Forest fires have a significant impact on terrestrial ecosystems, leading to harm to biodiversity and environment. To mitigate the ecological damage caused by forest fires, it was necessary to develop prediction models of fire risk. In this study, by evolving the optimal architecture and parameters using the particle swarm optimization (PSO) algorithm, a convolutional neural network (CNN) deep learning model was proposed to predict forest fire risk on a national scale. Utilizing fire data and fire risk factors from 2001 to 2020 in China, the PSO-CNN-based deep learning model (PSO-CNN) was utilized and tested. Compared to logistic regression, random forest, support vector machine, k-nearest neighbors, and CNN models, the PSO-CNN model exhibited superior performance with an accuracy of 82.2% and an AUC value of 0.92. These results clearly highlighted the effectiveness of the PSO-CNN model in enhancing the accuracy of forest fire prediction. Furthermore, the forest fire risk prediction level estimated by the proposed model on a national scale for the entire country was mostly consistent with actual fire data distribution, indicating its potential to be used as an important direction for deep learning in forest fire prediction research.

Evaporation is a key element of the water and energy cycle and is essential in determining the spatial and temporal variations of meteorological elements. In particular, evaporation is crucial for thoroughly understanding the climate variations of a region. In this study, we discussed evaporation, precipitation, and temperature by adopting Linyi City in Shandong Province, China, which is an important agricultural region, as a research case. Linear regression analysis, the empirical orthogonal decomposition function, and the Morlet wavelet function were used to reveal the trends, spatiotemporal modes, and multi-time scale characteristics of the three climate factors and provide a theoretical basis for the efficient use of climate resources in the future development of regional agriculture. Results showed that the precipitation (2.09 mm/a) and temperature (0.04 ℃/a) in Linyi City exhibited a synchronous growth trend. Conversely, evaporation (−6.47 mm/a) showed a decreasing trend and the evaporation paradox because of the considerable decrease in evaporation energy. Regional development of water-consuming agriculture in consideration of global warming is a key point for improving water use efficiency in Linyi City. In terms of spatial distribution, precipitation was dominated by the first mode wherein low precipitation was observed at the early stage, and high precipitation occurred at the late stage. The first mode was supplemented by the second mode wherein an inverse phase change occurred in the southeast-northwest direction. Large interannual fluctuations were observed only in Yinan County. Temperature exhibited a pattern of warming change with high homogeneity. Evaporation demonstrated obvious heterogeneity and was dominated by two major modes, and the difference in evaporation between Junan County and the other regions of Linyi City was large. Therefore, the local regional climate changes in Yinan and Junan should be given attention. All three meteorological elements showed interannual and interdecadal variations in the short (5 a), medium (16 a), and long (25 a) terms, with precipitation, temperature, and evaporation dominated by 16 a, 24 a, and 31 a, respectively. In the short-term future, the regional precipitation and temperature in Linyi will experience decrements that are above the multiyear average, and evaporation will increase to above the multiyear average. Given the changing trends of precipitation, temperature, and evaporation, urgent requirements for the regional development of efficient water-saving irrigation and the promotion of digital agriculture should be proposed.

As a typical lake in the early stage of eutrophication, the non-point source pollution (NSP) in Erhai Lake was highly concerned. Since 2017, protection and rescue actions have been implemented in the Erhai Lake basin which significantly improved the water quality of Erhai Lake. But little attention has been paid to the interception effects of these actions on NSP. Based on high resolution datasets, including refinement land use/land cover (LU/LC) data, China Meteorological Assimilation Driving Database for the SWAT (CMADS), and Advanced Land Observing Satellite-1 (ALOS-12) data, the interception effects of ecological ponds newly built in 2018 on NSP was discussed with the support of the soil and water assessment (SWAT) model. These three high resolution datasets can meet the needs for simulating NSP by SWAT model, which was conductive to better reveal the interception effects of ecological ponds on NSP. Besides, the pollution load of shrubbery on Total nitrogen (TN)/Total phosphorous (TP), which has never been involved in similar research works in the Erhai Lake basin, was measured. Compared the temporal and spatial variations in TN/TP concentration before and after 2018, the interception effects of ecological ponds on NSP in the Erhai Lake basin were revealed by a sharp decrease in TN/TP concentration and a delayed presence of annual peak value about 1 month late in TN concentration before and after 2018. The interception intensity of NSP was determined by the number and volume of ecological ponds in each sub-basin and their corresponding upstream.