Solar radiation is the Earth’s primary source of energy and has an important role in the surface radiation balance, hydrological cycles, vegetation photosynthesis, and weather and climate extremes. The accurate prediction of solar radiation is therefore very important in both the solar industry and climate research. We constructed 12 machine learning models to predict and compare daily and monthly values of solar radiation and a stacking model using the best of these algorithms were developed to predict solar radiation. The results show that meteorological factors (such as sunshine duration, land surface temperature, and visibility) are crucial in the machine learning models. Trend analysis between extreme land surface temperatures and the amount of solar radiation showed the importance of solar radiation in compound extreme climate events. The gradient boosting regression tree (GBRT), extreme gradient lifting (XGBoost), Gaussian process regression (GPR), and random forest models performed better (poor) prediction capabilities of daily and monthly solar radiation. The stacking model, which included the GBRT, XGBoost, GPR, and random forest models, performed better than the single models in the prediction of daily solar radiation but showed no advantage over the XGBoost model in the prediction of the monthly solar radiation. We conclude that the stacking model and the XGBoost model are the best models to predict solar radiation.

The Cellular Automata Markov model combines the cellular automata (CA) model’s ability to simulate the spatial variation of complex systems and the long-term prediction of the Markov model. In this research, we designed a parallel CA-Markov model based on the MapReduce framework. The model was divided into two main parts: A parallel Markov model based on MapReduce (Cloud-Markov), and comprehensive evaluation method of land-use changes based on cellular automata and MapReduce (Cloud-CELUC). Choosing Hangzhou as the study area and using Landsat remote-sensing images from 2006 and 2013 as the experiment data, we conducted three experiments to evaluate the parallel CA-Markov model on the Hadoop environment. Efficiency evaluations were conducted to compare Cloud-Markov and Cloud-CELUC with different numbers of data. The results showed that the accelerated ratios of Cloud-Markov and Cloud-CELUC were 3.43 and 1.86, respectively, compared with their serial algorithms. The validity test of the prediction algorithm was performed using the parallel CA-Markov model to simulate land-use changes in Hangzhou in 2013 and to analyze the relationship between the simulation results and the interpretation results of the remote-sensing images. The Kappa coefficients of construction land, natural-reserve land, and agricultural land were 0.86, 0.68, and 0.66, respectively, which demonstrates the validity of the parallel model. Hangzhou land-use changes in 2020 were predicted and analyzed. The results show that the central area of construction land is rapidly increasing due to a developed transportation system and is mainly transferred from agricultural land.

Tourism trend prediction facilitates estimation of tourism investment and revenue. Studies on tourism prediction have primarily relied on linear models and historical visitors; however, relationships between tourism trends and their factors may be nonlinear. This study constructed factors from internet search data and predicted tourism trends using a spatiotemporal framework based on the extreme gradient boosting (XGBoost) method. The study first sorted Baidu index data that is computed by weighting the search frequency. The spatial cluster analysis was conducted to incorporate spatial characteristics, and principal component analysis was further performed to identify factors. The next step derived variables using the weighted moving average method to reduce the lag effect between tourism internet search and actual behavior. We applied the proposed spatiotemporal XGBoost composite model to predict Beijing’s tourism trends. The R² scores of the simple XGBoost model, the autoregressive integrated moving average model, the spatial XGBoost model, and the spatiotemporal XGBoost composite model were 0.517, 0.625, 0.791, and 0.940, respectively. Compared to predictions from different models, the spatiotemporal XGBoost composite model has the best prediction ability. The findings also suggest that machine learning methods may not perform well without considering spatial properties, such as spatial autocorrelation and spatial heterogeneity.

Faced with the rapid growth of vector data and the urgent requirement of low-latency query, it has become an important and timely challenge to effectively achieve the scalable storage and efficient access of vector big data. However, a systematic method is rarely seen for vector polygon data storage and query taking spatial locality into account in the storage schema, index construction and query optimization. In the paper, we focus on the storage and topological query of vector polygon geometry data in HBase, and the rowkey in the HBase table is the concatenation of the Hilbert value of the grid cell to which the center of the object entity’s MBR belongs, the layer identifier and the order code. Then, a new multi-level grid index structure, termed Q-HBML, that incorporates the grid-object spatial relationship and a new Hilbert hierarchical code into the multi-level grid, is proposed for improving the spatial query efficiency. Finally, based on the Q-HBML index, two query optimization strategies and an optimized topological query algorithm, ML-OTQ, are presented to optimize the topological query process and enhance the topological query efficiency. Through four groups of comparative experiments, it has been proven that our approach supports better performance.