Inuence of vegetation cover change on the decadal trend of summer seasonal air temperature in the Pannonian Basin

The inuence of land surface vegetation on the atmosphere processes in the planetary boundary layer is of great importance for the study of weather and climatic conditions in the Earth’s climate system. Vegetation, as an integral part of the Earth’s climate system, has a great inuence on the exchange of energy between land and the atmosphere and, consequently, a signicant role in dening weather and climate patterns at the global, regional and local scales. However, due to the constant anthropogenic impact, this vegetation system is continuously changing mostly due to deforestation, afforestation, and forest res which make it dicult to present them during the research of the Earth’s climate system. The aim of this study is to examine the impact of the regional vegetation change on the seasonal surface air temperature and was performed using the Max-Planck-Institute Earth System Model. The region of our research is located in the Pannonian Basin and is one of many regions in which the anthropogenic impact on geophysical changes in the environment is considerable. The study was carried out over a ten-year period, from 2002 to 2011, during which we showed that the change in the presence percentage between certain types of vegetation leads to warming up as well as cooling down of air during the summer season. We have also shown to what extent this change in vegetation has an impact on the surface air temperature trend as well as on the change in the albedo and ux of sensible heat.


Introduction
Vegetation affects the Earth's climate through various processes that can be divided into two basic groups such as the bio-geophysical and biogeochemical. The main bio-geophysical processes include surface energy, moisture, and momentum uxes which are de ned by the physical properties of the vegetation such as albedo, aerodynamic roughness, and leaf area (Claussen et al., 2001;Betts, 2006;Bonan, 2008;Port et al., 2012). While the main chemical processes include the interaction with carbon dioxide CO 2 and ozone O 3 as well as aerosol production due to the release of hydrocarbons (Arneth et al., 2010). Each of these processes signi cantly affects the exchange of energy between the land and the atmosphere, resulting in changes in the regional and global circulation of the atmosphere and, thus, the impact on climate. For example, forest vegetation has a small albedo, so it absorbs a large part of the incoming shortwave solar radiation. Most of this radiation, through long-wave radiation, is spent on heating the air above forest vegetation through radiation, conduction, and convection (Betts and Ball, 1997). In the surface layers, the air is cooled by the process of evapotranspiration, while, due to the transfer of latent heat to higher layers of the atmosphere, it is heated during its release, i.e., we observe the production and transport of water vapor and its in uence as a component gas of the greenhouse effect on the air heating and cooling. Aerodynamically, forest vegetation is a very rough surface that increases turbulence and reduces wind speed in the surface layer of the atmosphere (Rotenberg and Yakir, 2010;Vautard et al., 2010). This increase in turbulence above forest vegetation encourages convection and, consequently, the formation of clouds and the frequency of precipitation over areas covered with forest vegetation (Pielke et al., 2007;Wulfmeyer et al., 2011;Ellison et al., 2012). Also, an increase in cloudiness above the forest vegetation can leads to an increase in the albedo of cloudiness, which affects local and regional air cooling (Pongratz et al., 2009;Teuling et al., 2017).
The intensity and duration of these processes mostly depend on the type of forest vegetation, which we can represent through the geographical area in which it is present. Due to this division, we can de ne the following types of forest vegetation: northern, temperate, and tropical forests (Bonan, 2008). For example, northern forests due to the small albedo have a great in uence on the regional air temperature throughout the year and relative to the other two types, they also have the greatest in uence on the global mean temperature, while the CO 2 absorption and evapotranspiration are lower (Betts, 2006;Bonan, 2008;Lee et al., 2011). The impact of temperate forests on the climate relative to the northern or tropical ones is more complex and uncertain, mainly due to its mixed composition and vegetation period of certain forest species (Bala et al., 2007;Bonan, 2008;Tölle et al., 2018). As a result, the in uence between cooling and heating varies regionally and seasonally . This type of forest is particularly interesting to us because it mitigates the effects of heatwaves over a long period of time in the areas of Central and Western Europe (Tueling et al. 2010;Stéfanon et al., 2012;Bastos et al., 2013;Bevan et al., 2013). It also covers the areas in the regions where the human population greatly in uences environmental change by changing local surface conditions. This is just a small overview of how important and complex the impact of forest vegetation on Earth's Climate System (ECS) is. However, due to the development of complex numerical climate models and weather forecasting models, we have a better insight into the research of the interaction of vegetation with the atmosphere by numerical simulations. With this method, various simulations were performed in which regional or global analyses of the in uence of the areas under vegetation on climate and weather conditions were performed. In general, the results obtained by this method showed that there is an increase in surface air temperature and a decrease in precipitation in areas where forest vegetation was removed. For example, Schneck and Mosbrugger (2011) studied the local and regional impact of deforestation in Southeast Asia, where they found that in areas where forest vegetation was removed, the surface air temperature increased and the precipitation decreased, and the circulation in the ocean changed locally. Pitman et al. (2009) performed a series of numerical simulations with the decrease in the areas under forest vegetation and concluded that some numerical models resulted in a decrease in the surface temperature, whereas others increased. The changes they obtained for total precipitation were generally small while a certain number of models gave a decrease in precipitation over deforested areas. Sanchez et al. (2007) observed the potential impact of changes in the area under forest vegetation in Europe where they concluded that precipitation in most of Europe was higher where the area under forest vegetation increased and the surface air temperature decreased by 1 to 3 degrees Kelvin scale. Anav et al. (2010) examined the impact of a decrease and increase in the area under forest vegetation on the climate in Europe. They concluded that there is an increase in the number of warm days due to the decrease in the area under forest vegetation and a decrease in the number of warm days due to the increase in the area under forest vegetation.
Depending on the type of research, we use models for numerical weather prediction (NWP), regional climate models (RCM), or global climate models (GCM). These models should contain numerical parts in which the dynamic development of plant mass is de ned, i.e., they should contain physical and chemical processes through which vegetation cycles are de ned, as well as the processes of photosynthesis and evapotranspiration throughout the year.
The role of such numerical parts is to simulate the physical and thermodynamic conditions that occur in the vegetation layer and the layer above it as accurately as possible and therefore they affect the accuracy of the result.

