.
(3) can be 158 written as (1 − ! ) . Furthermore, the sensible heat flux can be expressed in terms of the 159 aerodynamic conductance as = ! ! ( ! − ! ), where ! =1005 J/kg K is the specific heat capacity 160 of air, = 1.23 kg m -3 is air density and ! is the aerodynamic conductance. On including these 161 replacements in Eq. (3) we get an approximation for the surface temperature warming rate 162 where ! / ! is the air temperature warming rate. We can further simplify this expression by 165 considering the two terms in the denominator of Eq. (4). Considering !"# ~288 K, the term 166 ! 1 − ! varies by ~4.87 Wm -2 K -1 to ~0.54 Wm -2 K -1 from dry ( !~0 .1) to wet ( !~0 .9) conditions, 167 which is much smaller in magnitude compared to the term ! ! that is ~60 Wm -2 K -1 for a typical 168 cropland site ( ! =0.05 m s -1 ) and 250 W m -2 K -1 for a typical forest site ( ! =0.2 m s -1 ). Because of 169 these magnitudes, the term ! 1 − ! can be neglected. This leads to a further simplification of the 170 warming rate to 171 Eq. (5) shows that morning to noon warming of surface temperature is a function of evaporative 174 fraction, aerodynamic conductance and also of the warming rate of air temperature. 175 176 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License.
Finally, the sensitivity of the warming rate to changes in evaporative conditions is obtained by taking 177 the derivative of Eq. (5) with respect to evaporative fraction ( ! ). To express these derivatives with 178 respect to evaporative fraction, we use the apostrophe ( ! = ′). Therefore, ! ! / ! and 179 ! ! / ! represent the change in surface and air temperature warming rates due to a unit change in 180 evaporative fraction. Similarly, ! ! is the change in aerodynamic conductance from dry to wet 181 evaporative conditions. We obtain: 182 184 This model provides two important expressions that we test with observations. The first expression is 185 the warming rate of surface temperature, described by Eq. (5), which requires the information of the 186 warming rate of air temperature, aerodynamic conductance and evaporative fraction. On multiplying 187 these two equations with daily maximum solar radiation shall provide an approximation of DTR that 188 can also be validated with the observational data. The second expression is the response of the surface 189 temperature warming rate to evaporation, shown in Eq. (6), which is a negative quantity provided 190 ! ! / ! is small (or negative). The negative sign means that the surface temperature warming rate 191 decreases with increase in evaporative fraction. The amplitude of this decrease mainly depends on the 192 characteristic aerodynamic conductance ( ! ) of vegetation (the first term on the right hand side of Eq. 193 6) and also on its relative sensitivity to evaporative fraction ( ! ! / ! , the second term on the right hand 194 side). of cloud covers that result in reduced mean incoming solar radiation. An additional filter to remove 209 cloudy days is applied that is based on the quantile regression method using surface solar radiation and 210 potential solar radiation (Renner et al., 2019). This method was applied only from morning to noon, so 211 that if the day has clouds in the evening, it is still considered as a clear sky day. This does not influence 212 warming rates since they are calculated only from the morning to noontime variation of temperature. The vegetation type of each site is classified using the International Geosphere-Biosphere Programme 220 (IGBP) Data and Information System (Loveland and Belward, 1997). The IGBP land cover product is 221 available at 1 km resolution and was derived from the Advance Very High Resolution Radiometer 222 (AVHRR). Detailed information of each site with their location, number of days used in the analysis, 223 land use type and references is provided in the Appendix (Table A2). Vegetation are classified into 224 three types that is based on their typical vegetation height and coverage, see Table 1. Shorter vegetation 225 like croplands, grasslands, and shrublands are grouped into the 'short vegetation' type. Savanna 226 ecosystems are complex with heterogeneous vegetation height, which basically delineates the transition 227 of short vegetation to forests, and are grouped into the 'savanna' type. All forest types, including 228 deciduous broadleaf, evergreen broadleaf, evergreen needleleaf and mixed, are grouped in the 'forest' 229 type. 230

231
The geographic location of the selected 51 sites is shown in Figure 2. The color bar represents the 232 mean annual evaporative fraction derived from FLUXCOM data (Jung et al., 2019;Tramontana et al., 233 2016). Selected sites represent a wide range of ecosystems that is ideal for studying the generality of 234 the response of warming rates to differences in evaporative conditions and vegetation type.  Evergreen broadleaf forest 1 Evergreen needle leaf forest 9 Mixed forest 5

