Fig 3 - uploaded by Kilian Vos
Content may be subject to copyright.
Histogram computed from the same region of interest shown in Fig. 2, showing the probability density function (PDF) of MNDWI values for each of the four labelled classes ('water', 'sand', 'whitewater' and 'other land features'). Otsu's threshold (Otsu, 1979) is computed as the threshold that maximises inter-class variance between the 'sand' and 'water' classes, ignoring the 'white-water' and 'other land features' classes which do not contribute to identify the shoreline (i. e., sand/water interface).
Source publication
CoastSat is an open-source software toolkit written in Python that enables the user to obtain time-series of shoreline position at any sandy coastline worldwide from 30+ years (and growing) of publicly available satellite imagery. The toolkit exploits the capabilities of Google Earth Engine to efficiently retrieve Landsat and Sentinel-2 images crop...
Contexts in source publication
Context 1
... SWIR1 and G are the pixel intensity in the short-wave infrared band and green band, respectively. The MNDWI values range between À 1 and 1 as shown in Fig. 2c. Next, a histogram of MNDWI values is constructed with the labelled pixels located within a pre-defined distance (nominally set to 150 m in the toolkit) from 'sand' pixels as illustrated in Fig. 3. In the resulting histogram, the probability density function (PDF) of the 'sand' pixels is centred around positive values of MNDWI, while the 'water' pixels have negative MNDWI values. Accordingly, the 'sand'/'water' threshold is computed applying Otsu's thresholding algorithm (Otsu, 1979) to find the MNDWI value that maximises ...
Context 2
... 'other land features' class. The presence of 'white-water' is known to cause errors in the shoreline detection: Hagenaars et al. (2018) report seaward offsets of up to 40 m in the presence of wave-induced foam; and Pardo-Pascual et al. (2018) notes white-water as being one of the largest sources of error. Indeed, the histogram of MNDWI values in Fig. 3 reveals that the pixels belonging to the 'white-water' class span a broad range of values with no distinct peak and therefore do not assist in discriminating the shoreline and may in fact lead to false detections. The final step in the detection of individual shorelines is to compute the iso-valued contour on the MNDWI image for a ...
Similar publications
In this paper it is proposed to classify an image using a Machine Learning approach inside the Google Earth Engine platform and Jupyter Notebook: feeding the computer with training data (in our case being points/pixels having a label which represent the land-cover type), it will learn to recognize the type of pixel through a model built on the tech...
Frequent observations of surface water at fine spatial scales will provide critical data to support the management of aquatic habitat, flood risk and water quality. Sentinel-1 and Sentinel-2 satellites can provide such observations, but algorithms are still needed that perform well across diverse climate and vegetation conditions. We developed surf...
The results of the estimation of the daily design rainfall in Argentina are presented for 97 % of the continental national territory. This analysis was based on the processing of the Satellite-Derived Precipitation Data (DPDS) of the CHIRPS product (Climate Hazards Group InfraRed Precipitation with Station data), with a spatial resolution of 5 km a...
This study generates a process in GEE (Google Earth Engine) for SUHI (Surface Urban Heat islands) identification derived from TIRS (Thermal Infrared Sensor) and OLI (Operational Land Imager) sensors of Landsat 8 imagery in the area of Stuttgart, Germany. By comparing the temperature images in winter and summer seasons through a regression model, a...
Artificial intelligence (AI) and machine learning (ML) have been applied to solve various remote sensing problems. To fully leverage the power of AI and ML to tackle impactful remote sensing problems, it is essential to enable researchers and practitioners to understand how AI and ML models actually work and thus to improve the model performance st...
Citations
... Satellite images were acquired using the open-source CoastSat toolkit [76]. This allows the user to obtain public satellite data (Landsat 5,7,8,9 and Sentinel-2) by exploiting the capabilities of the Google Earth Engine. ...