Method And Data
We presented the results in a time as well as space framework. The seasonal period was taken to be the time frame, while for the spatial frame, we presented the results at the local and regional scale (Fig. 1). When de ning the seasonal period as a time frame, we used the meteorological de nition for the seasonal value, which is based on the annual surface air temperature cycle. That is, meteorologists and climatologists divide the seasons into groups of three months using a standard civil calendar based on the annual surface air temperature cycle. Thus, meteorological spring includes March, April, and May (MAM); meteorological summer includes June, July, and August (JJA); meteorological autumn includes September, October, and November (SON) and meteorological winter includes December, January, and February (DJF). Mean seasonal air temperatures (MST) are calculated as the arithmetic mean between the mean monthly air temperatures that make up the same season. For our research, we de ned the MST at the surface (MST 2m ) as well as MST at 13 vertical levels (MST lev ) of the atmosphere located near the Earth's surface, and we presented them with isobaric surfaces expressed in hecto Pascal (hPa). We de ned the multiannual or decadal average of MST as the arithmetic mean between all ten-time steps that make up the summer mean seasonal values, that is, for JJA season.
The results for the areas at the local scale marked with A1, A2, A3,…, A11, and A12 ( Fig. 1) are obtained as the arithmetic mean between 4 nearby model grid points marked with black circles. Depending on the geographical position, we have grouped these local areas into the southern area consisting of A1, A2, A3, and A4 local areas, the central area A5, A6, A7, and A8, and the northern area consisting of A9, A10, A11, and A12. For the area at the regional scale, we obtained the results as the arithmetic mean value between all model grid points that cover the spatial area of our research, and there are 20 in total, i.e., the region includes a resolution of 5x4 model points, 5 in the longitudinal and 4 in the meridional direction ( Fig.1).
To examine the impact of the change in the percentage between certain types of vegetation in the Pannonian Basin on the local and regional climate, we will use the method of numerical simulation using GCM developed at the Max Planck Institute ( For our research, we determined the time from January 1, 2002, to December 31, 2011, that is, we did a decadal period of numerical simulations for a given period. The numerical simulation was performed three times, i.e., the whole research was divided into three steps. In the rst step, we performed a numerical simulation for a given period with vegetation from 2002. In this step, we determined the control experiment through which we de ned the deviation of the output results of the MPI-ESM from the approximately real state. In the second step, the vegetation cover from the period of 2002 was replaced with the vegetation cover from the pre-industrial revolution, i.e., from the year 850, while in the third step, the vegetation cover from 2002 was replaced with the future vegetation cover from the 2050 year. Changes with the vegetation cover is done only in the area marked in Figure 1, while outside that area the vegetation cover is the same for all three steps in the experiment. Later in the text, we will mark these three numerical simulations with CF 2002 (Cover Fraction from the year 2002) as the rst step, CF 850 (Cover Fraction from the year 850) as the second step, and CF 2050 (Cover Fraction from the year 2050) as the third step in our research.
To simulate the initial physical and thermodynamic conditions that occur in the global ECS as accurately as possible, we used four-dimensional data assimilation based on the method of Newton's relaxation or known as the "nudging" method (Krishnamurti et al. 1991;Jeuken et al. 1996). Newton's relaxation method is a method that reduces the deviations of simulated trajectories from the trajectories obtained by observation or from simulated analyses in a given time and the entire space. These projections of trajectories can represent certain values of the state of the atmosphere, such as temperature, vorticity, divergence, and the logarithm of the surface pressure (Giorgeta et al., 2013a). The solution of Newton's relaxation equation in the ECHAM6 model is presented implicitly and explicitly. For our research, we used an implicit solution with standard periods of relaxation time (Lohmann and HooseI, 2009;Rast et al., 2013). In all three steps of the experiment, the assimilation will be continuously forced for four months during the duration of numerical integrations, i.e., from January 1, 2002, to April 31, 2002. During this period, we performed numerical simulations with the inclusion of nudging assimilation of atmospheric data such as surface pressure logarithm, divergence, vorticity, and air temperature which we have taken over by the European Centre for Medium-Range Weather Forecasts (ECMWF ERA-Interim) (Berrisford et al., 2011).
The advantage of the research done with the method of numerical simulations using the MPI-ESM model is that this model covers most of the ECS, i.e., we can say that it covers all relevant processes that de ne the global circulation in the atmosphere and oceans. For the needs of the research, we prepared the models in the following way.

