Abstract:Timely and large-area monitoring of terrestrial water bodies using remote sensing data is essential to protect water resource security for humans and ecosystems. Although efforts have been made to monitor inter-monthly water bodies at the regional, continental and global scales, surface open water bodies may experience rapid changes within a short duration. At the same time, the data gaps caused by the long revisit frequencies of satellite sensor imaging systems and the extremely cloudy and rainy weather in certain areas mean that a suitable operational method for recording such rapid changes over large areas does not yet exist. The recently launched operational geostationary satellite sensor GOES-16 ABI can provide multi-scale optical images every 10 min in the Americas, and it has great potential to provide a new solution for timely open surface water monitoring. In this research, a framework based on a multi-scale non-linear mixture model (NLMM) was proposed for the automatic estimation of 500 m daily open water body fraction maps across the full extent of the American continent from GOES-16 ABI imagery. Specifically, an exterior spectral library was created firstly from many pairs of training GOES-16 ABI images and 500 m water body fraction maps derived from Landsat images, and an inherited interior pure spectral library was then derived from a pair of the testing GOES-16 ABI image and the corresponding pure water and non-water (i.e., vegetation and soil/rock substrate) fractions extracted from the monthly water recurrence dataset. Furthermore, a synthetic interior mixed spectral library was built finally based on the linear and non-linear mixtures of randomly selected pure endmembers of water and non-water (i.e., vegetation and soil/rock substrate) in the testing image. All the exterior and interior pure and synthetic mixed spectral libraries were employed jointly in the NLMM machine learning method for estimating daily water body fractions with GOES-16 ABI images. Experiments show that the integration of all three different spectral libraries produced 500 m open water body fraction maps with the greatest accuracies. Compared to traditional linear spectral unmixing and a water index thresholding method, the proposed method produced more accurate results both quantitatively and visually. Moreover, the proposed method can produce time-series daily 500 m open water body fraction maps with minimal cloud coverage for the full extent of the Americas by combining dense GOES-16 ABI images at half-hour intervals, opening up great possibilities for near real-time rapid open water body changes monitoring, such as flooding and drought, over large areas, even with a high frequency of cloud cover.

Spectral Mixture Analysis for Bi-Sensor Wetland Mapping Using Landsat TM and Terra MODIS Data

Bi-Scale Analysis of Multitemporal Land Cover Fractions for Wetland Vegetation Mapping

Multiple endmember object spectral mixture analysis for high spatial resolution remote sensing imagery of urban areas

Linear Spectral Mixture Analysis of Landsat TM Data for Monitoring Invasive Exotic Plants in Estuarine Wetlands

Modified multiple endmember spectral mixture analysis for mapping impervious surfaces in urban environments

Estimating wide range Total Suspended Solids concentrations from MODIS 250-m imageries: An improved method

Fractional Monitoring of Desert Vegetation Degradation, Recovery, and Greening Using Optimized Multi-Endmembers Spectral Mixture Analysis in a Dryland Basin of Northwest China

Linear spectral mixture analysis of Landsat TM data for monitoring invasive estuarine vegetation

Wetland Mapping Using HJ-1A/B Hyperspectral Images and an Adaptive Sparse Constrained Least Squares Linear Spectral Mixture Model

Multidimensional soil salinity data mining and evaluation from different satellites

Semi-Analytically Estimating Suspended Particulate Matter Concentration in Inland Waters Using Four General Bands

A New Spectral Mixture Analysis Method Based on Spectral Correlation Matching

Integrating multisensor satellite data merging and image reconstruction in support of machine learning for better water quality management

Aquatic Plant Functional Type Spectral Characteristics Analysis And Comparison Using Multi-Temporal And Multi-Sensor Remote Sensing Over The Poyang Lake Wetland, China

Subpixel Mapping of Hyperspectral Image Based on Multiscale and Multifeature

Automatic mapping of 500 m daily open water body fraction in the American continent using GOES-16 ABI imagery

MBF: A Multi-Band Fusion Method for Sandy Water Extent Mapping Based on Multispectral Satellite Images

A simple and effective spectral-spatial method for mapping large-scale coastal wetlands using China ZY1-02D satellite hyperspectral images

A Surface Water Extraction Method Integrating Spectral and Temporal Characteristics

Inversion of Farmland Soil Moisture Based on Multi-Band Synthetic Aperture Radar Data and Optical Data

Fusion Classification of HSI and MSI Using a Spatial-Spectral Vision Transformer for Wetland Biodiversity Estimation