A Trend Analysis of Leaf Area Index and Land Surface Temperature and Their Relationship from Global to Local Scale

: Although the way in which vegetation phenology mediates the feedback of vegetation to climate systems is now well understood, the magnitude of these changes is still unknown. A thorough understanding of how the recent shift in phenology may impact on, for example, land surface temperature (LST) is important. To address this knowledge gap, it is important to quantify these impacts and identify patterns from the global to the regional scale. This study examines the trend and linear regression modeling of the leaf area index (LAI) and LST derived from the moderate resolution imaging spectroradiometer (MODIS) data, speciﬁcally to assess their spatial distribution and changing trends at the continental and regional scales. The change detection analysis of interannual variability in the global LAI and LST between two periods (2003–2010 and 2011–2018) demonstrates more positive LAI trends than negative, while for LST most changes were not signiﬁcant. The relationships between LAI and LST were assessed across the continents to ascertain the response of vegetation to changes in LST. The regression between LAI and LST was negative in Australia ( R 2 = 0.487 ***), positive but minimal in Africa ( R 2 = 0.001), positive in North America ( R 2 = 0.641 ***), negative in Central America ( R 2 = 0.119), positive in South America ( R 2 = 0.253 *) and positive in Europe ( R 2 = 0.740 ***). Medium temperatures enhance photosynthesis and lengthen the growing season in Europe. We also found a signiﬁcant greening trend in China ( trendp = 0.16 ***) and India ( trendp = 0.13 ***). The relationships between LAI and LST in these most prominent greening countries of the world are R 2 = 0.06 and R 2 = 0.25 for China and India, respectively. Our deductions here are twofold—(1) In China, an insigniﬁcant association appeared between greening trend and temperature. (2) In India, the signiﬁcant greening trend may be a factor in lowering temperatures. Therefore, temperature may stabilize if the greening trend continues. We attribute the trends in both countries to the di ﬀ erent land use management and climate mitigation policies adopted by these countries.

Land 2020, 9, 388 3 of 17 2003-2018 at the continental scale, (ii) find out the relationship between the LAI and LST at the continental scale, and (iii) explore the relationship between the LAI and LST, mean air temperature, total precipitation and specific humidity over the most prominent greening countries of the world (China and India) during the investigation period. The novelty of this research is the application of cloud computing through the Google Earth Engine (GEE) to assess the trends of LAI and LST, and the associations between them, from the global to local scale, with a focus on China and India.

Data
The Aqua daytime LST (MYD11A2 V6) from 2003 to 2018 was used in the study. The analysis covers data for the entire year. Cloud-contaminated LST was removed from this level of LST product by the data provider. MODIS is the main instrument aboard both the Terra and Aqua satellites [29]. Each pixel value of the LST data is the mean of an 8-day period [30]. The LAI product, "MCD15A3H V6" level 4, was utilized for this study. It is a composite of 4-day data with a spatial resolution of 500 m (Table 1). For product generation, the algorithm selected the best pixel available from the Terra and Aqua sensors over 4 days [31]. The specific humidity NOAA CFSV2 [32], mean air temperature and total precipitation ERA5 [33] products of India and China were used to find their relationship with LAI in order to assess how changes in greening impact these climatic variables.

Data Pre-Processing
Earth Engine Code Editor [34] was used to process and analyze the LST and LAI data from 2003 to 2018. Filtration in GEE was used to select the data collection ID, the period of study, and the bands of interest ( Table 1). The LST and LAI values were multiplied by scale factors 0.02 and 0.1, respectively [30,31]. The LST and LAI time series data were prepared based on the mean pixels value of 2003-2010 and 2011-2018, as the first and second period, respectively. Then, the values of LST and LAI were extracted for the globe, each continent, and the countries of interest (Tables A1-A3). For the analysis, a 1 km resolution was adopted. The LAI, air, total precipitation and specific humidity products were resampled in GEE to 1 km by the nearest neighbor method. Figures were created in the R program [35] by using packages such as "ggpubr", "tidyverse", "dplyr" and "ggplot2" [36][37][38][39].