Model setup
The resolution of the MPIOM model is given with 256 grid points in the zonal and 220 grid points in the meridional direction (Wetzel et al., 2011), where the North Pole is located in Greenland and the South Pole in Antarctica, and this grid of points is often called bipolar. The horizontal points of the grid are arranged in the Arakawa C-grid (Arakawa and Lamb, 1977). The vertical resolution is represented by 40 levels in the z-coordinate system, while the time step of the model is ∆t ocean =2700s. The initial conditions required for the MPIOM model were taken over from the Copernicus Marine Service (CMS).
The ECHAM6 model is set to horizontal resolution with 192 grid points in the zonal direction while in the meridional direction, there are 96 grid points. Vertical resolution is represented by 47 levels over the σ-hybrid coordinate de ned in a way that the model levels in the lower troposphere follow orography while the levels in the upper troposphere and stratosphere become approximate pressure coordinates with the last level at 0.01 hPa or about 80 km altitude. The time step of the model is ∆t atmosphere =450s, while the time step for transporting radiation through the atmosphere is ∆t radiation =1800s. The cumulus convection is parameterized with the mass ux transport (Tiedtke, 1989) using a modi cation for deep convection (Nordeng, 1994). The turbulent mixing of velocity, heat, humidity, and tracers is parameterized with vortex diffusion (Brinkop and Roeckner, 1995). For the concentration of greenhouse gases in all three numerical simulations, we always used the same RCP8.5 (Representative Concentration Pathways 8.5) climate scenario (Riahi et al. 2011) for the time during which our research was de ned. Input data for RCP8.5 we took over from We edited the JSBACH model (Reick et al., 2013) by including processes such as phenology, vegetation dynamics, vegetation albedo, and Bio-sphere-Energy-Transfer-Hydrology (BETHY) submodel. This model uses a mixed or mosaic approach (Koster and Suarez, 1992) in which each cell in the horizontal grid of the model is divided into tiles allowing the subnetwork representation of the extent of heterogeneity in the observed cell of the grid. The horizontal resolution is the same as in the ECHAM6 model, and each grid cell is divided into 11 tiles (Table 1). Each of the displayed tiles represents a speci c CF and is connected with one of 21 Land Cover Type (LCT), or with 2 LCT if they have similar characteristics as is the case with tiles 3, 4, 7, and 11, Table 2. It should be noted here that the LCT glacier was placed in combination with the LCT Tropical evergreen in tile 1 to simplify the operation of the model, Table 1. Table 1 represents a separate section that simulates the exchange of water, carbon, energy, and air movement between the land surface vegetation and atmosphere. These exchanges are summed for all tiles and presented as a total exchange for a single cell of the grid point. It was this way of presenting CF in the JSBACH model that enabled us to do our type of experiment, i.e., to examine whether there is an impact of vegetation cover on MST by replacing CF from 850 and 2050 in the area of the Pannonian Basin and if there is, how big of an impact is it.

