Timely and accurate forecasting of crop yields is crucial to food security and sustainable development in the agricultural sector. However, winter wheat yield estimation and forecasting on a regional scale still remains challenging. In this study, we established a two-branch deep learning model to predict winter wheat yield in the main producing regions of China at the county level. The first branch of the model was constructed based on the Long Short-Term Memory (LSTM) networks with inputs from meteorological and remote sensing data. Another branch was constructed using Convolution Neural Networks (CNN) to model static soil features. The model was then trained using the detrended statistical yield data during 1982 to 2015 and evaluated by leave-one-year-out-validation. The evaluation results showed a promising performance of the model with the overall R ${}^2$ and RMSE of 0.77 and 721 kg/ha, respectively. We further conducted yield prediction and uncertainty analysis based on the two-branch model and obtained the forecast accuracy in one month prior to harvest of 0.75 and 732 kg/ha. Results also showed that while yield detrending could potentially introduce higher uncertainty, it had the advantage of improving the model performance in yield prediction.

In the Loess Plateau (LP) of China, the vegetation degradation and soil erosion problems have been shown to be curbed after the implementation of the Grain for Green program. In this study, the LP is divided into the northwestern semi-arid area and the southeastern semi-humid area using the 400 mm isohyet. The spatial–temporal evolution of the vegetation NDVI during 2000–2015 are analyzed, and the driving forces (including factors of climate, environment, and human activities) of the evolution are quantitatively identified using the geographical detector model (GDM). The results showed that the annual mean NDVI in the entire LP was 0.529, and it decreased from the semi-humid area (0.619) to the semi-arid area (0.346). The mean value of the coefficient of variation of the NDVI was 0.1406, and it increased from the semi-humid area (0.1165) to the semi-arid area (0.1926). The annual NDVI growth rate in the entire LP was 0.0079, with the NDVI growing faster in the semi-humid area (0.0093) than in the semi-arid area (0.0049). The largest increments of the NDVI were from grassland, farmland, and woodland. The GDM results revealed that changes in the spatial distribution of the NDVI could be primarily explained by the climatic and environmental factors in the semi-arid area, such as precipitation, soil type, and vegetation type, while the changes were mainly explained by the anthropogenic factors in the semi-humid area, such as the GDP density, land-use type, and population density. The interactive analysis showed that interactions between factors strengthened the impacts on the vegetation change compared with an individual factor. Furthermore, the ranges/types of factors suitable for vegetation growth were determined. The conclusions of this study have important implications for the formulation and implementation of ecological conservation and restoration strategies in different regions of the LP.

Predicting crop maturity dates is important for improving crop harvest planning and grain quality. The prediction of crop maturity dates by assimilating remote sensing information into crop growth model has not been fully explored. In this study, a data assimilation framework incorporating the leaf area index (LAI) product from Moderate Resolution Imaging Spectroradiometer (MODIS) into a World Food Studies (WOFOST) model was proposed to predict the maturity dates of winter wheat in Henan province, China. Minimization of normalized cost function was used to obtain the input parameters of the WOFOST model. The WOFOST model was run with the re-initialized parameter to forecast the maturity dates of winter wheat grid by grid, and THORPEX Interactive Grand Global Ensemble (TIGGE) was used as forecasting period weather input in the future 15 days (d) for the WOFOST model. The results demonstrated a promising regional maturity date prediction with determination coefficient (R²) of 0.94 and the root mean square error (RMSE) of 1.86 d. The outcomes also showed that the optimal forecasting starting time for Henan was 30 April, corresponding to a stage from anthesis to grain filling. Our study indicated great potential of using data assimilation approaches in winter wheat maturity date prediction.

