Spatial Downscaling of Suomi NPP–VIIRS Image for Lake Mapping

Abstract

Abstract:Capturing the dynamics of a lake-water area using remotely sensed images has always been an essential task. Most of the fine spatial resolution data are unsuitable for this purpose because of their low temporal resolution and limited scene coverage. A Visible Infrared Imaging Radiometer Suite on board the Suomi National Polar-orbiting Partnership (Suomi NPP-VIIRS) is a newly-available and appropriate sensor for monitoring large lakes due to its frequent revisits and wide swath (more than 3000 km). However, it provides visible and infrared images at relatively coarse spatial resolutions, which would sometimes hamper the accurate mapping of lake shorelines. This study, therefore, proposes a two-step downscaling method that combines spectral unmixing and subpixel mapping to produce a finer resolution lake map from NPP-VIIRS imagery, which is then applied to delineate the shorelines of five plateau lakes in Yunnan Province, as well as the shoreline dynamics of Poyang Lake at three separate times. A newly published global water dynamic dataset is employed in this study to improve the downscaling method. Results suggest that the proposed method can generate a finer resolution lake map that exhibits more details of the shoreline than hard classification. The downscaling results of the Suomi NPP-VIIRS generally achieve higher than 75% accuracy, while the downscaling results of a Landsat-simulated fraction map could have accuracy higher than 85%. This reveals that errors and uncertainties exist in both procedures, but mainly come from the spectral unmixing procedure which retrieves water fractions from NPP-VIIRS data.

Detecting, Extracting, and Monitoring Surface Water From Space Using Optical Sensors: A Review

“…Most of them are common subpixel mapping algorithms, which can be used for any kind of land cover. For surface water subpixel mapping in particular, there are some specific methods, such as those for lake water area mapping (Huang et al, ; Shah, ; Zhang et al, ), and those integrated with terrain information (Huang, Chen, & Wu, ; Ling et al, ). All of these methods took advantage of the special characteristics of surface water, such as spatial distribution or terrain dependency, and produced better results than normal subpixel mapping algorithms.…”

“…So far, most of the aforementioned studies were concentrated on a single procedure, either pixel unmixing or subpixel mapping, because the uncertainties inevitably introduced in each procedure would be multiplied and hard to be evaluated once these two were combined. In combining these two procedures to spatially downscale lake mapping, Huang et al () found that errors and uncertainties existed in both procedures, but mainly came from the spectral unmixing procedure.…”

An Assessment of Surface Water Detection Methods for Water Resource Management in the Nigerien Sahel

“…In addition to pre-calculated global surface water datasets, algorithms for detecting surface water from remote sensing data also have been created and applied at global, regional, and local scales. These include the thresholding of single bands and spectral indices [5][6][7][8][9][10][11][12]; simple decision trees relying on knowledge of the spectral properties of water compared to other land cover types [13][14][15]; spectral mixture analysis (SMA) [16][17][18], and supervised and unsupervised classification schemes (including machine learning algorithms) [19,20].The most common and simple method of identifying surface water using remotely sensed data is calculation of spectral indices. Many indices have been developed specifically to exploit the unique spectral signature of water compared to other land cover types [5][6][7][8][9][10]12,13].…”

“…In addition to pre-calculated global surface water datasets, algorithms for detecting surface water from remote sensing data also have been created and applied at global, regional, and local scales. These include the thresholding of single bands and spectral indices [5][6][7][8][9][10][11][12]; simple decision trees relying on knowledge of the spectral properties of water compared to other land cover types [13][14][15]; spectral mixture analysis (SMA) [16][17][18], and supervised and unsupervised classification schemes (including machine learning algorithms) [19,20].…”

Multitemporal image analysis for monitoring fuzzy shorelines

“…Few studies exist on modelling shoreline using soft classification e.g. fuzzy c-means classification and the linear spectral mixture model (Dewi et al, 2016;Huang et al, 2017;Muslim et al, 2006;Taha and Elbeih, 2010). In our previous study, we used FCM classification to estimate the water memberships and then generate shoreline margin by using a choice of thresholds (Dewi et al, 2016).…”