The morphodynamics of the Leucate double-crescent bar system was studied over twenty years using bathymetric data supplemented by satellite images and video monitoring. Eleven different bar typologies were identified, mostly based on existing beach state classifications (Low-Tide Terrace, Transverse Bar and Rip, Rhythmic Bar and Beach), also including new heterogeneous typologies (TBR/LTT, RBB HP/RBB, TBR/RBB). The inner bar shows greater variability, with 10 different typologies observed, while the outer bar shows only three different typologies. Summer low-energy periods are dominated by TBR/LTT and TBR typologies, while RBB, although common throughout the year, dominates winter periods along with disrupted bar configurations. The return to less energetic periods in spring is associated with the establishment of heterogeneous typologies. The outer bar has a fairly stable position, although breaches at the embayments and slight movements of its horns can occur following particularly energetic episodes. The inner bar, on the other hand, is much more dynamic, with more common breaches at the embayments and significant cross-shore movement of the horns. Seasonal changes in bar typology do not lead to bar renewal through destruction/reconstruction. Overall, the morphological and typological characteristics of the bar system described here seem somewhat unique compared to the existing literature.
... The cross sections are digitized either through an interactive ad hoc tool, manually by the user, or from a separate file imported from the cross-section coordinates. A more detailed analysis of the CoastSat toolbox is provided in ref. [71]. The CoastSat toolkit is available via the link "https://github.com/kvos/CoastSat ...
... A more detailed analysis of the CoastSat toolbox is provided in ref. [71]. The CoastSat toolkit is available via the link "https://github.com/kvos/CoastSat ...
Climate change is increasingly affecting components of the terrestrial cryosphere with its adverse impacts in the Arctic regions of our planet are already well documented. In this context, it is regarded today as a key scientific priority to develop methodologies and operational tools that can assist towards advancing our monitoring capabilities and improving our decision-making competences in Arctic regions. In particular, the Arctic coasts are the focal point in this respect, due to their strong connection to the physical environment, society, and the economy in such areas. Geoinformation, namely Earth Observation (EO) and Geographical Information Systems (GISs), provide the way forward towards achieving this goal. The present review, which to our knowledge is the first of its kind, aims at delivering a critical consideration of the state-of-the-art approaches exploiting EO datasets and GIS for mapping the Arctic coasts properties. It also furnishes a reflective discussion on the scientific gaps and challenges that exist that require the attention of the scientific and wider community to allow exploitation of the full potential of EO/GIS technologies in this domain. As such, the present study also serves as a valuable contribution towards pinpointing directions for the design of effective policies and decision-making strategies that will promote environmental sustainability in the Arctic regions.
... The extraction of the beachfront was carried out using a series of Landsat 7/8 and 9 images (2000 to 2023) from which we extracted each coastline associated with each image using the Coastsat toolbox developed by Vos et al. 2019. This method provides the reference line from an external source of topographic data obtained from the National Mapping Institute. ...
The coast of Cameroon, located at the bottom of the Gulf of Guinea, is confronted with coastal hazards whose magnitude, distribution, and consequences are currently largely underestimated if not poorly understood. This study aims to fill this gap by proposing an integrated approach to coastal vulnerability assessment, combining simple traditional methods, multicriteria AHP (analytic hierarchy process) analysis, and machine learning techniques. Using geospatial data, field observations, and numerical models, we assessed the 402-km Cameroon coastline, taking into account interactions between physical, geological, and socio-economic factors. The results highlight geomorphology, slope, coastal erosion, and population density as the main contributors to vulnerability. The Integrated Coastal Vulnerability Index (IVCI) calculated by the simple method shows variable levels of vulnerability, with a predominance of “very low” and “low” in the northern sectors (S1 = 58%, S2 = 99%, and S3 = 87%) and “high” and “very high” in the south (S4 = 58% and S5 = 61%). The AHP method reveals a more balanced distribution of vulnerability levels, highlighting a sector (S3 = 96%) at “very strong” and “strong” risk. The application of six machine learning algorithms shows good predictive capabilities for ICVI, with the exception of the support vector machine (SVM). The artificial neural network (ANN) algorithm stands out for its superior accuracy, with an F-score of 0.9, ability to explain data variance (R = 0.95), accurate predictions (RMSE = 0.2), and excellent ability to distinguish classes (kappa coefficient of 0.9 and ROC AUC of 0.9). This study emphasizes the magnitude and complexity of interactions as indicators of the susceptibility of coastal populations to vulnerability.