Change Detection Method
For detecting changes in LAI and LST from 2003 to 2018, the data were divided into two periods: from 2003-2010 (reference period) and from 2011-2018 (experimental period). Simple subtraction image differencing was used in order to investigate the spatiotemporal variations in vegetation and temperature dynamics (Equation (1)). Note that only the statistically significant changes (p < 0.05) were shown and insignificant changes (p ≥ 0.05) were masked out. To investigate the influence of increasing LAI in the most greening countries [22], trends and relationships between LAI and LST were assessed in both China and India.
where image2 is the average LST or LAI from 2011 to 2018, and image1 is the average LAT or LAI from Land 2020, 9,388

Statistical Analyses
Two basic statistical operations (trend and regression analyses) were performed to evaluate changes in LAI and LST and their relationships at the continental and regional scales. All statistical operations were performed in R program.

Trend Analysis
Annual trend and trend during the period (trendp) of the study were calculated by using Equation (2), the Theil-Sen approach (TSA) in the "zyp" package [40]. The advantage of this trend estimator is that it yields accurate confidence intervals and it is resistant to outliers [41]. The significance of the trend was examined by using a Kendall test [42] where b is the estimation of the slope of the trend for LAI and LST, and x l is the lth observation. Trend: trendp/number of years.
. . x n represent n LAI and LST points, x j represents the LAI and LST points at time j and S is the Mann-Kendall test. A negative slope implies decreasing LAI and LST, while a positive slope shows increasing LAI or LST. A value of 0 shows no trend.

Relationship between LAI and LST
A simple linear relationship was used to assess the relationship between the LAI and climatic variables. The coefficient of determination (R-squared) values were extracted from Equation (4). R-squaring was used to assess the strength of the relationship between the LAI and LST, mean air temperature, total precipitation and specific humidity. The significance level was evaluated at α0.05. LAI was used as the response variable while climatic variables were used as the predictor variables where y* is the mean of y[i] if there is an intercept, and it is zero otherwise.

Results
We present results of the trends and relationships between LAI and LST at the global and continental scales, and for most greening countries, in the following subsections.

Mean of LAI and LST for 2018
The LAI product ("MODIS/006/MCD15A3H") was used to compute the spatial distribution of the global mean LAI for 2018 ( Figure 1a). In 2018, large greening was observed around the southeast region of North America, the Amazon in South America, Central Africa, and Southeast Asia. Likewise, the LST product ("MODIS/006/MYD11A2") was used to compute the spatial distribution of the global mean LST for 2018 ( Figure 1b). The regions with higher LST include Southwest America, northeastern Amazon, North Africa, West Africa, Australia, and some parts of the southwest of Asia. All regions Land 2020, 9,388 6 of 17 with the largest greening, as observed in Figure 1a, consistently demonstrate a moderate LST of 302 to 307 K.

Mean of LAI and LST for 2018
The LAI product ("MODIS/006/MCD15A3H") was used to compute the spatial distribution of the global mean LAI for 2018 ( Figure 1a). In 2018, large greening was observed around the southeast region of North America, the Amazon in South America, Central Africa, and Southeast Asia. Likewise, the LST product ("MODIS/006/MYD11A2") was used to compute the spatial distribution of the global mean LST for 2018 ( Figure 1b). The regions with higher LST include Southwest America, northeastern Amazon, North Africa, West Africa, Australia, and some parts of the southwest of Asia. All regions with the largest greening, as observed in Figure 1a, consistently demonstrate a moderate LST of 302 to 307 K.  (Figure 2a,b). Note that only the statistically significant changes (p < 0.05) are shown. Globally, the spatiotemporal variations of LAI with positive trends were more widespread than the negative trends during the study period (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018). In fact, the areas with the most notable differences in LAI between the reference and experimental period (demonstrating an increasing trend), were in South and North America, Europe, Central Africa, Southeast and Southwest Asia, and Australia ( Figure 2a). The observed LAI increases ranged between 0.016 and 3.149 m 2 /m 2 . LAI decreased in parts of North America, Africa, and the northeast of South America. This is demonstrated especially in South America by negative values (−4.669 to −1.545 m 2 /m 2 ) over a large proportion of its vegetated surface since 2003. Changes in mean LST between the reference and experimental periods were very much insignificant across most of the globe (Figure 2b). Areas with increased LST were observed in central South America (the Amazon), some parts of Europe and Southwest Asia. During this period, the increase in LST ranged between 0.089 and 2.931, while negative changes ranged from −3.699 to −0.805 K, as seen in the Amazon and eastern Central Africa (Figure 2b). northeast of South America. This is demonstrated especially in South America by negative values (−4.669 to −1.545 m 2 /m 2 ) over a large proportion of its vegetated surface since 2003. Changes in mean LST between the reference and experimental periods were very much insignificant across most of the globe (Figure 2b). Areas with increased LST were observed in central South America (the Amazon), some parts of Europe and Southwest Asia. During this period, the increase in LST ranged between 0.089 and 2.931, while negative changes ranged from −3.699 to −0.805 K, as seen in the Amazon and eastern Central Africa (Figure 2b).

