Land use transition is a manifestation of land use and land cover change (LUCC) and is also a major research focus of the Global Land Project (GLP), as well as land system science (LSS)

The analysis of urban land expansion and farmland loss is essential to adequately understand the land use change in a rapidly urbanizing China. We found that both urban expansion and farmland loss in Beijing experienced high- and low-speed stages and their spatial patterns were consistent during the past 35 years as most of the newly expanded urban land was converted from farmland. The area of farmland loss by urban expansion in Beijing is 12.6 km²/year, 39.86 km²/year, 23.38 km²/year, and 41.11 km²/year during the period of 1980–1990, 1990–2000, 2000–2010, and 2010–2015, respectively. The urban expansion in Beijing continuously preferred to consume “above average” quality farmland during 1980–2015. Meanwhile, although the urban expansion in Beijing was highly dependent on occupying farmland, the dependence of urban expansion on farmland consumption has declined over time. However, the contribution of urban expansion on farmland loss increased during 1980–2010 and decreased afterward. In order to protect the farmland from urban expansion, we call for more effort to improve the urban land use efficiency with rigid controls over areas of urban expansion.

Soil analysis using near-infrared spectroscopy has shown great potential to be an alternative to traditional laboratory analysis, and there is continuously increasing interest in building large-scale soil spectral libraries (SSLs). However, due to issues such as high non-linearity in soil spectral data and complexity in soil spatial variation, the establishment of robust prediction models for soil spectral libraries remains a challenge. This study aimed to investigate the performance of deep learning algorithms, including long short-term memory (LSTM) and LSTM–convolutional neural networks (LSTM–CNN) integrated models, to predict the soil organic matter (SOM) of a provincial-scale SSL, and compare it to the normally used local weighted regression (LWR) model. The Hebei soil spectral library (HSSL) contains 425 topsoil samples (0–20 cm), of which every 3 soil samples were collected from dry land, irrigated land, and paddy fields, respectively, in different counties of Hebei Province, China. The results show that the accuracy of the validation dataset rank as follows: LSTM–CNN (R²_p = 0.96, RMSE_p = 1.66 g/kg) > LSTM (R²_p = 0.83, RMSE_p = 3.42 g/kg) > LWR (R²_p = 0.82, RMSE_p = 3.79 g/kg). The LSTM–CNN model performed the best, mainly due to its comprehensive ability to effectively extract spatial and temporal features. Meanwhile, the LSTM model achieved higher accuracy than the LWR model, owing to its built-in memory unit and its advantage of faster feature band extraction. Thus, it was suggested to use deep learning algorithms for SOM predictions in SSLs. However, their performance on larger-scale SSLs such as continental/global SSLs still needs to be further investigated.

In recent decades, global social and economic development has resulted in substantial land-use transitions. This was first observed with respect to losses of forested land, attracting worldwide concern. Forest transitions have an important impact on global ecology, whilst farmland transitions are key in terms of global food security. However, research into farmland transitions is lacking, particularly with respect to mechanistic analysis. Using data on China’s farmland areas between 1950 and 2017, we investigated the transitional characteristics, and triggers, of farmland change through linear regression analysis. Furthermore, based on the Advocacy Coalition Framework, we reveal the internal mechanism of these transitions. Our main findings are as follows: (1) between 1950 and 2017, China’s farmland area exhibited significant growth, and there were two transitions, namely in 1984 and 2004; (2) macroscopic economic and social changes determine the overall evolution of the farmland area; (3) there were two advocacy coalitions in the farmland transition policy subsystem—the farmland supplement and farmland consumption coalitions; (4) under the influence of macroscopic economic and social development, external events play a catalytic role in the transitions, and relatively stable parameters have an indirect but lasting effect in terms of transition outcomes.

By mastering the spatial-temporal evolution of patterns of soybean production, a reference for optimizing a soybean production layout could be provided, ensuring food security. The variation coefficient method, and the comparative advantage and spatial autocorrelation models were used to analyze the spatial divergence regularities of soybean production, sown area and yield, spatial-temporal changes in the comparative advantages of soybean planting efficiency and soybean planting scale, and the spatial agglomeration characteristics in China from 1949 to 2019. The results indicate that (1) from 1949 to 2019, soybean production and yield changes in China remained constant with a fluctuating upwards trend, and soybean sown areas hardly changed, yet experienced a sharp fluctuation. (2) The Northeast China Plain (NECP) was the main soybean-producing area, and its main position was strengthened. In contrast, the main soybean production position of the Huang-Huai-Hai Plain (HHHP) has declined. The Northern arid and semiarid region (NASR), the Sichuan Basin and surrounding areas (SBSR), the Middle-Lower Yangtze Plain (MLYP), and the Yunnan-Guizhou Plateau (YGP) became new soybean production growth poles. (3) The spatial distribution of soybean planting efficiency-related comparative advantages in China extended from northern China to the whole country, and the soybean planting scale-related comparative advantages proceeded through three stages: steady expansion, relative stability, contraction, and stabilization. (4) The spatial agglomeration of soybean planting efficiency-related comparative advantages has weakened, and the spatial agglomeration of the soybean planting scale-related comparative advantages exhibited a strengthening-weakening-strengthening-weakening process. Through our research analysis, we propose a policy resource to fully utilize the soybean planting efficiency-related comparative advantages in southern China (SC), promote grain-soybean rotation patterns in the HHHP and NECP, improve the soybean cultivation subsidy system, and build a soybean industry chain in the NECP.

We demonstrated an innovative learning method of convolutional neural network (CNN) to identify landuse and land cover (LULC) patterns and extract features to disaggregate socio-economic factors by using remote sensing imageries at 30 m spatial resolution. The training labels were extracted from the historical LULC map to reduce the huge cost of labelling work, and to provide an inaccuracy but sufficient training dataset. The fully connected layer of the trained CNN was extracted as disaggregating features to map socio-economic factors of population and gross domestic product (GDP). Results indicate that current method can attain 92% overall agreement of LULC identification with the cross-validation of other products. The determination coefficient of disaggregating socio-economic factors can reach 0.945 for population density, and 0.876 for GDP density with the cross-validation at county level.