BACKGROUND: The timely, rapid, and accurate near real‐time observations are urgent to monitor the damage of corn armyworm, because the rapid expansion of armyworm would lead to severe yield losses. Therefore, the potential of machine learning algorithms for identifying the armyworm infected areas automatically and accurately by multispectral unmanned aerial vehicle (UAV) dataset is explored in this study. The study area is in Beicuizhuang Village, Langfang City, Hebei Province, which is the main corn‐producing area in the North China Plain.RESULTS: Firstly, we identified the optimal combination of image features by Gini‐importance and the comparation of four kinds of machine learning methods including Random Forest (RF), Multilayer Perceptron (MLP), Naive Bayesian (NB) and Support Vector Machine (SVM) was done. And RF was proved to be the most potential with the highest Kappa and OA of 0.9709 and 0.9850, respectively. Secondly, the armyworm infected areas and healthy corn areas were predicted by an optimized RF model in the UAV dataset, and the armyworm incidence levels were classified subsequently. Thirdly, the relationship between the spectral characteristics of different bands and pest incidence levels within the Sentinel‐2 and UAV images were analyzed, and the B3 in UAV images and the B6 in Sentinel‐2 image were less sensitive for armyworm incidence levels. Therefore, the Sentinel‐2 image was used to monitor armyworm in two towns.CONCLUSIONS: The optimized dataset and RF model are effective and reliable, which can be used for identifying the corn damage by armyworm using UAV images accurately and automatically in field‐scale. © 2022 Society of Chemical Industry.

Variability of the terrestrial water cycle, i.e. precipitation (P), evapotranspiration (E), runoff (Q) and water storage change (ΔS) is the key to understanding hydro-climate extremes. However, a comprehensive global assessment for the partitioning of variability in P between E, Q and ΔS is still not available. In this study, we use the recently released global monthly hydrologic reanalysis product known as the Climate Data Record (CDR) to conduct an initial investigation of the inter-annual variability of the global terrestrial water cycle. We first examine global patterns in partitioning the long-term mean $\stackrel{\mathrm{‾}}{P}$ between the various sinks $\stackrel{\mathrm{‾}}{E}$ , $\stackrel{\mathrm{‾}}{Q}$ and $\stackrel{\mathrm{‾}}{\mathrm{\Delta }S}$ and confirm the well-known patterns with $\stackrel{\mathrm{‾}}{P}$ partitioned between $\stackrel{\mathrm{‾}}{E}$ and $\stackrel{\mathrm{‾}}{Q}$ according to the aridity index. In a new analysis based on the concept of variability source and sinks we then examine how variability in the precipitation ${\mathit{\sigma }}_P^{\mathrm{2}}$ (the source) is partitioned between the three variability sinks ${\mathit{\sigma }}_E^{\mathrm{2}}$ , ${\mathit{\sigma }}_Q^{\mathrm{2}}$ and ${\mathit{\sigma }}_{\mathrm{\Delta }S}^{\mathrm{2}}$ along with the three relevant covariance terms, and how that partitioning varies with the aridity index. We find that the partitioning of inter-annual variability does not simply follow the mean state partitioning. Instead we find that ${\mathit{\sigma }}_P^{\mathrm{2}}$ is mostly partitioned between ${\mathit{\sigma }}_Q^{\mathrm{2}}$ , ${\mathit{\sigma }}_{\mathrm{\Delta }S}^{\mathrm{2}}$ and the associated covariances with limited partitioning to ${\mathit{\sigma }}_E^{\mathrm{2}}$ . We also find that the magnitude of the covariance components can be large and often negative, indicating that variability in the sinks (e.g. ${\mathit{\sigma }}_Q^{\mathrm{2}}$ , ${\mathit{\sigma }}_{\mathrm{\Delta }S}^{\mathrm{2}}$ ) can, and regularly does, exceed variability in the source (${\mathit{\sigma }}_P^{\mathrm{2}}$ ). Further investigations under extreme conditions revealed that in extremely dry environments the variance partitioning is closely related to the water storage capacity. With limited storage capacity the partitioning of ${\mathit{\sigma }}_P^{\mathrm{2}}$ is mostly to ${\mathit{\sigma }}_E^{\mathrm{2}}$ , but as the storage capacity increases the partitioning of ${\mathit{\sigma }}_P^{\mathrm{2}}$ is increasingly shared between ${\mathit{\sigma }}_E^{\mathrm{2}}$ , ${\mathit{\sigma }}_{\mathrm{\Delta }S}^{\mathrm{2}}$ and the covariance between those variables. In other environments (i.e. extremely wet and semi-arid–semi-humid) the variance partitioning proved to be extremely complex and a synthesis has not been developed. We anticipate that a major scientific effort will be needed to develop a synthesis of hydrologic variability.