CoastSat: A Google Earth Engine-enabled Python toolkit to extract shorelines from publicly available satellite imagery

Vos, Kilian; Splinter, Kristen D.; Harley, Mitchell D.; Simmons, Joshua A.; Turner, Ian L.

doi:10.1016/j.envsoft.2019.104528

Cited by 334 publications

(247 citation statements)

References 37 publications

Supporting

Mentioning

171

Contrasting

Unclassified

Order By: Relevance

“…These functions can be applied directly to large multi-dimensional satellite datasets, making them suitable for applying to large scale remote sensing workflows such as those enabled by Open Data Cube [3,4] or Google Earth Engine [5]. By enabling the automatic extraction of sub-pixel waterlines consistently across space and time, it is anticipated that these tools may supplement existing open source coastal monitoring software [70] and assist in scaling up current local-scale waterline analyses to provide insights into drivers of environmental change operating at regional, continental or global scales. Future work will focus on applying sub-pixel waterline extraction methods to provide accurate estimates of waterbody area and volume for inland reservoirs and lakes to monitor the impact of drought and water extraction within semi-arid inland Australia, and operationalising the approach to monitor fine-scale erosion and coastal change across complex coastal environments along the entire Australian coastline.…”

Section: Discussionmentioning

confidence: 99%

Sub-Pixel Waterline Extraction: Characterising Accuracy and Sensitivity to Indices and Spectra

et al. 2019

View full text Add to dashboard Cite

Accurately mapping the boundary between land and water (the 'waterline') is critical for tracking change in vulnerable coastal zones, and managing increasingly threatened water resources. Previous studies have largely relied on mapping waterlines at the pixel scale, or employed computationally intensive sub-pixel waterline extraction methods that are impractical to implement at scale. There is a pressing need for operational methods for extracting information from freely available medium resolution satellite imagery at spatial scales relevant to coastal and environmental management. In this study, we present a comprehensive evaluation of a promising method for mapping waterlines at sub-pixel accuracy from satellite remote sensing data. By combining a synthetic landscape approach with high resolution WorldView-2 satellite imagery, it was possible to rapidly assess the performance of the method across multiple coastal environments with contrasting spectral characteristics (sandy beaches, artificial shorelines, rocky shorelines, wetland vegetation and tidal mudflats), and under a range of water indices (Normalised Difference Water Index, Modified Normalised Difference Water Index, and the Automated Water Extraction Index) and thresholding approaches (optimal, zero and automated Otsu's method). The sub-pixel extraction method shows a strong ability to reproduce both absolute waterline positions and relative shape at a resolution that far exceeds that of traditional whole-pixel methods, particularly in environments without extreme contrast between the water and land (e.g., accuracies of up to 1.50-3.28 m at 30 m Landsat resolution using optimal water index thresholds). We discuss key challenges and limitations associated with selecting appropriate water indices and thresholds for sub-pixel waterline extraction, and suggest future directions for improving the accuracy and reliability of extracted waterlines. The sub-pixel waterline extraction method has a low computational overhead and is made available as an open-source tool, making it suitable for operational continental-scale or full time-depth analyses aimed at accurately mapping and monitoring dynamic waterlines through time and space.

show abstract

Section: Discussionmentioning

confidence: 99%

Sub-Pixel Waterline Extraction: Characterising Accuracy and Sensitivity to Indices and Spectra

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Both high-resolution modelled datasets were downloaded and analysed in the GEE code editor. To ensure spatial similarity in both gridded datasets, bilinear interpolation resampling technique was used [41]. To match the similar spatial information with station data, nearest point-pixel rainfall values of both the gridded datasets were extracted in GEE platform for station localities (Supplementary code 2) [33,36].…”

Section: Data Handling and Statistical Applications In Geementioning

confidence: 99%

An Analysis of Long-Term Rainfall Trends and Variability in the Uttarakhand Himalaya Using Google Earth Engine

