
This document elaborates on the full set of geospatial, thematic and auxiliary datasets adopted in the research of land use evolution, cropland restoration, pollution control and urban expansion simulation across the Erhai Lake Basin. Combined with the coupled Random Forest-Cellular Automata (RF-CA) model framework, these multi-source datasets serve as the fundamental data support for driving factor extraction, model training, spatial constraint setting, multi-scenario simulation and result verification. All datasets utilized in this study are publicly accessible standard scientific data released by authoritative domestic and international academic platforms, which comply with the data sharing specifications of geographical and ecological research fields. Each dataset is described in detail from the aspects of basic overview, spatial coverage, spatial resolution, data format, coordinate system, main application scenarios in this research, data preprocessing workflow, official access links and data usage specifications. The integrated datasets cover land use and cover information, topographic data, hydrological spatial data, socioeconomic data and ecological constraint data, forming a complete data system that matches the research objectives of analyzing land use transition mechanisms and simulating multi-path spatial patterns. All raw data have undergone unified projection transformation, resampling, clipping, mask extraction and missing value filling to guarantee spatial consistency and data quality before being imported into the RF-CA model. The standardized data processing procedures and reliable data sources ensure the scientificity, repeatability and credibility of the entire modeling, simulation and quantitative analysis process.

---

Full Dataset Documentation for Erhai Lake Basin Land Use Simulation Research
1 Overall Introduction of Dataset System
The research adopts five major categories of datasets, namely multi-temporal land use and cover remote sensing monitoring dataset, digital elevation model and derived topographic dataset, global hydrological spatial dataset, socioeconomic grid dataset and ecological constraint zoning dataset. All data are tailored to the geographical scope of the Erhai Lake Basin in Yunnan Province, China. After unified spatial preprocessing, all datasets are unified to the same coordinate system and spatial resolution to eliminate spatial mismatches. The main applications of these datasets include the identification of land use transition driving factors, construction of nonlinear conversion rules based on random forest, setting of multi-layer spatial constraints for cellular automata, division of four simulation scenarios, and quantitative evaluation of spatial conflicts and comprehensive benefits. The following chapters present detailed descriptions, technical parameters, preprocessing steps and official access addresses of each individual dataset.

2 Land Use and Cover Remote Sensing Monitoring Dataset
2.1 Basic Information
This is the core fundamental dataset of the whole research, which provides multi-temporal land use distribution maps for the Erhai Lake Basin and is used to extract historical land use transition characteristics, construct model training samples, and verify the simulation accuracy of the RF-CA model. The dataset classifies land cover into six primary categories including cropland, urban construction land, forest land, grassland, water area and ecological land, which fully matches the land use classification system required by this study. The data is produced and released by the Resources and Environmental Science Data Center of Chinese Academy of Sciences, a well-known authoritative domestic geospatial data platform, and has been widely applied in domestic and international land use change, ecological evolution and regional sustainable development research.
- Spatial Coverage: Erhai Lake Basin, Dali City, Yunnan Province, China
- Spatial Resolution: 30 meters (uniform resampled resolution for the whole study area)
- Data Format: Raster data in GeoTIFF standard format
- Coordinate System: Albers Equal-Area Conic Projection, unified with all auxiliary datasets
- Temporal Series: Multiple phases of historical remote sensing monitoring data, covering the base period for model training and the verification period for accuracy assessment

2.2 Main Application in Research
1. Extract historical land use conversion samples to train the random forest model and calculate the relative contribution rate of each driving factor;
2. Provide the base land use map for cellular automata initialization;
3. Serve as the real reference data for model accuracy verification including Overall Accuracy and Kappa coefficient;
4. Count the area changes of various land types under four simulation scenarios and calculate indicators such as cropland conversion ratio and ecological land growth rate.

2.3 Data Preprocessing Steps
After downloading the original data, the basin boundary vector file is used for spatial clipping to extract the range of the Erhai Lake Basin. Unify the projection coordinate system with other auxiliary datasets, remove invalid edge pixels and fill tiny missing values generated by remote sensing image splicing. Reclassify the secondary land use types in the original data into six major land use categories defined in this research to form a standardized land use raster dataset for model input.

