Aridity index (AI), defined as the ratio of annual potential evapotranspiration to annual precipitation, has been widely applied in dividing climate regimes and monitoring drought events. Investigating variation of AI and the role of climate variables are thus of great significant for managing agricultural water resource and maintaining regional ecosystem stability. In this study, with the well-corrected precipitation records and the optimized parameters in estimating solar radiation, we investigated the variation of AI and its climatic attribution in the Southwest China using the observed climate records from 135 meteorological stations during the period of 1993–2015. The results showed that the AI increased significantly (0.0053 year⁻¹, p < 0.05) from 1993 to 2015 and abruptly increasing occurred around 2002. Approximately 85% of stations showed an increasing trend, while 24% of stations reached up to a significant increasing level, which demonstrated that the Southwest China was getting drier for the recent 20 years. Overall, the AI was most sensitive to maximum air temperature and precipitation with values of 1.11 and −1.00, respectively. The following factors are vapor pressure, solar radiation, minimum temperature, and wind speed. The declining precipitation and ascending maximum air temperature were the key variables in dominating the increase in aridity index during the period of 1993–2015, which contribute 38.98% and 36.26% of AI variation, respectively. Our results highlighted the influence of climate variability on aridity in the Southwest China.

Reference evapotranspiration (ET₀) is an important parameter of agricultural activity and hydro-meteorological studies. In this study, we used the Food and Agriculture Organization Penman–Monteith equation to evaluate ET₀, and to investigate the change point of ET₀ in the Loess Plateau region from 1960 to 2013. The results showed that a change point was detected at approximately the year 1990 for annual ET₀ series from 108 meteorological stations using Cramer’s statistical test. The annual ET₀ decreased significantly (p < 0.05) by −1.22 mm year⁻¹ from 1960 to 1990, especially during the summer months, which contributed the most to the total annual reduction, while it increased significantly (p < 0.001) by 1.15 mm year⁻¹ from 1991 to 2013, with the spring months contributing the most. The ET₀ and its trend in the five integrated management divisions of the Loess Plateau have significant spatial heterogeneity. The highest and lowest ET₀ were found in the third and fifth divisions from 1960 to 2013. The ET₀ decreased significantly in the first division and increased in the fourth divisions (1960–1990 and 1991–2013, respectively). Using differential equations to quantitatively evaluate the contribution of various factors, the wind speed was mostly responsible for the variability in the ET₀ trend from 1960 to 1990, followed by solar radiation and vapor pressure. The positive effect of air temperature on ET₀ trend was offset by the other three factors, and the combined effects of the four climatic variables led to the decrease in the ET₀ trend. However, the rapidly increasing air temperature became the predominant factor in the change in the ET₀ trend after 1990. A spatiotemporal variation of predominant contribution to the ET₀ trend was identified. The temperature dominant region changed from the third to the first division for 1960–1990 and 1991–2013. Radiation did not change. Vapor pressure changed from the first to the fourth and the wind speed changed from the forth to the third. This study could contribute to a better understanding of the response of the spatial and temporal variation of ET₀ to increased climate change. Additionally, this research also provides scientific support to regional planning and management.

Accurate assessment of the extent of crop distribution and mapping different crop types are essential for monitoring and managing modern agriculture. Medium and high spatial resolution remote sensing (RS) for Earth observation and deep learning (DL) constitute one of the most major and effective tools for crop mapping. In this study, we used high-resolution Sentinel-2 imagery from Google Earth Engine (GEE) to map paddy rice and winter wheat in the Bengbu city of Anhui Province, China. We compared the performance of different popular DL backbone networks with the traditional machine learning (ML) methods, including HRNet, MobileNet, Xception, and Swin Transformer, within the improved DeepLabv3+ architecture, Segformer and random forest (RF). The results showed that the Segformer based on the combination of the Transformer architecture encoder and the lightweight multilayer perceptron (MLP) decoder achieved an overall accuracy (OA) value of 91.06%, a mean F1 Score (mF1) value of 89.26% and a mean Intersection over Union (mIoU) value of 80.70%. The Segformer outperformed other DL methods by combining the results of multiple evaluation metrics. Except for Swin Transformer, which was slightly lower than RF in OA, all DL methods significantly outperformed RF methods in accuracy for the main mapping objects, with mIoU improving by about 13.5~26%. The predicted images of paddy rice and winter wheat from the Segformer were characterized by high mapping accuracy, clear field edges, distinct detail features and a low false classification rate. Consequently, DL is an efficient option for fast and accurate mapping of paddy rice and winter wheat based on RS imagery.

Global warming and its associated changes in temperature and precipitation have significantly affected the ecosystem in Southwest China, yet studies that integrate temperature and precipitation changes are inadequate for quantitatively assessing the impacts of extreme events on ecosystems. In this study, the return period of concurrent climate extremes characterized by precipitation deficit and extreme temperature and the spatial and temporal dynamic patterns of their impacts on ecosystems were assessed by using high-precision temperature and precipitation data, as well as NDVI and NPP data collected for the 1985–2015 period. The results show that the 2009 concurrent event had a return period of about 200 years. The return periods of individual climate factors are significantly overestimated or underestimated. Concurrent events significantly reduced the spring and annual Normalized Difference Vegetation Index (NDVI) and net primary productivity (NPP) in Southwest China. The magnitude of the reduction in vegetation greenness and productivity increased with the intensity of concurrent events. Concurrent events beginning in autumn 2009 reduced spring NDVI and NPP by 8.8% and 23%, and annual NDVI and NPP by 2.23% and 7.22%, respectively. Under future climate scenarios, the return period of concurrent events could be significantly shortened, which would have a more severe impact on regional ecosystems.