Using state-of-the-art remote sensing methods, including Sentinel-1 and Sentinel-2 images, environmental and terrain datasets, and four machine learning models using the Google Earth Engine platform, a 10 meter resolution 2022 land cover map was created, which includes 12 vegetation categories and 3 non vegetation categories (referred to as TP_LC10-2022).
Satellite data: Sentinel-2 and Sentinel-1.
Terrain data: SRTM, used to create a global elevation model with a horizontal accuracy of 16 meters and a vertical accuracy of 6 meters, with a spatial resolution of 30 meters.
Precipitation Data: Climate Hazards Infrared Precipitation and Station Data (CHIRPS) is a comprehensive dataset that records global precipitation from 1981 to the present. CHIRPS integrates satellite imagery with in-situ station data, allowing for the generation of gridded rainfall time series suitable for trend analysis and seasonal drought monitoring at a resolution of 0.05 °.
Temperature data: ERA5 Land, the dataset provides a comprehensive reanalysis of land variables, using hourly temperature data with a resolution of 0.1 ° in 2022.
Land cover classification: includes four main steps: (1) sampling strategy, (2) data preprocessing and feature construction, (3) classification model comparison, and (4) accuracy evaluation and mutual comparison.
TP_LC10-2022 achieved an overall accuracy of 86.5% and a kappa coefficient of 0.854% when using the RF model, outperforming other classification models including GTB, MD, and SVM. The comparison between TP_LC10-2022 and four widely used land cover products (GLC_FCS30-2020, FROMGLC30-2015, FROMGLC10-2017, and WorldCover2021) shows that TP_LC10-2022 has higher overall accuracy and reflects the local scale variation of vegetation types with latitude.
This work is licensed under a
Creative
Commons Attribution 4.0 International License.
CSTR:11738.11.NCDC.ZENODO.DB6668.2024
10.5281/zenodo.10875021