... Its ability to identify, map, and monitor land cover types (Kuenzer et al. 2011) offers the advantage of frequent revisits and the capacity to retrieve information over large areas using different electromagnetic spectrum ranges (Cárdenas et al. 2017). The accessibility to a growing archive of open-access Earth observation data, coupled with advancing technological development, provides cost-effective solutions for assessing changes in marine ecosystems worldwide (Vos et al. 2019). These technologies have demonstrated success in various applications, such as seagrass monitoring (Chowdhury et al. 2023), tracking changes within estuaries and coastal areas (Dahdouh-Guebas 2022), and monitoring coral reefs (Lubin et al. 2001). ...
Mangroves, integral to ecological balance and socioeconomic well-being, are facing a concerning decline worldwide. Remote sensing is essential for monitoring their evolution, yet its effectiveness is hindered in developing countries by economic and technical constraints. In addressing this issue, this paper introduces MANGLEE (Mangrove Mapping and Monitoring Tool in Google Earth Engine), an accessible, adaptable, and multipurpose tool designed to address the challenges associated with sustainable mangrove management. Leveraging remote sensing data, machine learning techniques (Random Forest), and change detection methods, MANGLEE consists of three independent modules. The first module acquires, processes, and calculates indices of optical and Synthetic Aperture Radar (SAR) data, enhancing tracking capabilities in the presence of atmospheric interferences. The second module employs Random Forest to classify mangrove and non-mangrove areas, providing accurate binary maps. The third module identifies changes between two-time mangrove maps, categorizing alterations as losses or gains. To validate MANGLEE’s effectiveness, we conducted a case study in the mangroves of Guayas, Ecuador, a region historically threatened by shrimp farming. Utilizing data from 2018 to 2022, our findings reveal a significant loss of over 2900 hectares, with 46% occurring in legally protected areas. This loss corresponds to the rapid expansion of Ecuador’s shrimp industry, confirming the tool’s efficacy in monitoring mangroves despite cloud cover challenges. MANGLEE demonstrates its potential as a valuable tool for mangrove monitoring, offering insights essential for conservation, management plans, and decision-making processes. Remarkably, it facilitates equal access and the optimal utilization of resources, contributing significantly to the preservation of coastal ecosystems.
... The otsu's method involves finding an optimal threshold based on the observed distribution of pixel values [51,52] and an acceptable method for achieving stable temporality classification for time series analysis. Otsu's segmentation method can effectively separate land from water along the coast [51,53,54]. ...
Coastal zone is the most dynamic environment and requires regular monitoring to maintain sustainable coastal resource management. Employing remote sensing technology to monitor changes involves an essential process referred to as coastline extraction. This study aims to conduct a comparative analysis of five commonly employed techniques for coastline extraction: Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), Water Ratio Index (WRI), and Automated Water Extraction Index (AWEI). These methods are evaluated by automatically delineating sandy beaches and coastlines using highresolution imagery from the Sentinel-2 satellite for Bali, Indonesia. The Otsu algorithm is utilized to determine the optimal threshold value. Compared to the other indices, the results indicate that MNDWI proved to be the most effective in highlighting water bodies, with the average distance from the validated point of MNDWI being 12.9 m and a Mean Absolute Error (MAE) of 7.11, whereas NDVI demonstrated a high level of proficiency in detecting coastal vegetation. This study highlights the potential of utilizing both Sentinel-2 satellite imagery and the Google Earth Engine (GEE) platform for efficient coastline monitoring. This study also provides scientific evidence supporting the reliability and accuracy of coastline extraction through spectral water indices.