Each tile in
It should be noted that for CF 2002 forest and agricultural vegetation in the Pannonian Basin make up over 90 % of the total CF both locally and regionally. That means it is the biggest changes of CF were made with CF 850 , and CF 2050 in tiles 3, 4, 9, and 11, Table 3. Tiles type 1, 2, and 6 are not contained in the region of the Pannonian Basin at all. While the tiles of type 5, 7, 8, and 10 are contained signi cantly less part from the total part of CF and which amounts to 0.04 %, 0.64 %, 0.36 %, and 1.85 %, respectively. Forest vegetation is mostly of the type "Temperate broad-leaf deciduous" and "Deciduous conifer," while agricultural vegetation is of type C3 and C4. The types of CF, as well as their percentage representation outside the Pannonian Basin region, are the same for CF 850 and CF 2050 and were taken from the control period, i.e., from CF 2002 . Figure 2 gives an example of what the CF change for tile 4 looks like.
The data exchange between ECHAM6 and MPIOM models is performed every 24 hours, while the data exchange between ECHAM6 and JSBACH models is performed in each model time step de ned in the ECHAM6 model and equals ∆t atmosphere =450s. Initial conditions required for the initialization of ECHAM6, JSBACH, and MPIOM models are the same for all three steps of our experiment. The data we used to initialize the starting conditions in the ECHAM6 and JSBACH models were taken over by the ECMWF ERA-Interim (Berrisford et al., 2011). To de ne the anthropogenic impact on surface vegetation change, we used data based on the land-use harmonization (LUH1) protocol developed by Hurtt et al., (2011), and are placed by MPI (Hagemann, 2002;Pongratz, J., et al., 2008).

Trend
To determine the trend for the increase or decrease of the MST, we used the method of least squares to determine the linear trend where the slope showed the average change over time. The linear trend is given by the equation where the independent variable t represents the time and the dependent variable Y the phenomenon for which we de ne the trend, while c and m are our regression coe cients.
N represents duration of the experiment, N=10 year and i is counter of years. A standard statistical estimation of the output model data for CF 2002 was done using the methods: Pearson Correlation Coe cient (R), root mean square error (RMSE), and mean bias (MB).
Applying the MB method to the results of the control experiment CF 2002 , we obtained the results from the model reanalysis of ERA-Interim for MST 2m and MST lev . Therefore, every time we compared the results from CF 850 and CF 2050 with CF 2002 , we actually compared the results with the data from the given reanalysis.
The existence of a trend is determined using the Mann-Kendall test (Mann 1945, Kendall 1975, where we took a value for the level of signi cance 5%.

