Land, a core natural resource, profoundly impacts human economic activities and the health and stability of ecosystems1. Habitat quality is crucial, indicating the survival conditions and suitability for biological species in biomes2. Changes in land use and patterns driven by population growth and urbanization significantly affect the structure and function of natural environments and biological habitats3,4,5,6. These highlight the deep connection between socio-economic activities and the natural environment, which plays a pivotal role in influencing habitat quality7,8,9. In northwestern China, Zhongwei is essential for maintaining biodiversity, conserving soil and water, purifying the air, and controlling wind and sand. Located in a transition zone between arid and semi-arid climates, with complex geographic conditions and a fragile ecological environment, Zhongwei faces multiple ecological challenges, including vegetation degradation, land desertification, biodiversity decline, and soil erosion. Therefore, analyzing the characteristics of land use changes and the drivers of land expansion in Zhongwei is critically important for protecting ecosystems and managing ecological risks. This analysis helps predict the trends and causes affecting habitat quality10,11,12.
Land use prediction refers to the analysis of historical data and the prediction of the future land use trends, which is vital for stipulating scientific land management policies13,14. With the technological development in GIS, remote sensing, machine learning, and artificial intelligence, great breakthroughs have been made in the theory, method and application of land use prediction10,11,12,13,14. In current research, the most commonly used models for land use prediction are the coupled models such as Markov-FLUS and CA-Markov3. These excel in processing large-scale spatial data, simulating time series changes in land use, integrating multiple influencing factors, and providing dynamic simulations3,4,12. However, these models still exhibit significant limitations in their strategy for mining transformation rules and in the simulation of landscape dynamics3,4. The PLUS model, an emerging tool in land use change simulation, integrates deep learning and multi-scale spatial analysis to enhance adaptability and accuracy in managing complex landscape dynamics, while preserving detailed spatial data. This model addresses the deficiencies of the traditional coupled models in transformational rule mining strategies and the simulation of landscape dynamics, thus providing a scientific basis for land resource management and planning3,4,10,12,15. Habitat quality assessment involves the systematic analysis and evaluation of the conditions of biological habitats within a specific area, assessing factors such as ecosystem structure and function, species diversity, and overall habitat health. Typically, this assessment is conducted by constructing ecological indicators for habitats using various ecological models, supported by field surveys and remote sensing techniques16,17. Currently, Recently, the InVEST model, which accounts for threatening factors, has been widely used in habitat quality assessment due to its advantages in spatial visualization and accuracy18,19,20. However, studies that combine the PLUS model and the InVEST model to predict land use and habitat quality in the arid zone of northwestern China have not yet been reported.
Zhongwei, an important ecological security barrier in China, has a fragile ecological environment and serious land sandification problems21. The unique geographic and climatic conditions of the region make it an ideal subject for studying land use changes and their impacts on the ecosystem. Additionally, the distinctive natural conditions provide a representative and valuable basis for research in land use, ecological protection, and agricultural development. In recent years, as the economy and society in this region have rapidly developed, particularly in the fields of tourism and agriculture, land use has undergone significant changes. These changes have placed considerable pressure on land resources22. Consequently, a critical challenge now facing this region is how to utilize land resources scientifically while protecting and enhancing habitat quality. Simulating and predicting land use changes and their ecological impacts under different development scenarios is crucial for assessing and optimizing conservation measures. This promotes ecological security and sustainable development in the region. Therefore, we utilized land use data from 1980, 2000, and 2020 to couple the PLUS and InVEST models. This allowed us to predict and analyze the spatial and temporal evolution, characteristics, and expansion drivers of land use patterns and habitat quality in Zhongwei. We considered three scenarios: natural development, cultivated land protection, and ecological protection. This study aims to provide a reference and share experiences relevant to sustainable land management and development in Zhongwei and the broader Northwest China.