et al. 2020

View full text Add to dashboard Cite

This paper analyses the spatio-temporal trends and variability in annual, seasonal, and monthly rainfall with corresponding rainy days in Bhilangana river basin, Uttarakhand Himalaya, based on stations and two gridded products. Station-based monthly rainfall and rainy days data were obtained from the India Meteorological Department (IMD) for the period from 1983 to 2008 and applied, along with two daily rainfall gridded products to establish temporal changes and spatial associations in the study area. Due to the lack of more recent ground station rainfall measurements for the basin, gridded data were then used to establish monthly rainfall spatio-temporal trends for the period 2009 to 2018. The study shows all surface observatories in the catchment experienced an annual decreasing trend in rainfall over the 1983 to 2008 period, averaging 15.75 mm per decade. Analysis of at the monthly and seasonal trend showed reduced rainfall for August and during monsoon season as a whole (10.13 and 11.38 mm per decade, respectively); maximum changes were observed in both monsoon and winter months. Gridded rainfall data were obtained from the Climate Hazard Infrared Group Precipitation Station (CHIRPS) and Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR). By combining the big data analytical potential of Google Earth Engine (GEE), we compare spatial patterns and temporal trends in observational and modelled precipitation and demonstrate that remote sensing products can reliably be used in inaccessible areas where observational data are scarce and/or temporally incomplete. CHIRPS reanalysis data indicate that there are in fact three significantly distinct annual rainfall periods in the basin, viz. phase 1: 1983 to 1997 (relatively high annual rainfall); phase 2: 1998 to 2008 (drought); phase 3: 2009 to 2018 (return to relatively high annual rainfall again). By comparison, PERSIANN-CDR data show reduced annual and winter precipitation, but no significant changes during the monsoon and pre-monsoon seasons from 1983 to 2008. The major conclusions of this study are that rainfall modelled using CHIRPS corresponds well with the observational record in confirming the decreased annual and seasonal rainfall, averaging 10.9 and 7.9 mm per decade respectively between 1983 and 2008, although there is a trend (albeit not statistically significant) to higher rainfall after the marked dry period between 1998 and 2008. Long-term variability in rainfall in the Bhilangana river basin has had critical impacts on the environment arising from water scarcity in this mountainous region.

show abstract

“…The first has been proposed by Luijendijk et al [113] for the global-scale mapping of sandy beaches and establishing their dynamics. The second approach has been employed by Vos et al [114] to document shoreline dynamics (some examples also deal with sandy beaches). The validity and the importance of the both above-mentioned studies are indisputable, and the availability of any similar approach would help in the global mapping of coastal megaclast deposits.…”

Section: Automatic Detection Of Coastal Megaclast Deposits: the Currementioning

confidence: 99%

Finding Coastal Megaclast Deposits: A Virtual Perspective

Ruban

2020

JMSE

View full text Add to dashboard Cite

Coastal megaclast deposits are dominated by detrital particles larger than 1 m in size. These attract significant attention of modern researchers because of the needs of sedimentary rock nomenclature development and interpretation of storm and tsunami signatures on seashores. If so, finding localities that exhibit coastal megaclast deposits is an important task. Field studies do not offer a quick solution, and, thus, remote sensing tools have to be addressed. The application of the Google Earth Engine has permitted to find four new localities, namely Hondarribia in northern Spain (Biscay Bay), the Ponza Island in Italy (Tyrrhenian Sea), the Wetar Island in eastern Indonesia (Banda Sea), and the Humboldt o Coredo Bay at the Colombia/Panama border (eastern Pacific). In these localities, coastal megaclast deposits consisting of blocks (1–10 m in size) and some megablocks (>10 m in size) are delineated and preliminary described in regard to the dominant size of particles, package density, mode of occurrence, etc. The limitations of such virtual surveys of coastal megaclast deposits are linked to an insufficiently high resolution of satellite images, as well as ‘masking’ effects of vegetation cover and cliff shadows. However, these limitations do not diminish the importance of the Google Earth Engine for finding these deposits. Consideration of some tourism-related information, including photos captured by tourists and bouldering catalogues, facilitates search for promising areas for subsequent virtual surveying of megaclast distribution. It is also established that the Google Earth Engine permits quantitative analysis of composition of coastal megaclast deposits in some areas, as well as to register decade-long dynamics or stability of these deposits, which is important to interpret their origin. The current opportunities for automatic detection of coastal megaclast deposits seem to be restricted.

show abstract

CoastSat: A Google Earth Engine-enabled Python toolkit to extract shorelines from publicly available satellite imagery

Cited by 334 publications

References 37 publications

Sub-Pixel Waterline Extraction: Characterising Accuracy and Sensitivity to Indices and Spectra

Sub-Pixel Waterline Extraction: Characterising Accuracy and Sensitivity to Indices and Spectra

An Analysis of Long-Term Rainfall Trends and Variability in the Uttarakhand Himalaya Using Google Earth Engine

Finding Coastal Megaclast Deposits: A Virtual Perspective

Contact Info

Product

Resources

About