Statistical estimation of the output model data
In Table 4, we present the results for MB, R, and RMSE from CF 2002 for the decadal average of MST 2m for JJA season. At the regional scale, MB is -0.85 º C and we noticed that is mostly negative both on a regional and local scale, which means that the model for this period gave us smaller results than the reanalysis. R is negative and indicates a lack of linear correlation between model values and reanalysis for the JJA season. After applying the MB method to the results with CF 850 and CF 2050 , we obtained R with a weak negative linear correlation between the data. For RMSE, we found that it ranges around 1.8 at the regional scale, while at the local scale it ranges between 1.8 to 2.6, Table 4.
Page 7/16 3.2 Trend and decadal values of mean seasonal temperature For our research, we de ned the trends of MST 2m (Y MST2m ) whose values are shown through JJA season at the local and regional scale (Table 5) while the graphical presentation is given for local areas in Fig.3. At the regional scale (Fig.   7), Y MST2m is negative as in most local areas, and this means that we mostly have a decrease in MST 2m . This reduction mainly occurs in the central and northern areas, while in the northern area it is more intense, i.e., there is a faster cooling of the air, Fig.3. As for the southern area, we mostly got a positive Y MST2m , i.e., we have a warming of MST 2m . We only got a slight decrease in MST 2m in A4, and an increase in A5 with CF 2050 . Using the MK test, the existence of the Y MST2m , as well as it's decreasing and increasing, has been con rmed both at the local and regional scale for CF 850 and CF 2050 .
As for the decadal average of MST 2m (Table 6) at the regional scale, we obtained a temperature increase of about 1.5 º C. With CF 850, we got an increase of 1.7 º C at the regional scale while at the local scale, it ranges between 0.6 º C and 2.6 º C. These increases are only for the southern area because we obtained a positive Y MST2m for it (Fig.3), where there is an increasing trend of MST 2m . For the central and northern areas, there is a negative Y MST2m so, despite the increase we have received, there is a decreasing trend of MST 2m . While with CF 2050, we got an increase of around 1.3 º C at the regional scale, and this increase at the local scale ranges between 0.7 º C to 1.8 º C. The difference of the decadal average of MST 2m between CF 850 and CF 2050 for the JJA season ranges between 0.1 º C and 0.9 º C at the local and 0.4 º C at the regional scale, and we got a match in the northern area, Table 6.
We presented the decadal average of the MST lev in Fig. 4 at 13 vertical levels which we showed over isobaric surfaces (air pressure) from 1000 to 700 hPa with a vertical increment of -25 hPa. On the isobaric surface of 1025 hPa, we presented the value of the decadal average of MST 2m . In the given gure it can be seen that the decadal average of MST lev up to 900 hPa of surface follows the temperature ratio as given with the decadal average of MST 2m through season between CF 850 and CF 2050 . While at levels higher than 900 hPa, MST lev slowly equalizes between the CF 850 and CF 2050 experiments which is most pronounced in the northern area. The ratio of the differences we obtained for the decadal averages of MST 2m (Table 6) between CF 2002 with CF 850 and CF 2050 is very similarly represented for the decadal averages of MST lev in the atmosphere layer between 1000 hPa and 900 hPa of the isobaric surface.

Decadal values of aerodynamic roughness length and leaf area index
The initial changes in CF concentrations that we performed for CF 850 led to a signi cant change in the value of the aerodynamic roughness length (z 0 ) at the regional and local scale, which is roughly increased between 50% and 85% (Table 7). For CF 2050 this increase is much smaller and it is around 0% at the regional scale, while at the local scale we notice a slight decrease, mainly in the southern area, which is in line with the initial conditions for this step of our experiment. These changes in the z 0 are important because they indirectly affect the increase or decrease of the turbulent transport of uxes from the surface to the atmosphere, (Giorgetta et al., 2013a,b). Also, the higher z 0 at the Earth's surface reduces the wind speed (Rotenberg and Yakir, 2010;Vautard et al. 2010) and thus the preconditions for increasing turbulent transport.
The inclusion of vegetation dynamic and phenology module in the JSBACH model was a condition to obtain the production of leaf area index (LAI) which is very important in de ning surface albedo (Otto et al., 2011). The increase in LAI was obtained with CF 850 (Table 8) at both local and regional scale and is approximately between 1.5% and 13% at the local scale, and 6% at the regional. With CF 2050, we generally recorded a reduced LAI of 1.1% to 7% at the local scale and 3% at the regional.