2.4 Official Access Link & Usage Rules
Official platform website: https://www.resdc.cn
Access Instructions: Users need to complete real-name registration and login on the platform. Ordinary registered users can download a certain number of datasets for free every day in accordance with the platform regulations. All datasets are for academic research and non-commercial use. Please follow the platform's data citation specifications when quoting in academic papers. Relevant DOI and citation information can be obtained on the dataset details page of the platform.

3 Digital Elevation Model (DEM) and Derived Topographic Dataset
3.1 Basic Information
The digital elevation model dataset is the core topographic driving data of this research, which is used to extract elevation and slope indicators, and construct topographic constraint conditions for the cellular automata model. The original DEM data is derived from the Geospatial Data Cloud platform, which is operated by the Computer Network Information Center of Chinese Academy of Sciences. The platform provides a variety of global and regional elevation products with reliable spatial accuracy. On the basis of original DEM data, slope data is generated through GIS spatial analysis tools, and both elevation and slope are taken as key natural driving factors of land use transition.
- Spatial Coverage: Erhai Lake Basin, consistent with the land use dataset
- Spatial Resolution: 30 meters (ASTER GDEM standard resolution)
- Data Format: GeoTIFF raster data
- Coordinate System: Unified Albers Equal-Area Conic Projection
- Derived Datasets: Slope raster dataset calculated based on original DEM

3.2 Main Application in Research
1. Take elevation and slope as two core natural driving factors and import them into the random forest model to quantify their contribution to land use transformation;
2. Establish topographic constraint rules for the cellular automata model, set slope threshold restrictions for cropland reclamation and urban expansion, and simulate the spatial inhibition effect of terrain on land use change;
3. Analyze the terrain differentiation characteristics of cropland to wetland conversion in the basin, and explain the spatial distribution law of ecological restoration projects.

3.3 Data Preprocessing Steps
Clip the original DEM data according to the basin boundary, perform depression filling and hydrological correction to eliminate the influence of abnormal terrain points. Use the spatial analysis module of GIS software to calculate the slope raster dataset. Normalize elevation and slope data to the range of 0 to 1 to adapt to the input requirements of the random forest model. Finally, match the spatial pixel size and coordinate system with the land use dataset.

3.4 Official Access Link & Usage Rules
Official platform website: https://www.gscloud.cn
Access Instructions: User registration and login are required before data retrieval and download. The platform supports functions such as online data preview, regional retrieval and data clipping. All elevation datasets are open for academic research. Commercial use is prohibited without authorization. Users shall abide by the data sharing agreement of the platform.

4 Global Hydrographic Dataset (HydroSHEDS)
4.1 Basic Information
The HydroSHEDS global hydrographic dataset is adopted to extract river network distribution, distance to river and other hydrological indicators, which are used to analyze the water proximity effect of the Erhai Lake Basin. This dataset is jointly developed by World Wildlife Fund and relevant international research institutions, and provides seamless global river networks, watershed boundaries and hydrological grid data, which is widely used in hydrological simulation, water ecological protection and watershed land use research. In this study, river network vector data is extracted from the dataset to calculate the Euclidean distance from each pixel to the nearest river, which is identified as the primary driving factor of land use transition.
- Spatial Coverage: Global scope, clipped to Erhai Lake Basin in practical application
- Data Type: Vector river network data and raster flow direction data
- Data Format: Standard GIS vector format and raster format
- Spatial Characteristic: Multi-scale seamless global hydrographic data, stable spatial consistency

4.2 Main Application in Research
1. Calculate the distance to river for each pixel in the basin, and take it as a key hydrological driving factor to participate in the training of random forest model;
2. Analyze the influence law of water proximity effect on the spatial distribution of cropland and urban land;
3. Provide basic hydrographic support for the setting of pollution control related scenarios in multi-scenario simulation.

4.3 Data Preprocessing Steps
Extract the river network vector within the Erhai Lake Basin from the original global hydrographic dataset. Use the Euclidean distance analysis tool to generate the raster dataset of distance to river. Unify the spatial resolution and coordinate system with other datasets, and perform normalization processing to form standardized driving factor data.

