Strong influence of trees outside forest in regulating microclimate 1 of intensively modified Afromontane landscapes

. Climate change is expected to have detrimental consequences on fragile ecosystems, threatening biodiversity 9 as well as food security of millions of people. Trees are likely to play a central role in mitigating these impacts. The 10 microclimatic conditions below tree canopies usually differ substantially from the ambient macroclimate, as vegetation 11 can buffer temperature changes and variability. Trees cool down their surroundings through several biophysical 12 mechanisms, and the cooling benefits occur also with trees outside forest. The aim of this study was to examine the effect 13 of canopy cover on microclimate in an intensively modified Afromontane landscape in Taita Taveta, Kenya. We studied 14 temperatures recorded by 19 microclimate sensors under different canopy covers, and land surface temperature (LST) 15 estimated by Landsat 8 thermal infrared sensor. We combined the temperature records with high–resolution airborne laser 16 scanning data to untangle the combined effects of topography and canopy cover on microclimate. We developed four 17 multivariate regression models to study the joint impacts of topography and canopy cover on LST. The results showed a 18 negative linear relationship between canopy cover percentage and daytime mean ( R 2 = 0.65) and maximum ( R 2 = 0.75) 19 temperatures. Any increase in canopy cover contributed to reducing temperatures. The average difference between 0% 20 and 100% canopy cover sites was 5.27 ˚C in mean temperatures and 10.2 ˚C in maximum temperatures. Canopy cover 21 reduced LST on average by 0.05 ˚C/%CC. The influence of canopy cover on microclimate was shown to vary strongly 22 with elevation and ambient temperatures. These results demonstrate that trees have substantial effect on microclimate, 23 but the effect is dependent on macroclimatic conditionse, highlighting the importance of maintaining tree cover 24 particularly in warmer conditions. Hence, we demonstrate that trees outside forests can increase climate change resilience 25 in fragmented landscapes, having strong potential for regulating regional and local temperatures. and slope (°), on temperature. We examined the relationships between the variables first with Pearson’s correlation using elevation, slope and CC as explanatory variables in a multiple regression model. Elevation and CC were the only statistically significant variables. We corrected the daytime mean temperatures according to the altitudinal lapse 165 rates, which were 7.26 ° C km -1 for soil temperature (T soil ), 8.09 ° C km -1 for surface temperature (T surface ) and 8.06 ° C km - 166 1 for air temperature (T air ). In the case of diurnal analysis, we applied separate lapse rates for each hour that were derived from the regression analyses. The lapse rates were 6.1 ° C–8.2 ° C km -1 in T soil , 3.8 ° C–10.4 ° C km -1 in T surface , and 3.3 ° C– 10.2 ° C km -1 in T air .In the case of diurnal analysis, we applied separate lapse rates for each hour. These varied from 6.1 ° C to 8.2 ° C km -1 in T soil , 3.8 ° C to 10.4 ° C km -1 in T surface and 3.3 ° C to ° C km -1 in T air . To find the relationships between temperature, CC and topographic variables, we conducted statistical analyseis, including descriptive statistics, linear regression and Pearson’s correlation. We used standard deviation (SD) to describe the variability of temperatures. We used RStudio (R Core Team, 2019) for all statistical analyseis. collection hemispherical at field for validating the raster. proceeding the analysis using ALS. Our results demonstrate a consistent but heterogeneous influence of canopy cover on the microclimate of highly diverse tropical ecosystems. Daytime temperatures correlated inversely with canopy cover, the effect being strongest on surface temperatures. During the hottest daysIn hotter environments, the difference between sites of high and low canopy cover 556 became most notable. The cooling effect did not exist only with high canopy cover, but even intermediate canopy cover 557 and trees outside forest buffered the hottest temperatures. Our results thus provide robust evidence that any efforts in the direction of preserving, restoring or increasing vegetation cover can have a substantial impact in creating more stable and cooler microclimates. Satellite- based land surface temperatureLST was a suitable proxy for assessing microclimatic variables surface- and near-ground temperatures, particularly in heterogeneous regions, where the network of field measurements cannot cover the spatial microclimate variability. the effect our results indicate that warmer and drier regions are likely to the most trees.

4 relationship with LST due to evapotranspiration causing increased latent heat loss from the canopy (Goward et al., 1985; 91 Goward and Hope, 1989;Nemani and Running, 1997). Canopies' cooling effect has different magnitudes at different   (Korhonen et al., 2006) and it is the most important 103 variable used in defining forests or other land with tree cover (FAO, 2015). ALS can assess tree cover over large areas 104 more precisely than field measurements can. Hence, when ALS is combined with either field-based or remotely sensed 105 temperatures, we can study the influence of trees on temperature in a new way of that is both nuanced and large scale.

