Daily Global Solar Radiation in China Estimated From High‐Density Meteorological Observations: A Random Forest Model Framework

Abstract

Accurate estimation of the spatiotemporal variations of solar radiation is crucial for assessing and utilizing solar energy, one of the fastest‐growing and most important clean and renewable resources. Based on observations from 2,379 meteorological stations along with scare solar radiation observations, the random forest (RF) model is employed to construct a high‐density network of daily global solar radiation (DGSR) and its spatiotemporal variations in China. The RF‐estimated DGSR is in good agreement with site observations across China, with an overall correlation coefficient (R) of 0.95, root‐mean‐square error of 2.34 MJ/m2, and mean bias of −0.04 MJ/m2. The geographical distributions of R values, root‐mean‐square error, and mean bias values indicate that the RF model has high predictive performance in estimating DGSR under different climatic and geographic conditions across China. The RF model further reveals that daily sunshine duration, daily maximum land surface temperature, and day of year play dominant roles in determining DGSR across China. In addition, compared with other models, the RF model exhibits a more accurate estimation performance for DGSR. Using the RF model framework at the national scale allows the establishment of a high‐resolution DGSR network, which can not only be used to effectively evaluate the long‐term change in solar radiation but also serve as a potential resource to rationally and continually utilize solar energy.

Dual Effects of Synoptic Weather Patterns and Urbanization on Summer Diurnal Temperature Range in an Urban Agglomeration of East China

“…Compared with traditional statistical methods, RF has a better fitting effect on nonlinear data, and can analyze the importance of variables. The specific steps of the RF model construction process are as follows (Zeng et al, 2020): 1) In the N total samples, n times of replacement are randomly selected, n new training sets are obtained, and the unextracted part constitutes "out-of-bag" (OOB) data; 2) Each training set generates a decision tree, each node of the decision tree selects mtry from the independent variables, and branches grow according to the principle of minimum node impurity; 3) Repeat step (2) n times to obtain n decision trees to form a random forest; 4) The result of the random forest is the result obtained by the simple averaging method for each decision tree, and the prediction accuracy is determined by the average OOB of each decision tree.…”

“…Based on the RF model, we obtain the DTR difference between urban and rural areas and explore and analyze the importance of each factor influencing the DTR U-R . We validated the RF model using a 10-fold cross validation (CV) method to repeatedly estimate the expected model performance based on each subset of training data in general during prediction (Wang H. et al, 2019;Yang X. et al, 2019;Zeng et al, 2020). The method of 10-fold CV involved cutting the sample into 10 subsets, reserving one subset for testing the accuracy of the model, and using the other nine subsets for training the model.…”

“…So, we further employed the RF model to check the drivers of △T max (Figures 8D-F), that is, T max was selected as the dependent variable, and △LST D , △LST N , △AHF, △NDVI, △Altitude, △rainfall, and △SSD as independent variables. In addition to the direct retrieval of solar radiation, SSD also includes the influence of aerosols, cloud cover, and precipitation on the daily global solar radiation (Wang K. et al, 2012Zeng et al, 2020).…”

“…As shown in Figure 8C, △T max is more important to DTR U-R than △T min , implying that the comprehensive effect of urbanization (changed land use and soil water content) and aerosol pollution (changed solar radiation and cloud fraction) has a more obvious impact on daytime air temperature in urban areas, relative to rural areas (Dai et al, 1999;Sun et al, 2019;Yang et al, 2020b). The development of urbanization results in elevated concentrations of the aerosols produced by human activities, which leads to increased cloud cover (Wang K. et al, 2012Yang et al, 2020b;Zeng et al, 2020). The existence of aerosols and clouds can have an impact on solar radiation.…”

“…The most important environmental factors in this regard are Altitude, AHF, and NDVI, followed by rainfall, SSD, LST N and LST D ,. Urban-rural differences of AHF play an important role in the magnitude of DTR U-R (Wang K. et al, 2012Zeng et al, 2020). The development of urbanization will also change the DTR by changing the nature of the urban underlying surface.…”