239
The evaporative condition is quantified by evaporative fraction. One of the advantages of evaporative 240 fraction is its stability for daylight hours such that it can be assumed to be constant over a day 241 (Shuttleworth et al., 1989). Daily evaporative fraction is obtained by the linear regression of half hourly 242 morning to noon values of the ratio of the latent heat flux to the total turbulent heat fluxes. Similarly, a 243 linear regression of half hourly warming rate and evaporative fraction values is used to quantify the 244 response of the warming rate to evaporative fraction. 245

246
We use the term air temperature for the temperature measured above the canopy. Surface temperature 247 is calculated from the upwelling longwave radiation using the Stefan-Boltzmann law, such that the 248 surface temperature is the skin temperature of the vegetation. The aerodynamic conductance ( ! ) is 249 obtained from the observed frictional velocity ( * ) and wind speed ( ) by ! = * ! (see, e.g., 250 Verma (1989)). For simplicity, the conductance of heat fluxes and momentum are assumed to be 251 identical (Mallick et al., 2016). The primary advantage of warming rate over DTR is its suitability to compare sites with different solar 257 energy input. This is apparent from Figure 3, where we show the probability density distribution of the 258 observed daily warming rates of surface (a) and air temperatures (b) for short vegetation, savanna, and 259 forest. We look at the surface and air temperature warming rates to determine if they carry any 260 information on vegetation type. In general, the surface temperature warming rate of short vegetation is 261 larger by almost a factor of two compared to the surface temperature warming rate of forests. Savanna 262 covers the range in surface temperature warming rates, reflecting their characteristics being positioned 263 between short vegetation and forests. Hence, the vegetation type clearly affects the surface temperature 264 warming rate. Surprisingly, this is not true for air temperature warming rates. Short vegetation, savanna 265 and forests show similar distributions of air temperature warming rate. The air temperature warming 266 rate of short vegetation is smaller than its surface temperature warming rate. Conversely, in forests, the 267 magnitudes are similar, indicating the strong coupling between surface and air temperature. Site-specific information on warming rates is provided in Figure A1  rates, such that it is higher for dry and lower for wet sites. On the other hand, despite these differences 281 in the mean, the sites contain days with a good range of evaporative fractions (see Figure A2 in 282 Appendix). The range of evaporative fractions is important to calculate the sensitivity of warming rates 283 to evaporative fraction. 284 285 Next, we quantify the response of surface and air temperature warming rates to changes in evaporative 286 fraction from dry to wet conditions. The warming rate response to evaporative fraction is obtained from 287 the linear regression of daily warming rates to daily evaporative fractions for each site. Figure 4 shows 288 the mean response of the surface (orange) and air (blue) temperature warming rate to evaporative 289 fraction for short vegetation, savanna and forest. For site-specific responses, see Figure A2 in the 290 Appendix. It is noticeable that regardless of the magnitudes of the warming rates and different mean 291 evaporative conditions, the response of warming rates to evaporative fraction is almost consistent for 292 the different vegetation types. For instance, the surface temperature warming rate of short vegetation 293 shows a consistent decrease of ~23 x 10 -3 K/W m -2 from dry to wet days. However, the air temperature 294 warming rate decreases only by ~5 x 10 -3 K/W m -2 . In our earlier work, similar responses were 295 observed for a cropland site (See Appendix, site 8). We find a similarly weak response for savanna and 296 forests. In savanna, the surface temperature warming rate still decreases by ~12x10 -3 K/W m -2 from dry 297 to wet conditions, but the air temperature warming rate remains almost the same. In forests, both, The error bars represent the standard error in the mean of all sites in the respective type. 304

305
Besides evaporation, the aerodynamic conductance also influences the diurnal variation of temperature. 306 The aerodynamic conductance governs the ventilation of energy and mass from the surface to the 307 atmosphere (Thom, 1972). Figure 5 shows the mean aerodynamic conductances for the vegetation 308 types. The mean aerodynamic conductance is usually a characteristic of vegetation height but 309 variations might occur due to changing evaporative conditions. In general, it is observed that the 310 aerodynamic conductance of short vegetation is much lower than the aerodynamic conductance of 311 forest. Savannas show relatively higher aerodynamic conductances compared to short vegetation. Some 312 woody savannas have comparable aerodynamic conductances to forests. Forests have generally high 313 aerodynamic conductances. 314