4.4 Official Access Link & Usage Rules
Official platform website: https://www.hydrosheds.org
Access Instructions: The dataset is fully open and free for global academic users. Users can directly download core products such as river networks and watershed boundaries on the official website. The data supports multiple GIS formats and has no additional access restrictions. It is required to mark the data source and platform in the research results and academic papers.

5 Socioeconomic Grid Dataset
5.1 Basic Information
The socioeconomic dataset includes population density grid data and road accessibility data, which belong to human activity driving factors. The data is sourced from the socioeconomic thematic datasets of the Resources and Environmental Science Data Center of Chinese Academy of Sciences. Population density data reflects the intensity of human activities, and road accessibility characterizes the location advantages of regional land, both of which are important man-made driving forces affecting urban expansion and land use conversion.
- Spatial Coverage: Erhai Lake Basin
- Spatial Resolution: 30 meters
- Data Format: GeoTIFF raster data
- Coordinate System: Unified Albers Equal-Area Conic Projection

5.2 Main Application in Research
1. Import population density and road accessibility into the random forest model to analyze the contribution of socioeconomic factors to land use transition;
2. Provide basic data for the simulation of urban expansion trend in cellular automata;
3. Assist in identifying high-intensity human activity areas and potential spatial conflict zones between urban expansion and cropland restoration.

5.3 Data Preprocessing Steps
Clip the original socioeconomic data according to the basin boundary, eliminate abnormal extreme values in the grid, and carry out min-max normalization. Match pixels and coordinates with other datasets to ensure spatial superposition consistency.

5.4 Official Access Link & Usage Rules
Official platform website: https://www.resdc.cn
Access Instructions: Same as the land use dataset. After real-name registration, users can search and download population, road and other socioeconomic thematic data. The data is limited to academic research use.

6 Ecological Constraint Zoning Dataset
6.1 Basic Information
The ecological protection redline dataset is a rigid policy constraint dataset for this research, released by the National Earth System Science Data Center. This platform is a national-level comprehensive scientific data platform covering multiple disciplines such as ecology, geography and earth science. The ecological redline is designated as a prohibited development area, which is set as a hard constraint rule in the cellular automata model to simulate the intervention effect of ecological protection policies on land use change.
- Spatial Coverage: Erhai Lake Basin ecological protection redline scope
- Data Type: Vector boundary data
- Data Format: Standard GIS vector format
- Attribute Information: Contains protection level and constraint type attributes

6.2 Main Application in Research
1. Construct rigid spatial constraints for the CA model, prohibit urban expansion and arbitrary cropland occupation within the ecological redline;
2. Take the ecological redline as a policy driving factor and calculate its contribution rate through the random forest model;
3. Support the setting of cropland restoration and pollution control scenarios, and delineate key areas for ecological restoration.

6.3 Data Preprocessing Steps
Convert the ecological redline vector data into binary constraint raster data (1 for restricted area, 0 for unrestricted area), unify the resolution and coordinate system, and superimpose it on other spatial datasets as a constraint layer.

6.4 Official Access Link & Usage Rules
Official platform website: https://www.geodata.cn
Access Instructions: Users need to register and log on to the platform to retrieve ecological constraint datasets. The data is managed in accordance with national ecological protection data specifications. When using the data, relevant national ecological protection management regulations shall be complied with, and the data source shall be clearly cited in academic achievements.

7 Integrated Data Processing Specification & Compatibility Description
All the above five categories of datasets have completed unified spatial processing before being imported into the RF-CA model. The unified coordinate system is Albers Equal-Area Conic Projection, the unified spatial resolution is 30 meters, and all raster data adopt GeoTIFF format which is compatible with MATLAB and mainstream GIS software. All vector data can be converted into raster data according to research needs. The datasets have good spatial matching and data compatibility, which effectively avoids spatial dislocation and data errors in the process of model training and simulation. All data fully comply with the open sharing agreements of each official platform, and all data sources can be marked and cited in academic papers, research reports and journal articles.