106
The complexity of the issue with climate change requires attention at both spatial resolutions.

107
The primary objective of this study was to examine how different levels of CC can contribute to lower temperatures and 108 more stable microclimates across a highly heterogeneous Afromontane landscape in Kenya. We based our analysis on 109 micro-climatological measurements and CC estimates retrieved from ALS data. Microclimate sensors cannot entirely 110 capture the spatial variability of temperatures, especially in heterogeneous landscapes. Therefore, we used satellite 111 thermal data to provide a comprehensive and spatially continuous representation of the relationship between CC and 112 temperature.  (Fig. 1). The elevation in the study area varies from 640 m in the 120 lowlands to the highest peak of the hills, Vuria, at 2208 m. Climate is mainly semi-arid. According to the Kenya Ministry 121 of Agriculture, Livestock and Fishery (MoALF), annual precipitation averages 650 mm, but differences between hills and 122 lowlands are notable: lowlands receive 500 mm annually compared to 1500 mm in the hills. Two rainy seasons control 123 the climate and growing seasons: long rains from March to June, and short rains from October to December (Pellikka et        We applied an ALS-based Digital Elevation Model (DEM) raster at 1 m resolution and a CC raster at 30 m resolution.

144
The ALS data for the hills were acquired in February 2014 and February 2015, and the data for lowland areas in March 145 2014. The mean pulse density of the ALS data in the hills was 3.1 pulses/m -2 and mean return density 3.4 returns/m -2 , for 146 the lowlands the pulse density was 1.04 pulses/m -2 . The ALS data used in this study are described in detail in Adhikari et

148
We resampled the DEM to 30 m resolution to fit to the spatial resolution of the Landsat 8 image, and utilized it to derive 149 topographic factors slope degree (°) and aspect (°) using ArcGIS Pro spatial analyst tools.

151
Based on the CC raster derived from the ALS data, we selected a total of 19 field plots representing different CC levels 152 (Table 1). In the plots, we installed TOMST TMS-4 microclimate sensors to measure temperature at three different 153 heights: soil at 6 cm below ground, surface at 2 cm above ground, and air temperature at 15 cm above ground (Tsoil, Tsurface 154 and Tair, respectively) (Wild et al., 2019). from June 13 to July 10, 2019. The sensors were deployed in places that were 155 as flat as possible to reduce the effect of slope, and that received both sunlight and shade during the day with the changing 156 sun angles. In high CC sites, the sensors were shaded most of the day, while in the open areas, the sensors were exposed 157 to sunlight all day.

158
The sensors measured parameters every 15 minutes from June 13 to July 10, 2019. We calculated daytime temperature 159 aggregates between sunrise and sunset, local time 06.30-18.30 UTC + 3h. We calculated maxima as the mean of daily 160 maxima, and minimum temperatures as the mean of minimum temperatures based on the 24 hour cycle.

161
To isolate the influence of CC on microclimate, we quantified and later removed the effect of topography, such as 162 elevation (m) and slope (°), on temperature. We examined the relationships between the variables first with Pearson's 163 correlation using elevation, slope and CC as explanatory variables in a multiple regression model. Elevation and CC were 164 the only statistically significant variables. We corrected the daytime mean temperatures according to the altitudinal lapse 165 rates, which were 7.26 °C km -1 for soil temperature (Tsoil), 8.09 °C km -1 for surface temperature (Tsurface) and 8.06 °C km -166 1 for air temperature (Tair). In the case of diurnal analysis, we applied separate lapse rates for each hour that were derived 167 from the regression analyses. The lapse rates were 6.1 °C-8.2 °C km -1 in Tsoil, 3.8 °C-10.4 °C km -1 in Tsurface, and 3.3 °C-168 10.2 °C km -1 in Tair.In the case of diurnal analysis, we applied separate lapse rates for each hour. These varied from 6.1 169 °C to 8.2 °C km -1 in Tsoil, 3.8 °C to 10.4 °C km -1 in Tsurface and 3.3 °C to °C km -1 in Tair. To find the relationships between 170 temperature, CC and topographic variables, we conducted statistical analyseis, including descriptive statistics, linear 8 regression and Pearson's correlation. We used standard deviation (SD) to describe the variability of temperatures. We 172 used RStudio (R Core Team, 2019) for all statistical analyseis.

173
The ALS data was 4-5 years older than the field measurements. Moreover, the ALS data was collected during the short 174 dry season, in contrast to the field measurements, which we carried out during the start of the long dry season in June 175 2019. To address the mismatch between the data collection dates, we acquired hemispherical photography at each field 176 plot for validating the CC raster. The differences in CC were not statistically significant and we considered the estimates 177 consistent enough for proceeding the analysis using CC from ALS. In the case of Mwatate river plot, CC was retrieved 178 by hemispherical photography only, because the plot was outside of the ALS coverage. The methodology is described in 179 Appendix A.Because the ALS data was 4-5 years older than the field measurements, we acquired hemispherical 180 photography at each field plot for validating the CC raster. Moreover, the ALS data was collected during the short dry 181 season, in contrast to the field measurements, which we carried out during the start of long dry season in June 2019. For

182
Mwatate river plot, CC was retrieved by hemispherical photography only, as the plot was laying outside of the ALS 183 coverage. The methodology is described in the supplementary material.