Trends of Continental LAI
The temporal pattern of trends in LAI for different continents is shown in Figure 3a

Trends of Continental LAI
The temporal pattern of trends in LAI for different continents is shown in Figure 3a-h. Both the annual trend and trend during the period (trendp) were estimated for each continent. In all cases, during the study period (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018), all continents demonstrated a greening trend even though the rate of change in LAI varied. The year to year trend of LAI ranged between 0.002 and 0.005, while for the trend period it was 0.026 to 0.081. The Asian continent had the most significant trendp (0.081). Overall, the average LAI for all the continents exhibited a significant greening over the study period. Land 2020, 9, x FOR PEER REVIEW 7 of 16 the trend period it was 0.026 to 0.081. The Asian continent had the most significant trendp (0.081). Overall, the average LAI for all the continents exhibited a significant greening over the study period.

Trend of Continental LST
The temporal trends in LST for different continents can be seen in Figure 4a-h. Both trend and trendp were estimated for each continent. The changing trends of LST varied by continent. For example, in Australia, no trend was observed, however there was a very sharp drop in LST in 2010. In Africa (trendp = 0.229) and North America (trendp = 0.756), an increasing trend was observed. In contrast, a decreasing trend of LST was observed for Central America. Although the increasing trend of LST was greater in South America (trendp = 0.320) compared to Africa, Europe (trendp = 0.992) and Asia (trendp = 0.902) had the most significant increase in LST over the study period.

Trend of Continental LST
The temporal trends in LST for different continents can be seen in Figure 4a-h. Both trend and trendp were estimated for each continent. The changing trends of LST varied by continent. For example, in Australia, no trend was observed, however there was a very sharp drop in LST in 2010. In Africa (trendp = 0.229) and North America (trendp = 0.756), an increasing trend was observed. In contrast, a decreasing trend of LST was observed for Central America. Although the increasing trend of LST was greater in South America (trendp = 0.320) compared to Africa, Europe (trendp = 0.992) and Asia (trendp = 0.902) had the most significant increase in LST over the study period.
Land 2020, 9, x FOR PEER REVIEW 7 of 16 the trend period it was 0.026 to 0.081. The Asian continent had the most significant trendp (0.081). Overall, the average LAI for all the continents exhibited a significant greening over the study period.

Trend of Continental LST
The temporal trends in LST for different continents can be seen in Figure 4a-h. Both trend and trendp were estimated for each continent. The changing trends of LST varied by continent. For example, in Australia, no trend was observed, however there was a very sharp drop in LST in 2010. In Africa (trendp = 0.229) and North America (trendp = 0.756), an increasing trend was observed. In contrast, a decreasing trend of LST was observed for Central America. Although the increasing trend of LST was greater in South America (trendp = 0.320) compared to Africa, Europe (trendp = 0.992) and Asia (trendp = 0.902) had the most significant increase in LST over the study period.

LAI and LST Continental Relationship
The relationship between LAI and LST was assessed for each continent to ascertain the response of vegetation to changes in LST (Figure 5a-h). In Australia, a moderate inverse relationship was observed between the LAI and LST (R 2 = 0.487). No significant relationship was found in Africa (R 2 = 0.001), while a strong/significant relationship was observed in North America's LST (R 2 = 0.641). This demonstrates high sensitivity of photosynthetic vegetation to LST on this continent. In Central America, a weak inverse relationship was found (R 2 = 0.119). Again, a weak but significant relationship was observed for South America (R 2 = 0.253). A strong positive relationship was observed in Europe (R 2 = 0.740), and this was the most significant relationship compared with all the other continents. This may indicate that vegetation responds strongly to changes in LST in Europe. In Asia, a moderate positive relationship was observed (R 2 = 0.442). The relationship between the average LAI and the LST for all continents showed a positive relationship (R 2 = 0.306).

