This dataset uses the Google Earth Engine (GEE) platform to analyze the spatiotemporal evolution characteristics of China's coastline, including Hainan and Taiwan, from 1990 to 2019 using long time series of images from the Land Satellite Theme Imager (TM), Enhanced Theme Imager Plus (ETM+), and Land Imager (OLI).
1. We retrieved all available Level 1 Earth Remote Sensing satellite orthorectified images with terrain correction (L1T) from 1990 to 2019, including the Topographic Imager (TM), Enhanced Topographic Imager Plus (ETM+), and Land Imager (OLI). A total of 60527 land remote sensing satellite images (Figure 1b) are archived on the Google Earth Engine (GEE) platform as part of the United States Geological Survey (USGS) image collection.
2. Select 17 key tidal stations, including 8 warm temperate tidal stations, 7 subtropical tidal stations, and 2 tropical tidal stations (Jia et al., 2021). The tide tables for these sites can be found online（ https://www.chaoxibiao.net/ Last visit date: November 12, 2024. In order to eliminate the influence of precipitation, we selected tidal data from the rainy season in coastal areas of China, spanning from April to September (Liu et al.) Based on the high tide time and satellite transit time, we selected 1583 images to extract the coastline.
3. Three global coastline datasets were used as reference datasets. A new global coastline vector with a spatial resolution of 30 meters (GSV; Sayre et al., 2019) was developed based on the annual composite map of 2014 Landsat satellite images. The 2015 Global Multiscale Coastline Dataset (GMSSD_2015; Liu et al., 2019) was generated using Google Earth imagery. This dataset includes global vector coastline data with a resolution of meters, which was obtained through human-computer interaction and analysis of Google Earth images. Coastline_CS (Li et al., 2019) is a spatiotemporal dataset of the continental coastline and its types in the East China Sea region. This dataset covers a 5-year time interval from 1990 to 2015, generated using Landsat images and single band edge detection methods: 1990, 2000, and 2015.
Adopting a threshold segmentation method based on exponential time series reconstruction. Firstly, a high-quality reconstructed Modified Normalized Difference Water Index (MNDWI) time series was constructed using the Time Series Harmonic Analysis Algorithm (HANTS). Secondly, using the Otsu algorithm, the land and water in coastal areas are separated based on the MNDWI value of high tide levels. Finally, a 30-year coastline dataset was generated.
Evaluated the consistency between the extracted coastline and the reference dataset. The results showed that over 50% of the sample points had an offset of less than 1 px. More than 70% of the sample points had an offset of less than 2 px, and over 80% had an offset of less than 3 px. The maximum average absolute deviation of the coastline was 57.98 meters, and the minimum average absolute deviation was 29.35 meters. The coastline extracted by this method has good consistency with the reference dataset, which to some extent confirms the reliability of the results.
This work is licensed under a
Creative
Commons Attribution 4.0 International License.
https://cstr.cn/31253.11.sciencedb.16228
https://doi.org/10.57760/sciencedb.16228