... These datasets were computed from the waterlines derived from optical satellite imagery along 4 km of coastline at Truc Vert (see panel a in Fig. 1). The extraction of the instantaneous waterline position , was performed using the python toolkit CoastSat (Vos et al., 2019b) which is freely available on GitHub (https://github.com/kvos/ CoastSat). ...
Modelling and predicting the future of sandy shorelines is a key challenge in coastal research and is critical for sustainable coastal management. However, currently the most skillful shoreline models strongly rely on data to calibrate the free parameters, and are thus restricted to a few well monitored sites in the world. Here we address the challenges and opportunities offered by optical satellite imagery to provide useful information for equilibrium shoreline model calibration on cross-shore transport dominated sites. We focus on Truc Vert beach, southwest France, where previous work showed good equilibrium model skill to reproduce shoreline change from the time scales of hours (storms) to decades. Satellite derived waterlines are extracted over 11 years (2009–2020) and further transformed into satellite derived shorelines (SDS) with different water level corrections (e.g. tide and/or run up) and varying alongshore averaging lengths, and thus different uncertainties, in order to test model performance. Successively the timeseries duration and sampling frequency required for model calibration were also investigated. The model calibrated using the SDS data showed similar skill as the model calibrated using in-situ alongshore averaged shoreline positions, even for the uncorrected SDS dataset which Root Mean Square Error (RMSE) are approximately 30 m. Alongshore averaging was found to be the only necessary processing of the SDS data while any other site-specific corrections did not significantly improve model skill. Finally to further investigate the effect of sampling frequency and noise in the dataset we performed an analysis using a synthetic shoreline. Our results suggest that the effect of noise is negligible as long as the sampling frequency remains high (dt
30 days). Pending further validation, results show the strong potential of using uncorrected SDS dataset for shoreline model calibration at cross-shore transport dominated sandy coasts.
... The GEE archives Earth observation datasets that provide timely, accurate high-resolution insights into the world's habitats and ecosystems. It enables a real-time monitoring system on land features based on Earth Engine Apps for the large community (Zhang et al. 2020;Vos et al. 2019). Monitoring water quality parameters across a large spatial coverage of rivers, lakes, streams, and ponds, on the other hand, is the most important prerequisite for sustaining ecosystem health and habitat life. ...
Turbidity is an essential biogeochemical parameter for water quality management because it shapes the physical landscape and regulates ecological systems. It varies spatially and temporally across large water bodies, but monitoring based on point-source field observations remains a difficult task in developing countries due to the need for logistics and costs. In this study, we present a novel semi-analytical approach for estimating turbidity from remote sensing reflectance (Rrs)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$({R}_{{\text{rs}}})$$\end{document} in moderate to highly turbid waters in the lower part of the Godavari River (i.e., locations near Rajahmundry). The proposed method includes two sub-algorithms—Normalized Difference Turbidity Index (NDTI) and semi-empirical single-band turbidity (Ts\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T}_{{\text{s}}}$$\end{document}) algorithm—to retrieve spectral reflectance information corresponding to the study locations for turbidity modeling. Sentinel-2 Multi-Spectral Imager data have been used to quantify the turbidity in the Google Earth Engine (GEE) platform. The correlation analysis was observed between spectral reflectance values and in situ turbidity data using cubic polynomial regression equations. The results indicated that the Ts\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${T}_{{\text{s}}}$$\end{document}, which uses the only red-edge wavelength, identified turbidity as the most accurate across all locations (highest R² = 0.91, lowest RMSE = 0.003), followed by NDTI (highest R² = 0.85, lowest RMSE = 0.05), respectively. The remote sensing data application provides a better way to monitor turbidity at large spatio-temporal scales in attaining the water quality standards of the Godavari River.
