Subseasonal-to-seasonal (S2S) prediction of winter wheat yields is crucial for farmers and decision-makers to reduce yield losses and ensure food security. Recently, numerous researchers have utilized machine learning (ML) methods to predict crop yield, using observational climate variables and satellite data. Meanwhile, some studies also illustrated the potential of state-of-the-art dynamical atmospheric prediction in crop yield forecasting. However, the potential of coupling both methods has not been fully explored. Herein, we aimed to establish a skilled ML–dynamical hybrid model for crop yield forecasting (MHCF v1.0), which hybridizes ML and a global dynamical atmospheric prediction system, and applied it to northern China at the S2S time scale. In this study, we adopted three mainstream machining learning algorithms (XGBoost, RF, and SVR) and the multiple linear regression (MLR) model, and three major datasets, including satellite data from MOD13C1, observational climate data from CRU, and S2S atmospheric prediction data from IAP CAS, used to predict winter wheat yield from 2005 to 2014, at the grid level. We found that, among the four models examined in this work, XGBoost reached the highest skill with the S2S prediction as inputs, scoring R² of 0.85 and RMSE of 0.78 t/ha 3–4 months, leading the winter wheat harvest. Moreover, the results demonstrated that crop yield forecasting with S2S dynamical predictions generally outperforms that with observational climate data. Our findings highlighted that the coupling of ML and S2S dynamical atmospheric prediction provided a useful tool for yield forecasting, which could guide agricultural practices, policy-making and agricultural insurance.

Subseasonal-to-seasonal (S2S) prediction of winter wheat yields is crucial for farmers and decision-makers to reduce yield losses and ensure food security. Recently, numerous researchers have utilized machine learning (ML) methods to predict crop yield using observational climate variables and satellite data. Meanwhile, some studies also illustrate the potential of state-of-the-art dynamical atmospheric prediction in crop yield forecasting. However, the potential of coupling both methods has not been fully exploited. Herein, we aim to establish a skilled ML–dynamical hybrid model for crop yield forecasting (MHCF v1.0), which hybridizes ML and a global dynamical atmospheric prediction system, and apply it to northern China at the S2S time scale. MHCF v1.0 demonstrates that crop yield forecasting with S2S dynamical predictions generally outperforms that with observational climate data. The coupling of ML and S2S dynamical atmospheric prediction provides a useful tool for yield forecasting, which could guide agricultural practices, policy-making and agricultural insurance.

As the population increases and climate extremes become more frequent, the pressure on food supply increases. A better understanding of the influence of climate variations on crop yield in China would be of great benefit to global food security. In this study, gridded, daily meteorological data and county-level annual yield data were used to quantify the climate sensitivity of corn, rice, and spring wheat yields, and identify the spatiotemporal variation relationship between climate and yields from 1980 to 2014. The results showed that rice and corn were more sensitive to climate variations than spring wheat, both spatially and temporally. Photosynthetic active radiation (PAR) was found to be beneficial to rice in northeast China and the Yangtze River basin, as well as corn in the south and spring wheat in Xinjiang, but not to rice in the south of the Yangtze River and spring wheat in the southeast coast. The temperature centroid shift was the main driving factor causing the movement of the centroid of the three crops. For every 1 km shift of the temperature centroid, the corn and rice yield centroids moved 0.97 km and 0.34 km, respectively. These findings improve our understanding of the impacts of climate variations on agricultural yields in different regions of China.

Timely understanding and quantitative analysis of the changing trend in natural ecosystems in arid and semi-arid areas and their response to the ecological water supply process are of great significance for maintaining the health of oasis ecosystems. Taking the Eichmann Lake wetland of the Aksu River Basin in Xinjiang as the research area, the temporal-spatial distribution characteristics of the lake and the response of ecological water in recent years were studied based on remote sensing images and monitoring data. The results show that: (1) The water surface area of Eichmann Lake is shrinking, from 61.57 km² in 1996 to 27.76 km² in 2020. The changes in water surface area have experienced three stages: rapid decline, slow decline, and slow recovery. After the ecological water supply, the water surface area has obvious seasonal changes with hysteresis; (2) In areas with a low average water level, the ecological water supply has a significant impact on the groundwater level. The higher the water supply is, the higher the groundwater level will be. There is a significant lag effect between the change in the groundwater level and the response of the ecological water supply, which is 1–2 months; (3) The response characteristics of different natural vegetation to the ecological water supply were different in interannual, seasonal, and spatial contexts. The response of Populus euphratica to the ecological water supply is obvious, and its growth is the best within the range of 100–500 m from the water supply outlet. This research can provide the basis for the rational allocation of the Aksu River Basin’s water resources, and also act as a valuable reference for the restoration and reconstruction of surrounding vegetation in the Aksu River irrigation area.