315
In addition to the mean aerodynamic conductance we also observed its response to evaporative 316 fraction. The change in aerodynamic conductance due to the change in evaporative fraction is denoted 317 by ! ! that is derived from the linear regression of their observed daily values. The negative sign of 318 ! ! reflects the decrease in ! from dry to wet days so that the aerodynamic conductance is enhanced 319 on days with low evaporative fraction. Site-specific values of ! and ! ! are provided in Figure A3 in aerodynamic conductance whereas the enhanced aerodynamic conductance is just a secondary factor 327 whose impact on warming rate is further analyzed using our model.  To explain these findings we hypothesize that the high aerodynamic conductance of forest and its 347 enhancement on dry conditions lowers the diurnal warming of surface temperature. Consequently, the 348 warming rates of surface temperature of forests are less sensitive to evaporation. Our hypothesis is 349 based on the observational based findings and the interpretation of the model equations where one can 350 determine the contribution of ! , ! and ! ! in shaping the warming rates. This can already be 351 anticipated from Eq. 6, evaluated with the median values provided in Table 3. The first term on the 352 right-hand side of Eq. 6 is about -21 x 10 -3 K/(W m -2 ) for short vegetation, but only -7 x 10 -3 K/(W m -2 ) 353 for forests, similar to what is shown in Figure 4. In the next section we verify our hypothesis using the 354 modeled expression for surface temperature warming rate and its response to evaporative conditions. In this section we estimate the surface temperature warming rate and its response to evaporative 362 fraction using our model, which is then compared to observations. Then we use the model to quantify 363 the contribution of evaporative fraction and of aerodynamic conductance to the diurnal temperature 364 range. 365

366
To model the warming rate we use Eq. (5), in which the vegetation type is captured by ! , and 367 evaporative conditions by ! . The model sensitivity of surface temperature warming rate to evaporative 368 fraction and aerodynamic conductance is shown in Figure 5a. The model shows a stronger gradient of 369 the warming rate with evaporative fraction for low aerodynamic conductances. As in the observations 370 warming rates for low aerodynamic conductances are greater compared to high aerodynamic 371 conductances. Observed warming rates for short vegetation, savannas and forests are also plotted in 372 increases with decreasing evaporative fraction. Contrarily, the forest sites show no such strong 378 variation in warming rate. This is consistent with the study by Diak and Whipple (1993) We next tested the model by estimating the daily surface temperature warming rate for each site using 394 Eq. (5) from daily values of observed ! , ! and ! ! . Since ! ! is similar for all sites, the 395 diurnal variation of air temperature does not seem to depend on the diurnal variation of surface 396 temperature, and vice versa. Figure 6b shows the comparison of the modeled surface temperature 397 warming rates to those derived from observations. The model performs very well for all sites for the 398 given information. The coefficient of determination (r 2 ) is also high for savanna and forests, pointing at 399 the functionality of our model for complex and taller vegetation. However, short vegetation shows 400 slightly weaker r 2 because our model underestimates the surface temperature warming rate at a few 401 short vegetation sites. We speculate that these are the sites with non-vegetated surfaces where the 402 ground heat flux contribution to diurnal surface temperature variations is significant (Saltzman and 403 Pollack, 1977) which is currently neglected in our model. 404

405
It is apparent from Figure 6 that the response of the surface temperature warming rate to evaporative 406 fraction is predominantly governed by the aerodynamic conductance. The expression for ! ′ ! in 407 Eq. (6) quantifies this. Note that here we do not assume a constant aerodynamic conductance since ! 408 in the observation is enhanced on dry days. Our model reproduces the response of the warming rates to 409 evaporative conditions (r 2 = 0.6) for all types, Figure 7a compensating the response of warming rates to evaporative conditions. For this, we compare the 428 modeled ! ′ ! with and without the inclusion of enhanced aerodynamic conductance term (the 429 second term on the right-hand side of Eq. 6), such that ! ′ ! when ! ! = 0 only captures the 430 contribution of mean aerodynamic conductance and ! ′ ! when ! ! ≠ 0 additionally shows the 431 contribution of the enhanced aerodynamic conductance on drier days. For the comparison of the two 432 cases it is important to recognize that the more negative the values of ! ′ ! , the stronger the 433 imprint of evaporation is in the diurnal variation of temperature. 434