... This study was based on the algorithm of a thresholding segmentation technique called CoastSat used to detect and extract the shorelines. CoastSat was developed by Vos et al. [34] and has been shown to have good performance for beaches [35,36], reservoirs [37] and tidal flats [38]. The shoreline detection algorithm implemented in CoastSat combines sub-pixel edge segmentation with image classification components, refining the segmentation into four different categories: sand, water, whitewater, and other land features. ...
This study presents an in-depth analysis of the dynamic beach landscapes of Hainan Island, which is located at the southernmost tip of China. Home to over a hundred natural and predominantly sandy beaches, Hainan Island confronts significant challenges posed by frequent marine natural disasters and human activities. Addressing the urgent need for long-term studies of beach dynamics, this research involved the use of CoastSat to extract and analyze shoreline data from 20 representative beaches and calculate the slopes of 119 sandy beaches around the island for the period from 2013 to 2023. The objective was to delineate the patterns of beach evolution that contribute to the prevention of sediment loss, the mitigation of coastal hazards, and the promotion of sustainable coastal zone management. By employing multi-source remote sensing imagery and the CoastSat tool, this investigation validated slope measurements across selected beaches, demonstrating consistency between the calculated and actual distances despite minor anomalies. The effective use of the finite element solution (FES) in the 2014 global tidal model for tidal corrections further aligned the coastlines with the mean shoreline, underscoring CoastSat’s utility in enabling precise coastal studies. The analysis revealed significant seasonal variations in shoreline positions, with approximately half of the monitored sites showing a seaward progression in summer and a retreat in winter, which were linked to variations in wave height. The southern beaches exhibited distinct seasonal variations, which contrasted with the general trend due to differing wave impacts. The western and southern shores showed erosion, while the northern and eastern shores displayed accretion. The calculated slopes across the island indicated that the southern beaches had steeper slopes, while the northern areas exhibited more pronounced slope variations due to wave and tidal impacts. These findings highlight the critical role of integrated coastal management and erosion control strategies in safeguarding Hainan Island’s beaches. By understanding the mechanisms driving seasonal and regional shoreline changes, effective measures can be developed to mitigate the impacts of erosion and enhance the resilience of coastal ecosystems amidst changing environmental conditions. This research provides a foundational basis for future efforts aimed at the sustainable development and utilization of coastal resources on Hainan Island.
... Costs of a professional ArcGIS license may restrict usage of this approach. Open source options include the Analysing Moving Boundaries in R (AMBUR) tool, and the QGIS based EPR4Q, and the Python based Coastsat (Jackson and Short, 2020;Vos et al., 2019). Each generate SCA statistics, but require considerable manual user input and do not automate satellite image based outputs, and in the case of Coastsat can only map changes of 10 meters or above. ...
... The use of waterline processors for the automatic detection of instantaneous waterline Vos et al., 2019b) is one of the simplifications of both global and local assessments based on globally available satellite-derived coastlines. These processors are usually calibrated with information from a limited set of beaches, and consequently, the application of these tools to areas with different conditions may limit the successful detection of the coastline and coastal change assessment (Castelle et al., 2021;Ceccon et al., 2021;Vos et al., 2019bVos et al., , 2023. ...
... The use of waterline processors for the automatic detection of instantaneous waterline Vos et al., 2019b) is one of the simplifications of both global and local assessments based on globally available satellite-derived coastlines. These processors are usually calibrated with information from a limited set of beaches, and consequently, the application of these tools to areas with different conditions may limit the successful detection of the coastline and coastal change assessment (Castelle et al., 2021;Ceccon et al., 2021;Vos et al., 2019bVos et al., , 2023. ...
... Satellite-derived waterlines were obtained by means of the opensource tool CoastSat to each of the pilot sites (Vos et al., 2019b). Only waterlines within a maximum time lag of ±6 days from the dates of the in-situ measurements were considered. ...