LAI and LST Continental Relationship
The relationship between LAI and LST was assessed for each continent to ascertain the response of vegetation to changes in LST (Figure 5a-h). In Australia, a moderate inverse relationship was observed between the LAI and LST (R 2 = 0.487). No significant relationship was found in Africa (R 2 = 0.001), while a strong/significant relationship was observed in North America's LST (R 2 = 0.641). This demonstrates high sensitivity of photosynthetic vegetation to LST on this continent. In Central America, a weak inverse relationship was found (R 2 = 0.119). Again, a weak but significant relationship was observed for South America (R 2 = 0.253). A strong positive relationship was observed in Europe (R 2 = 0.740), and this was the most significant relationship compared with all the other continents. This may indicate that vegetation responds strongly to changes in LST in Europe. In Asia, a moderate positive relationship was observed (R 2 = 0.442). The relationship between the average LAI and the LST for all continents showed a positive relationship (R 2 = 0.306).   (Figure 6b). The LST for the two also varied. In China, there was no significant trend (trendp = −0.13) in LST (Figure 6c). In India, there was a negative trend (trendp = −0.69) in LST (Figure 6d).

Relationship between LAI and LST in China and India
The relationships between the LAI and LST for two countries were assessed to ascertain if LAI responds to change in LST (Figure 6e,f). Despite the significant positive change in LAI over the study period in China, the relationship between the LAI and LST was not significant (R 2 = 0.063). This is possibly due to there being no significant change in LST for this particular country over the study period. Conversely, although India showed a weak inverse relationship between the LST and LAI (R 2 = 0.254), this relationship is stronger than what was found for China. This is because the country demonstrates a decreasing LST trend especially, from 2010 to 2015 (Figure 6d).

Relationship between LAI and LST in China and India
The relationships between the LAI and LST for two countries were assessed to ascertain if LAI responds to change in LST (Figure 6e,f). Despite the significant positive change in LAI over the study period in China, the relationship between the LAI and LST was not significant (R 2 = 0.063). This is possibly due to there being no significant change in LST for this particular country over the study period. Conversely, although India showed a weak inverse relationship between the LST and LAI (R 2 = 0.254), this relationship is stronger than what was found for China. This is because the country demonstrates a decreasing LST trend especially, from 2010 to 2015 (Figure 6d).

Relationship between LAI and Air Temperature, Humidity and Precipitation in China and India
The relationships between the LAI and air temperature, specific humidity, and total precipitation for China and India were assessed to determine if climatic variables are associated with changes in LAI (Figure 7a-c). Despite the significant positive change of LAI over the study period in India, the relationship between the LAI and selected climatic variables (mean air temperature, annual specific humidity, and the annual total precipitation) was not significant. In China, a weak positive relationship between LAI and air temperature (R 2 = 0.33) and specific humidity (R 2 = 0.336) was

Relationship between LAI and Air Temperature, Humidity and Precipitation in China and India
The relationships between the LAI and air temperature, specific humidity, and total precipitation for China and India were assessed to determine if climatic variables are associated with changes in LAI (Figure 7a-c). Despite the significant positive change of LAI over the study period in India, the relationship between the LAI and selected climatic variables (mean air temperature, annual specific humidity, and the annual total precipitation) was not significant. In China, a weak positive relationship between LAI and air temperature (R 2 = 0.33) and specific humidity (R 2 = 0.336) was observed. The country displayed increasing air temperature and specific humidity in association with increasing LAI during the study period (Figure 7a,c). Meanwhile, no relationship was observed between total precipitation and LAI trend in both India and China (Figure 7b).

Discussion
This study examined the trend and linear regression modeling of level 4 LAI and LST derived from MODIS data, specifically to assess their spatial distribution and changing trends at the continental and regional scales. We employed a change-detection approach to quantify the interannual variability in the global LAI and LST between two periods (2003-2010 and 2011-2018). To address some of the peculiarities among regions, such as that of the most prominent greening countries in the world, the relationships between the LAI and LST over China and India were investigated. An investigation into the changing trends of LAI and LST, and their interannual variability at the regional and global levels, could reveal ecological processes and biosphereatmosphere interactions, setting the basis for monitoring vegetation and climate feedback [12,19,28].
The current study highlights the spatial details of the changing global LAI and LST by mapping the spatial patterns of linear trends between two periods (2003-2010 and 2011-2018). Globally, the spatiotemporal variations in LAI demonstrate more positive trends than negative (Figure 2a). This is consistent with the findings of previous research that observed an increase in growing season integrated LAI (greening) from 25% to 50% of the global vegetated area during 1982-2009 [12]. In fact, the areas with notable differences in LAI between the reference period (2003-2010) and 2011-2018, showing an increasing trend, are in South and North America, Europe, Central Africa, and Southeast and Southwest Asia. However, Eurasia demonstrates a more subtle greening than North America [43]. Recent evidence suggests that some of these regions (e.g., China) have an increased greening trend [22], and historically, most of these regions have records of growing LAI [11,12,44]. This study found greening and browning trends of the global LAI in the ranges of 0.016 to 3.15 and To understand the spatial pattern of the global LAI and LST more effectively, we assessed their temporal trends and relationships per continent. The trend of the LAI and LST varied by continent. In general, all continents demonstrate a greening trend. The most significant greening trend (trendp = 0.081) was found in Asia. This can be explained by the increase in land use management activities (e.g., agriculture) in this part of the world [22]. For the LST, the most significant trends were found in South America, Europe and Asia. The relationship between the LAI and LST further indicates how differences in LST within continents affect the LAI trend. The strong inverse relationship between the LAI and the LST in Australia (Figure 5), and the absence of trend in Africa, are likely due to the increasing anthropogenic activities (e.g., fires, deforestation, drought, etc.) and high interannual