435
In general, for most of the sites the enhanced aerodynamic conductance plays a small, but noticeable 436 role in weakening the response of the warming rate to evaporative fraction. This is evident since the 437 data points lie above the 1:1 line and tend to be less negative for the case when ! ! ≠ 0. This effect is, 438 however, more consistent for forests compared to short vegetation and savannas (see the inset bar plot 439 which summarizes the mean ! ′ ! for two cases). For short vegetation sites, ! ′ ! decreases 440 only by 6% when ! ! ≠ 0 is considered. For savannas, the decrease is 32 %, and it is highest with 53% 441 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License. for forests. This suggests that along with the inherent high aerodynamic conductance of forests, its 442 enhanced aerodynamic conductance is also responsible for the absence of evaporation imprints in the 443 diurnal variation of temperature. The higher aerodynamic conductance of forests is responsible for 444 reducing 74 % of the imprints of evaporation in diurnal surface temperature when compared to the 445 short vegetation. We next link our model for surface temperature warming rates back to the diurnal variation in surface 456 temperature. To understand how solar radiation, the aerodynamic conductivity of the different 457 vegetation types and evaporative fraction individually influence the diurnal variation in temperature, 458 we can obtain the diurnal surface temperature range (DT s R) by multiplying the expression for warming 459 rate given by Eq. 5 with the daily maximum in absorbed solar radiation. To quantify the sensitivity of 460 DT s R to its three main contributors, we considered four cases. In first case, we assume that the diurnal 461 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License. variation in surface temperature is solely driven by solar radiation, such that there is no evaporation 462 ( ! = 0) and the surface has no vegetation, represented by a very low aerodynamic conductance of 463 ! = 0.02. Figure 8a shows that in this scenario, DT s R is overestimated for all vegetation types with 464 poor r 2 ≤ 0.3. This greater warming indicates that vegetation and evaporation cools surface 465 temperatures. In second case we added the information on aerodynamic conductance of each vegetation 466 type along with solar radiation (Figure 8b). The DT s R estimates for forests (r 2 =0.54) and to some extent We demonstrate a robust way of characterizing the diurnal variation of temperature using their morning 484 to noon warming rates, which are derived from the half hourly temperatures and solar radiation. The 485 warming rate is suitable for the comparison of locations with different solar energy input whereas other 486 metrics like diurnal temperature range depends on solar radiation (Makowski et al., 2009). 487 Consequently, temperature warming rates for specific vegetation types are comparable for sites at 488 different geographic locations. Our surface energy balance model can reproduce the warming rate and 489 shows the physical significance of evaporation and aerodynamic conductance. The model can capture 490 the diurnal variation of temperature quite well. These approximations can further be improved by a 491 more detailed formulation of net longwave radiation (which could, for instance, include optical 492 properties of the atmosphere) and the ground heat flux. Warming rates are also sensitive to clouds and 493 might not capture the information of evaporation and vegetation on cloudy days. Also, we did not 494 provide a way to calculate warming rates of air temperature. These could represent topics for future 495 research. 496

497
One of the main findings of our study is the different response of diurnal surface and air temperature to 498 evaporation. The air temperature warming rate does not contain any imprints of evaporation whereas, 499 for short vegetation, the surface temperature warming rate decreases strongly with evaporative fraction. 500 This finding is consistent with our previous work where we explained the role of boundary layer in 501 https://doi.org/10.5194/hess-2020-95 Preprint. Discussion started: 16 March 2020 c Author(s) 2020. CC BY 4.0 License.
compensating imprints of evaporative conditions in the diurnal variation of air temperature. We found 502 that the diurnal variation of air temperature is similar for all vegetation types irrespective of their 503 aerodynamic conductance and evaporative conditions. We anticipate that our hypothesis of the 504 compensating effect of boundary layer might also be true for forests, but this would need further 505 research. 506

