Data for: High-spatiotemporal-resolution dynamic water monitoring using LightGBM model and Sentinel-2 MSI data

The product quantified monthly surface water dynamics (excluding northern hemisphere winter) in the middle farming-pastoral ecotone of Northern China (M-FPENC) based on Sentinel-2 images using the light gradient boosting machine (LightGBM) at 10 m resolution from 2016 to 2021 with high accuracy. A description of the methodology and details can be found in the manuscript. Please do not hesitate to send an email to libohao@mail.bnu.edu.cn if you have any other questions. This dataset includes Monthly Surface Water Maps for each year from March to November between 2016 and 2021, as well as Annual Water Inundation Frequency Maps. Data is stored in GeoTiff format. Introduction to documents (1) Monthly surface water maps are named as “year+month.tif”, e.g. “201603.tif”. (2) In the monthly surface water maps, pixel values 0, 1, and 2 refer to invalid observations, valid observations but not water, and water, respectively. (3) Water inundation frequency maps are named as “WIF+year.tif”, e.g. “WIF2016.tif”. (4) To reduce the size of the data, the pixel values of the water inundation frequency maps are scaled up by a factor of 1000 and stored and converted to INT16, e.g. the pixel value of 1000 represents the true value of 1, i.e. the water inundation frequency is 100%.