Discussion
This study examined the trend and linear regression modeling of level 4 LAI and LST derived from MODIS data, specifically to assess their spatial distribution and changing trends at the continental and regional scales. We employed a change-detection approach to quantify the interannual variability in the global LAI and LST between two periods (2003-2010 and 2011-2018). To address some of the peculiarities among regions, such as that of the most prominent greening countries in the world, the relationships between the LAI and LST over China and India were investigated. An investigation into the changing trends of LAI and LST, and their interannual variability at the regional and global levels, could reveal ecological processes and biosphere-atmosphere interactions, setting the basis for monitoring vegetation and climate feedback [12,19,28].
The current study highlights the spatial details of the changing global LAI and LST by mapping the spatial patterns of linear trends between two periods (2003-2010 and 2011-2018). Globally, the spatiotemporal variations in LAI demonstrate more positive trends than negative (Figure 2a). This is consistent with the findings of previous research that observed an increase in growing season integrated LAI (greening) from 25% to 50% of the global vegetated area during 1982-2009 [12]. In fact, the areas with notable differences in LAI between the reference period (2003-2010) and 2011-2018, showing an increasing trend, are in South and North America, Europe, Central Africa, and Southeast and Southwest Asia. However, Eurasia demonstrates a more subtle greening than North America [43]. Recent evidence suggests that some of these regions (e.g., China) have an increased greening trend [22], and historically, most of these regions have records of growing LAI [11,12,44]. This study found greening and browning trends of the global LAI in the ranges of 0.016 to 3.15 and −4.669 to −1.545, respectively. Conversely, changes in LST between 2003 and 2010 and 2011 and 2018 were very much insignificant across a large proportion of the globe (Figure 2b). Areas with positive increases in LST were observed in Central America, the Amazon, some parts of Europe and Southwest Asia. The increases and decreases in LST ranged from 0.089 to 2.931 and −3.699 to −0.805 K, respectively.
To understand the spatial pattern of the global LAI and LST more effectively, we assessed their temporal trends and relationships per continent. The trend of the LAI and LST varied by continent. In general, all continents demonstrate a greening trend. The most significant greening trend (trendp = 0.081) was found in Asia. This can be explained by the increase in land use management activities (e.g., agriculture) in this part of the world [22]. For the LST, the most significant trends were found in South America, Europe and Asia. The relationship between the LAI and LST further indicates how differences in LST within continents affect the LAI trend. The strong inverse relationship between the LAI and the LST in Australia (Figure 5), and the absence of trend in Africa, are likely due to the increasing anthropogenic activities (e.g., fires, deforestation, drought, etc.) and high interannual variability [45][46][47]. In contrast, LAI and LST in North America (R 2 = 0.64 ***), Europe (R 2 = 0.4 ***), and Asia (R 2 = 0.64 ***) have indicated significant relationships ( Figure 5). The socio-economic and biophysical characteristics of these regions differ compared to other regions, such as medium temperature in Europe, which enhances plants' photosynthetic activities. Zhu et al. [12] and Zhu et al. [12] disclosed that in the northern high latitudes, rising temperatures, which enhance photosynthesis and lengthen the growing season, are the dominant climatic factor contributing to greening trends. In Asia, for example, increased agricultural activities in countries such as China and India contribute substantially to greening trends. The conversion of the relationship between LST and LAI from positive to inverse in different regions could be attributed to decreasing photosynthesis when the temperature exceeds the temperature optimum of photosynthesis (20 to 30 • C) [48,49].
From our results, we observed a significant greening trend, with trendp = 0.16 *** and trendp = 0.13 *** (Figure 6a,b) for China and India, respectively. These results together confirmed the findings of a recent study by Chen et al. (2019), despite a slight difference in the temporal observations of the satellite data assessed by the two studies. However, this study found an insignificant LST trend, R 2 = 0.002, for China (Figure 6c), and a weak negative trend (R 2 = 0.13) for India (Figure 6d). In addition to quantifying the relationship between the LAI and LST in these countries, this study also provides an important insight into how the vegetation dynamics in these greening parts of the globe affect the LST derived from data from a 16-year MODIS time series. Since the photosynthesis of vegetation requires an optimum temperature, changes in temperature also affect vegetation dynamics. An insignificant relationship was found between LAI and LST in China. In contrast to the relationship observed in India, a weak inverse (R 2 = 0.24) was observed. A recent study by Yu et al. [50] studied the impact of recent vegetation greening on temperature and precipitation over China. The study found a surface cooling of 0.11 • C and a precipitation increase of 0.02 mm/d during summer in response to China's widespread greening. The study assessed the regional variability of climatic responses to greening. Northern China showed significant cooling, while no significant temperature changes were observed in southern China. They discussed that vegetation greening is influencing the occurrence of extreme climate events by decreasing the number of heatwave days (by 1-3 days, particularly in northern China), reducing drought events in northern China, and minimizing the risk of floods in southern China.
The differences between China and India, in terms of their changes in LAI and the relationship between LAI and LST, could be due to the differences in the nature of land use/cover (e.g., forest or cropland) that were to found to contribute substantially to the greening of these two countries. This is because the feedback of vegetation to climate ais mediated by phenology, which is largely influenced by factors such as albedo, surface roughness and canopy conductance [25]. China accounts for 25% of the global net increase in leaf area, and a large proportion of the contribution comes from forests (42%) and croplands (32%). This assists us in stabilizing the climate by lowering the temperature (Yu et al. [50]), while in India, the greening is mostly from the croplands (82%), and only a negligible proportion is from forests (4.4%) [22]. The phenomenon of an increase in LAI leading to decreasing LST in India can be explained based on the influence of all the factors mentioned above, but of particular importance is the surface roughness length. Richardson et al. [25] disclosed that increases in surface roughness length are associated with an increase in the efficiency with which sensible heat is transferred from the surface, but surface roughness length varies according to seasonal changes in canopy structure. In agricultural crops, surface roughness increases substantially over the course of the growing season, as there is transition from bare ground to tall plants. In India, the greening trend was high from 2011 to 2018. The poor relationship between the LAI and precipitation may be due to the influence of ecosystem management practices in China and India [12]. We attribute these changes in greening to land use management and the climate mitigation options adopted by these countries.