507
The notion that diurnal surface and air temperature variations respond differently to evaporation should 508 be considered when developing air temperature products from remotely sensed surface temperature 509 (Cresswell et al., 1999;Fu et al., 2011;Hengl et al., 2012;Jang et al., 2004;Kilibarda et al., 2014;Zhu 510 et al., 2013). Typically, these products are primarily based on the assumption that surface temperature 511 is proxy of air temperature. Generally, these approaches overestimate daytime air temperature (Oyler et 512 al., 2016;Zhang et al., 2011). This finding is consistent with our results, which show a greater 513 variation of surface temperature depending on vegetation type and evaporative fraction (cf. Eq. 5). 514

515
Our study shows that surface and air temperature warming rates are similar in forests, which indicates 516 the strong coupling between the two temperatures. This finding is in agreement with the previous study 517 by Li et al., 2015 andMildrexler et al., 2011 , where evaporative cooling and high aerodynamic 518 conductance of forests were identified as the responsible factors for the strong coupling of surface and 519 air temperature. However, we also show that this coupling remains persistent irrespective of the 520 evaporative conditions of the forest. Using our model and observations we show that the aerodynamic 521 conductance of forest increases on dry days resulting in reduced warming of surface temperature and 522 hence its stronger coupling to air temperature. These findings complement the recent studies on the 523 convector effect where its role in lowering the surface temperature for a semi-arid forest is discussed 524 (Banerjee et al., 2017(Banerjee et al., , 2018Brugger et al., 2019;Kröniger et al., 2018;Rotenberg and Yakir, 2010). 525 Our demonstration of enhanced aerodynamic conductance on dry days is similar to what these authors 526 describe as the convector effect. 527 528 Unlike forests, the surface temperature warming rate in short vegetation responds strongly to changes 529 in evaporative conditions. In observations, the warming rate decreases by ~23 x10 -3 K/W m -2 from dry 530 to wet days. In general, this decrease is comparable for all the short vegetation sites and we anticipate 531 that some spread is due to their somewhat different aerodynamic properties. Another source of 532 ambiguity is the unequal distribution of days of different evaporative fractions which also influences 533 the computation of ! ′/ ! and ! ! . This constraint requires longer time series of observations to 534 obtain a greater sampling range of dry and wet days. Overall, our results nevertheless show that the 535 surface temperature warming rate is a promising indicator of evaporative fraction, especially for short 536 vegetation. 537

538
The other implication of our study is a better physical understanding of the processes that govern the 539 diurnal temperature range. Our model is capable of capturing the contribution of solar radiation, Temperature and evaporation are among the foremost-discussed variables in hydrology and climate 547 science. Our study contributes information on the relationship between diurnal temperature variations, 548 evaporative conditions, and vegetation. To measure the diurnal variation, we introduce the morning to 549 noontime warming rate of temperature. This rate has advantages over mean, maximum or minimum 550 temperatures for conducting a multisite analysis because it removes the effect of solar radiation. We 551 demonstrated that the warming rate and its response to evaporation is reproducible from the 552 simplification of surface energy balance. In doing so, we can address the two major questions that we 553 formulated in the introduction. First, our observational analysis shows no imprints of evaporation in 554 the air temperature warming rate across vegetation types. However, the surface temperature response to 555 evaporation is rather vegetation dependent, being stronger for short vegetation and absent in forests. 556 These findings provide insights for the second question about the role of aerodynamic conductance. 557 We showed that the aerodynamic conductance is very important in reducing the diurnal variation of 558 surface temperature. It is mostly the high aerodynamic conductance of forests, which compensates their 559 response to evaporative fraction. In addition, the aerodynamic conductance in itself is sensitive to 560 evaporative conditions. Using observational and model-reproduced findings we demonstrate that along 561 with the high aerodynamic conductance of forests their aerodynamic conductance roughly doubles on 562 dry days. The higher aerodynamic conductance results in more efficient transport of heat from the 563 surface to the atmosphere and compensates for the diurnal rise in surface temperature, which is 564 reflected in their lower surface temperature warming rate. 565 566 To conclude, our results imply that diurnal temperature variations can be understood and predicted by 567 relatively few factors, solar radiation, aerodynamic conductance and evaporative fraction. Surprisingly, 568 diurnal air temperature carries little information of vegetation type and evaporative conditions of the 569 land surface, while surface temperatures carry a stronger imprint of evaporation, but only for short