Rice yield is not only influenced by factors of varieties and managements, but also by environmental factors. In this study, agronomic trait data of rice and climate data in eastern China were collected, and rice yields were predicted using a variety of algorithms, including the non-linear tool of feed-forward backpropagation neural networks (FFBN) and the linear model of partial least squares regression (PLSR). The results showed that both the agronomic traits and the climate data were significantly related with rice yield. The PLSR model showed that covariates occurred among the parameters, and modifications should be considered for climate data-based modelling. The FFBN model demonstrated better prediction performance than that of PLSR, in which the relation coefficient (R²) and root mean square error (RMSE) were 0.611 vs. 0.374 and 0.578 vs. 0.865 ton/ha using climate data, respectively; and 0.742 vs. 0.689 and 0.556 vs. 0.608 using agronomic trait data, respectively. When using fused data the R² and RMSE improved to 0.843 vs. 0.746 and 0.440 vs. 0.549, respectively. The optimum architecture of the FFBN consisted of one hidden layer with 29 neurons. Therefore, the FFBN algorithm is an effective option for the prediction of rice yield in complex systems of rice production.

Soil erosion poses a global threat to arable land and its sustainability, particularly in China, where the most severe soil erosion exists worldwide. No-tillage (NT) and mulching (NTS) are considered the most effective soil management techniques for reducing erosion, but only 10% of the global area utilizes them. Therefore, in comparison to conventional tillage (CT), we conducted a comprehensive national assessment of NT and NTS to evaluate their impact on water erosion across China’s croplands for the period spanning 2000 to 2018, through using Revised Universal Soil Loss Equation (RUSLE); subsequently, we projected the temporal and spatial erosion distribution, and examined their effects of various underlying driving factors by using a random-forest model. Nationally, the average soil loss rates were 1085, 564, and 396 t km⁻² a⁻¹ for the CT, NT, and NTS, respectively, across the entire arable land over a span of 18 years. This represents a reduction of 48% and 64% in the NT and NTS, respectively, compared to CT. From 2000 to 2018, water erosion-induced soil loss exhibited a slightly increasing trend with a wavelike pattern in CT, NT, and NTS. The spatial distribution of water erosion in China’s arable land was primarily influenced by local precipitation, accounting for 45% to 52% of the total impact on CT, NT, and NTS. Additionally, the soil slope degree played a role, contributing 29% to 36% of the erosion patterns. Overall, NT and NTS demonstrated superior performance in mitigating the soil erosion in the southern regions of China, including the Central South, Southwest, and East China, owing to the substantial local rainfall and steep terrain. In contrast, NT and NTS exhibited a lower but still significant reduction in soil loss in the northern regions of China due to the flat topography and limited rainfall. However, considering the trade-off between economic losses (yield) and ecosystem benefits (erosion control), we recommend implementing NT and NTS primarily in the northern parts of China, such as the Northeast, North China, and Northwest.

One of the primary challenges of our time is to feed a growing and more demanding world population with degraded soil environments under more variable and extreme climate conditions. Conservation tillage (CS) and deep tillage (DT) have received strong international support to help address these challenges but are less used in major global food production in China. Hence, we conducted a large-scale literature search of English and Chinese publications to synthesize the current scientific evidence to evaluate the effects of CS and DT on soil protection and yield maintenance in the Northeast China Plain, which has the most fertile black soil (Mollisols) and is the main agricultural production area of China. As a result, we found that CS had higher soil bulk density, strong soil penetration resistance, greater water contents, and lower soil temperature, and was well-suited for dry and wind erosion-sensitive regions i.e., the southwest areas of the Northeast. Conversely, DT had better performance in the middle belt of the Northeast China Plain, which contained a lower soil temperature and humid areas. Finally, we created an original dataset from papers [dataset 1, including soil physio-chemical parameters, such as soil water, bulk density, organic carbon, sand, silt, clay, pH, total and available nitrogen (N), phosphorus (P), and potassium (K), etc., on crop biomass and yield], by collecting data directly from publications, and two predicted datasets (dataset 2 and dataset 3) of crop yield changes by developing random forest models based on our data.