206
where w = water vapor content, T0 = air temperature and RH = relative humidity.

207
The SC formula is shown in Eq.

218
We conducted similar topographic correction with the Landsat image as with microclimate sensors to exclude the effect 219 of topography on LST. Topographic variables (elevation, slope and aspect), CC, and their interaction terms were included 220 as independent factors and LST as the dependent factor in four multiple regression models (Table 2). and LST were 221 included in a multiple regression model. We classified aspect to nine classes indicating eight cardinal directions (south, 222 south-west, west, north-west, north, north-east, east, south-east), and flat surface. The classes were treated as dummy 223 variables due to their categorical nature. We also classified elevation to three classes: below 1000 m, 1000-1500 m, and 224 above 1500 m. We used the LST at elevation of 880 m, slope of 0 ° and aspect class north as reference. Following class "north". We used linear regression to study how much CC percentage and topographic variables affected 230 microclimate and LST. In total, we estimated four different models for LST (Table 2).

253
T'soil showed a more gradual decrease than T'soil and T'surface, where SD decreased substantially only in high CC sites ( Fig.   254 3). The SD of maximum temperatures were higher than in mean temperatures.

255
Based on the regression coefficients, which indicate the magnitude of the influence of CC on temperature, the cooling 256 effect of CC was stronger on maximum temperatures than mean. Additionally, whereas CC affected mean T'soil more than 257 mean T'air, in maximum temperatures the situation was the opposite, and T'air was more affected by CC than T'soil (  fluctuation even during the hot day streaks: differences remained even less than 1 °C between hottest and coolest days.

267
When comparing the three measurement heights, the coldest mean temperatures were measured in T'air and the hottest in

276
Minimum temperatures showed no explicit relationship with CC, and sites with similar CC had high temperature 277 variability. R 2 were low (< 0.2) at all measurement heights, and correlations between temperatures and CC were 278 insignificant. All results from the regression analyses are summarized in Table 3. 284 Figure 4 shows the intra-daily temperature variability based on study period means. T'soil were more stable than T'surface 285 and T'air that showed higher peaks and drops. In the morning, temperatures at all measurement heights started to rise

332
showing little fluctuation even during the hot day streaks: differences in mean temperatures remained even less than 1 °C 333 between hottest and coolest days.

334
The cooling effect of CC varied throughout the study period: on hot days, the cooling effect (described by CC's regression 335 coefficient in Fig. 4)

390
Model 4 was built up on Model 2 by adding interaction terms between slope and aspect classes and CC (Table 4)

397
In summary, including either of the elevation factors (DEM or elevation zones) in the model showed that elevation 398 affected CC's cooling effect significantly, having almost two times higher impact in the lowlands compared to the hills.

399
The dependence of CC's impact on elevation is demonstrated in Fig. 77

488
In addition to dew capture, agroforestry has shown to contribute to improved soil moisture (Rhoades 1995; Siriri et al.

491
The pressure on tropical forests in sub-Saharan Africa is caused by many reasons, fuelwood collection being significant  Nevertheless, water is scarce especially in the lowland areas, and trees' vast need for water must be taken into account.

496
The phenomenon is paradoxical, since because trees improve the water cycle, in general, but are consumes high amounts 497 of water (Ong et al., 2006). Water balance also affects the temperature buffering capacity of trees (Davis et al., 2019). In 498 areas with water scarcity, the competition forof water resources with between crops, animals and people may be a limiting

581
Thimonier et al., 2010). From the two images, we used the less exposed one in the analysis. For the calculation of canopy 582 cover, we used the images taken from eye-level, because they were more comparable to the ALS-based canopy cover, 583 and the photographs in cardinal directions were all taken at eye-level. We classified the image pixels to sky and canopy 584 by determining a threshold value to separate dark and light pixels in the image. For most images, we used the automatic 585 threshold method by Nobis and Hunziker (2005). In the case of some images, the algorithm clearly produced errors due 586 to overexposure and direct sunlight, therefore the algorithm by Ridler and Calvart (1978) was applied, or a manual 587 threshold was determined. We used only the blue band in the analysis, apart from photographs where the classification 588 was failing and using all the bands produced the best result (Heiskanen et al., 2015a). The gamma correction was γ = 2.2.

589
Only the zenith angle range of 0-15° was analyzed, because errors in canopy cover accuracy increase with larger angles 590 (Paletto and Tosi, 2009

868
Field Code Changed