Revealing cropping pattern shifts and environmental impacts in Northeast China using multi-source satellite data

Abstract

Remote sensing technology offers unique opportunities for monitoring agricultural systems, gaining significance amidst global challenges such as climate change, population expansion, and heightened demands for agricultural products. As a crucial breadbasket in China, Northeast China contributes around a quarter of the nation's total grain output and one-third of grain commodity output. The cropping structure in this vital region has witnessed notable shifts from a diverse crop planting structure to a monoculture grain production system, aiming at ensuring the nation’s food security. However, the shift has come at a cost, the degradation of black soil and the depletion of water resources. Despite the importance of these changes, a comprehensive understanding of the shifting cropping patterns and their environmental consequences in this region remains limited, largely due to the lack of long-term crop distribution information. This study aims to develop a crop mapping method to generate accurate long-term annual crop type maps in Northeast China dating back to 2000 and reveal cropping patterns, rotation patterns, reclamation time, drivers, and environmental consequences of crop shifts. The work carried out in this thesis is as follows:
This study established a crop mapping method to map long-term crop distribution by using multi-source remote sensing data and the Google Earth Engine platform. By using multi-source satellite data from Sentinel and fused Landsat and MODIS, two crop map products were provided with reasonable accuracies, including a 10-m crop map product from 2017 to 2023 with overall accuracies ranging from 78.1% to 91.3%, and a 30-m crop map product from 2001 to 2022 with overall accuracies ranging from 80.1% to 89.3%. Stable and high consistency between mapped results in this study and statistical data at the prefecture-level administrative units also indicated the reliability of the crop map products. The crop map products were used for analysis on cropping patterns shifts, policy evaluation, and environmental impacts.
Pixel-based methods can capture fine details using fine-resolution imagery but can produce noisy results, while the object-based methods offer additional contextual information and thus might have superior mapping results. This study proposed a hybrid method integrating both object and pixel-based mapping strategies, considering planting patterns, planting habits, field size, and landscape attributes, achieving higher mapping performance than using one method alone with overall accuracies improved from 91.3% to 92.2%. The hybrid method combined the advantages of the object-based method in providing superior spatial details and fewer noticeable errors at the parcel level and pixel-based in detecting imperceptible changes of maize and soybean intercropping practice in Northeast China.
This study revealed dramatic shifts in cropping patterns in Northeast China from 2001 to 2022, by using 30-m long-term crop map products generated in this study. We found maize planting areas increased by 110.5%, and paddy rice cultivation expanded by 80.7%, while soybean planting areas declined by 35.1% during the past two decades. Continuous cropping is dominant in Northeast China, with the proportion of maize soybean rotation less than 5 times as high as 83%. Paddy rice was predominantly cultivated in Sanjiang Plain and Liaohe River Delta in 2001, and then cultivated along major rivers in Songneng Plain. Maize cultivation moved northward, while soybean reclamation exhibits a clear southward trend during the past two decades.
Focusing on soybeans, with 57% of national soybean production produced in Northeast China, the soybean planting areas shrunk continuously due to low profit and international trade market. This study evaluated the effectiveness of the recent soybean producer subsidy policy released in 2017 to regulate soybean planting areas and cropping patterns in Northeast China by using 10-m crop map products generated in this study. We found the policy has not revitalized but only stabilized the soybean planting areas, due to the high comparative benefits of maize cultivation. The routine maize-soybean rotation system might dominate the cropping patterns instead of the soybean subsidy policy, and the effects of the routine rotation were more prevalent in the 2300–2700 ℃ d of the ≥10℃ accumulated temperature.
The dramatic shifts in cropping patterns in Northeast China would bring environmental risks to local agricultural system. This study revealed the environmental impacts of the shifts regarding groundwater security and soil organic carbon content in Northeast China during the past two decades. We found ground water storage(GWS) exhibited an overall declining trend of -0.42 cm/year in Northeast China during the past two decades, with 70.7% of regions displaying a significant (P < 0.05) decreasing trend. Rice and maize cultivation were the primary drivers of groundwater depletion. The average soil organic carbon (SOC) content decreased from 21.86 g/kg in 2000 to 20.69 g/kg in 2020 in Northeast China, reflecting a decline rate of -5.32%, and Liaoning Province experienced the highest SOC loss with a rate of 20.06%. Our study also revealed that crop rotation frequency significantly influences SOC content, with high-intensity crop rotation leading to nearly five times the SOC recovery compared to continuous cropping practices.
This study provides new methods of large-scale and long-term crop mapping and offers a valuable dataset that supports guidance for black soil protection, the optimization of regional crop planting structures, the improvement of crop rotation practices, the formulation of agricultural policies, and the promotion of sustainable agricultural development in this crucial grain-producing region of China.

Date of Award	13 Jul 2025
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Fangfang Zhu (Supervisor) & Ping Fu (Supervisor)