Trends of albedo and surface sensible heat ux
Numerous studies with climate models indicate that temperate-type forests cool the air compared to areas where there are pastures or agricultural plantations, while other studies show the opposite (Jackson et al., 2008 andAnav et al., 2010). Some of these contradictions may be related to the season, water availability, and soil moisture levels. In our research, we got a heating trend only in the southern area, while in the central and northern areas, we have a cooling trend (Fig. 3).
One of the ways of heating the ground layer of air with the forests of temperate type is to distribute more heat in the atmosphere because they are darker and absorb more sunlight, i.e., they have a small albedo and heat the air through the radiation, conduction, and convection (Lee et al., 2011). For this type of forest, albedo ranges from 0.15 to 0.18 (Barry and Chorley, 1992), which means that between 15% and 18% of the total incoming solar radiation returns to the atmosphere. In our research, we obtained values of albedo, which at the regional scale is approximately 12.7% for the JJA season. These changes in albedo that we obtained in our experiments are directly related to the distribution of the surface sensible heat ux (sshf), Fig. 5, as one of the key processes by which forests can modify air temperature. This can be seen in Fig. 3 for the JJA season where the slope of trend Y MST2 is very similar to the slope of trend sshf (Y sshf ), However, if we look at the trend of albedo (Y albedo ) in Fig. 6 we see that with CF 2050 it's mostly slightly negative on the local and regional scale (Fig. 7), i.e. tends to reduce albedo. While with CF 850 for the JJA season we got that Y albedo is positive, Fig. 6, where it is most pronounced in the northern and central area. In the south area, this trend is very slight which again agrees well with Y MST2m and Y MSTlev for these areas and regions.
In the JJA season albedo has a signi cant role in the regulation of surface air heating via sshf because these two trends are mostly negatively linearly correlated. That is when Y albedo increases Y sshf decreases and vice versa. But here appears the second process that cools the ground air, because of an indirect increase in albedo which we got with CF 850 . In the JJA season, we have an increase of albedo but due to an increase in vegetation cover, we should have a decrease in albedo and thus heating of surface air.
This happens due to the increase of cloudiness above the forest areas (Arneth et al., 2010) and consequently, we got an increase in cloud albedo, and thus the surface albedo which further causes a decrease in surface air heating.
With CF 850 and CF 2050 , we got a positive trend of cloud cover for the northern area, while for the central area it is mostly slightly positive with CF 2050 and slightly negative with CF 850 which is in line with the trends of albedo and MST 2m . This increase in cloudiness over forest covers during the summer season is in line with the results they have obtained (Teuling et al., 2017).
For the trend of MST 2m cooling that we got at the regional scale during the JJA season (Fig. 7), we can say that it mainly happens due to the changes of forest cover of temperate type, because we made in the central and northern area an increase of 30% to 60% while for the southern area this increase is between 15% and 25% for CF 850 , Table 3. For CF 2050 , this increase is signi cantly smaller and ranges from 2% to 10% in the northern area, and for central and southern areas, it ranges from 1% to 3%. Tölle et al., (2018) also recorded air cooling during the summer and warming during the winter season for mid-Europe due to the in uence of increasing forest cover in their research. This cooling effect we obtained is consistent with remote and surface observations in Central Europe (Alkama and Cescatti, 2016;Bright et al., 2017). Using satellite data for approximately the same period, (Tang et al., 2018) showed that there is a difference between areas under forest vegetation and open land on seasonal variations of maximum and minimum surface air temperatures on the European continent. That is, forest areas have a trend of cooling the surface air temperature during the summer season and warming it during the winter. They also showed that these trends are increasing as we go from low to high latitudes, which is in line with our results because we also got a bigger trend of cooling in the northern than in the central area for the summer season, Fig. 3. While for the southern area we got a warming trend which is again in line with Tölle et al., (2018) and their results for surface air heating in southern Europe due to the increase of forest cover. For similar seasonal period Heck et al. (2001) have concluded that an increase in forest cover leads to a cooler spring and warmer summer which are partially contradictory with our results, that is, we obtained that in the summer season there was warming but with a tendency of further cooling.

Conclusion
The change in the concentration of CF types that we performed in our study unequivocally shows that it led to a change in MST 2m and MST lev both at the local and regional scales. These changes are caused by a change in the energy balance on the surface that we obtained as a result of a decrease and increase in albedo. This increase in albedo is caused by a signi cant increase in forest cover and a decrease in agricultural and grass regions with CF 850 , while this change is smaller with CF 2050 . Due to the change, we have made in CF, there was a change in the surface energy balance caused by the changes in transport of the ux of sensible heat from the surface layers into the atmosphere.
Depending on the area this led to a tendency of heating or cooling of surface air for the summer season. We got a cooling trend in the northern region of the Pannonian basin which is also the largest and amounts -0.3 [ 0 C/year]. In the central area, we also got a cooling trend but it is less intense and amounts to about -0.1 [ 0 C/year], while we got a warming trend in the southern region which amounts to about 0.1 [ 0 C/year]. At the regional scale, we also got a trend of surface air cooling of approx -0.1 [ 0 C/year]. The cooling effect that we got for the summer season is mainly due to the increase in albedo, which occurs due to the increase in cloudiness above the areas covered with temperate-type forest vegetation and is consistent with recent research on the subject.

Declarations
Funding (this research is not funded by any institution or organization)