Conclusions
This study presented the trends of and linear regression modeling for level 4 LAI and LST derived from MODIS data, specifically to assess their spatial distribution and their changing trends at the continental and regional scales. For example, China and India were determined as the most greening parts of the globe. The following conclusions can be drawn: The current study highlights the spatial details of the changing global LAI and LST by mapping the spatial patterns of linear trends between two periods (2003-2010 and 2011-2018). Globally, the spatiotemporal variations in LAI between both periods demonstrate more positive trends than negative; II.
The continental patterns of LAI and LST trends, and their linear relationships, indicated that LAI and LST varied by continent. In general, during the study period (2003-2018), all continents demonstrated a greening trend, with the most significant greening trend (trendp = 0.081) in Asia. The relationships between the LAI and LST in North America (R 2 = 0.64 ***), Europe (R 2 = 0.4 ***) and Asia (R 2 = 0.64 ***) were very significant; III.
We observed a significant greening trend for China and India, respectively (trendp = 0.16 *** and trendp = 0.13 ***), from 2003 to 2018. These results together confirmed the results presented by Chen et al. [22], despite a slight difference in the temporal observations of the satellite data product assessed by the two studies. Besides this, the study also found an insignificant LST trend for China (trendp = -0.13) and a weak negative trend for India (trendp = -0.69). IV.
An insignificant relationship between the LAI and LST in China (R 2 = 0.057) is indicative of the impact of the positive greening trend on LAI. In contrast, a weak inverse relationship (R 2 = 0.24) was observed for India and is attributed to a strong positive change in LAI in recent times, in this particular country, as observed in this study. V.
A weak positive relationship between LAI and mean air temperature and specific humidity was observed in China. However, no relationship was found between LAI and selected climatic variables in India.