Improved Bathymetric Mapping of Coastal and Lake Environments Using Sentinel-2 and Landsat-8 Images

Abstract

The bathymetry of nearshore coastal environments and lakes is constantly reworking because of the change in the patterns of energy dispersal and related sediment transport pathways. Therefore, updated and accurate bathymetric models are a crucial component in providing necessary information for scientific, managerial, and geographical studies. Recent advances in satellite technology revolutionized the acquisition of bathymetric profiles, offering new vistas in mapping. This contribution analyzed the suitability of Sentinel-2 and Landsat-8 images for bathymetric mapping of coastal and lake environments. The bathymetric algorithm was developed using an empirical approach and a random forest (RF) model based on the available high-resolution LiDAR bathymetric data for Mobile Bay, Tampa Bay, and Lake Huron regions obtained from the National Oceanic and Atmospheric Administration (NOAA) National Geophysical Data Center (NGDC). Our results demonstrate that the satellite-derived bathymetry is efficient for retrieving depths up to 10 m for coastal regions and up to 30 m for the lake environment. While using the empirical approach, the root-mean-square error (RMSE) varied between 1.99 m and 4.74 m for the three regions. The RF model, on the other hand, provided an improved bathymetric model with RMSE between 1.13 m and 1.95 m. The comparative assessment suggests that Sentinel-2 has a slight edge over Landsat-8 images while employing the empirical approach. On the other hand, the RF model shows that Landsat-8 retrieves a better bathymetric model than Sentinel-2. Our work demonstrated that the freely available Sentinel-2 and Landsat-8 imageries proved to be reliable data for acquiring updated bathymetric information for large areas in a short period.

COVID-19 and its impact on environment: Improved pollution levels during the lockdown period – A case from Ahmedabad, India

“…This manual process is not only time consuming and expensive, but also not suitable for analysis over a regional scale ( Duan et al, 2013 ; Ritchie et al, 2003 ). Recently, remote sensing techniques have been widely used to undertake these studies with accuracy, and advantageous of having both cost and time over the manual approaches ( Yunus et al, 2019 ; Avtar et al, 2019 ). The use of remotely sensed data helps not only to monitor and identify the water quality over large regions, but temporal changes can also be assessed regularly.…”

“…Thus, empirically calibrated models, machine learning and artificial intelligence, etc. have been developed to fit the reflectance spectrum with in-situ profiles ( Doxaran et al, 2002 ; Yunus et al, 2019 ; Davies-Colley and Smith, 2001 ; Luis et al, 2019 ). Among these, machine learning has become popular recently in many fields of satellite remote sensing and GIS analysis ( Dou et al, 2020 ; Merghadi et al, 2020 ).…”

Evaluating scale effects of topographic variables in landslide susceptibility models using GIS-based machine learning techniques

“…While logistic regression (LR) is a parametric machine learning algorithm (learning model that summarizes data with a set of parameters of fixed size - no matter how much data we input at a parametric model, it won’t change its mind); both support vector machine (SVM) and random forest (RF) are non-parametric models (algorithms that do not make strong assumptions about the form of the mapping function; also the complexity grows as the number of training samples increases) 19,34 . Among these two non-parametric models, RF does not need any real hyperparameters to tune, whereas SVM requires tuning for the right kernel, regularization penalties, and the slack variable 13,35 . Detailed description and computation of each ML algorithm are provided in the following sections.…”

Advances in Quantitative Earth Remote Sensing: Past, Present and Future

“…The second paper, by Yunus et al [2], presents an improved bathymetric mapping of coasts and lakes using multidecadal Landsat and newly established European Sentinel satellite observations. Data from both satellites are available freely through a variety of portals on the world wide web and the Google Earth Engine, which also offers processing of large datasets, such as the entire Landsat records.…”