Data-Driven Shoreline Modeling: Historical Remote Sensing and Prediction Methods DOI
Mark Lundine, Arthur C. Trembanis

Опубликована: Янв. 1, 2024

This study describes a novel, data-driven approach for extracting satellite-derived shorelines, constructing shoreline change timeseries, and projecting future positions. First, to build datasets, an image-to-image translation generative adversarial network was developed map RGB satellite imagery into segmented land/water images, after which shorelines are extracted via contour tracing algorithm (Marching Squares). model incorporates over 20,000 LANDSAT 5, 7, 8 Sentinel-2 images from the Delmarva coast (USA), New Jersey Shore Long Island Northern Tuscany Littoral Cell (Italy), spanning 1984 present time. Next, shore-normal transects generated at set alongshore intervals, then used compute intersection points with construct timeseries of cross-shore position along given transect. With these yearly linear trends computed through ordinary least squares, maps illustrating across broad (several km tens km) coastal regions constructed. Last, long short-term memory networks incorporated on individual datasets learn behavior each project positions (with bootstrapped uncertainties). For select regions, projected transect merged together full two-dimensional uncertainty polygons), time period. All software components this open-source freely available, as trained models labelled imagery. In future, methodology should be extended other coasts worldwide, well integrate higher spatial temporal resolution

Язык: Английский

Shoreline Change of Western Long Island, New York, from Satellite-Derived Shorelines DOI Creative Commons

Catherine N. Janda,

Jonathan A. Warrick, Daniel Buscombe

и другие.

Coasts, Год журнала: 2025, Номер 5(1), С. 2 - 2

Опубликована: Янв. 2, 2025

Shoreline measurement techniques using satellite-derived imagery can provide decades of observations shoreline change. Here we apply these to the western south shore Long Island, New York, which has three distinct beaches, Rockaway Peninsula, Beach, and Jones Beach are 18, 15, 24 km in length, respectively. These beaches recreation areas for millions regional residents include several groin fields, sediment dredging nourishment operations, a coastal wave climate that includes winter northeasterly storms summer hurricanes. The shorelines along ends have been accreting at ~4 m/yr during observation record (1984–2022) resulting from net westward longshore drift. central 10–12 lower change rates, rates generally lowest within fields (0.5–1.5 m/yr). also evidence propagating accretion erosion waves durations years. projects shown significantly influence accretion, this is commonly followed by significant retreat subsequent

Язык: Английский

Процитировано

0
Historical Coast Snaps: Using Centennial Imagery to Track Shoreline Change DOI Creative Commons
Fátima Valverde, Rui Taborda, Amy E. East

и другие.

Remote Sensing, Год журнала: 2025, Номер 17(8), С. 1326 - 1326

Опубликована: Апрель 8, 2025

Understanding long-term coastal evolution requires historical data, yet accessing reliable information becomes increasingly challenging for extended periods. While vertical aerial imagery has been extensively used in studies since the mid-20th century, and satellite-derived shoreline measurements are now revolutionizing change studies, ground-based images, such as photographs picture postcards, provide an alternative source of data earlier periods when other datasets scarce. Despite their frequent use documenting qualitative morphological changes, these valuable sources have rarely supported quantitative assessments evolution. This study demonstrates potential ground-oblique images quantitatively assessing position change. Using Conceição-Duquesa Beach (Cascais, Portugal) a case study, we analyze over 92 years by applying novel methodology to postcards. The approach combines image registration, detection, coordinate transformation, rectification while accounting positional uncertainty. Results reveal significant counterclockwise rotation between 20th 21st centuries, exceeding estimated uncertainty thresholds. highlights feasibility using reconstruct positions quantify is straightforward, adaptable, offers promising avenue extending temporal range datasets, advancing our understanding

Язык: Английский

Процитировано

0
eo-tides: Tide modelling tools for large-scale satellite Earth observation analysis DOI Creative Commons
Robbi Bishop‐Taylor,

C. Phillips,

Stephen Sagar

и другие.

The Journal of Open Source Software, Год журнала: 2025, Номер 10(109), С. 7786 - 7786

Опубликована: Май 29, 2025

Язык: Английский

Процитировано

0
Shoreline evolution and morphological trends of Swarnadwip island through GIS approaches and Digital Shoreline Analysis System DOI
Numan Al Kibriya, Naila Matin, G. M. Jahid Hasan

и другие.

Marine Geodesy, Год журнала: 2024, Номер unknown, С. 1 - 36

Опубликована: Дек. 11, 2024

A comprehensive analysis of coastline changes was conducted over a span two decades for Swarnadwip, an offshore island Bangladesh in the Bay Bengal. Located near Meghna estuary, is influenced by vast sedimentary contributions Ganges-Brahmaputra-Meghna (GBM) river system. This makes Swarnadwip focal point research aimed at deepening our understanding coastal morpho-dynamics. In this research, three different shoreline detection strategies were explored, and suitability each method assessed to identify most optimum geographically complex region. Changes island's area calculated time interval, further change statistics (NSM, EPR LRR) estimated using Digital Shoreline Analysis System (DSAS) on GIS platform. The study revealed that from 2003 2022, length extended 50.6 74.15 km (1.24 km/yr), net land accretion 10,230.20 ha (538.43 ha/yr). Between 2006 2010, witnessed substantial expansion. However, much newly acquired mass lost again, as erosional forces governed 2010 especially northern western sections. These findings provide critical insights future management highlight need continuous monitoring safeguard Bangladesh's dynamic coastlines.

Язык: Английский

Процитировано

1
Data-Driven Shoreline Modeling: Historical Remote Sensing and Prediction Methods DOI
Mark Lundine, Arthur C. Trembanis

Опубликована: Янв. 1, 2024

This study describes a novel, data-driven approach for extracting satellite-derived shorelines, constructing shoreline change timeseries, and projecting future positions. First, to build datasets, an image-to-image translation generative adversarial network was developed map RGB satellite imagery into segmented land/water images, after which shorelines are extracted via contour tracing algorithm (Marching Squares). model incorporates over 20,000 LANDSAT 5, 7, 8 Sentinel-2 images from the Delmarva coast (USA), New Jersey Shore Long Island Northern Tuscany Littoral Cell (Italy), spanning 1984 present time. Next, shore-normal transects generated at set alongshore intervals, then used compute intersection points with construct timeseries of cross-shore position along given transect. With these yearly linear trends computed through ordinary least squares, maps illustrating across broad (several km tens km) coastal regions constructed. Last, long short-term memory networks incorporated on individual datasets learn behavior each project positions (with bootstrapped uncertainties). For select regions, projected transect merged together full two-dimensional uncertainty polygons), time period. All software components this open-source freely available, as trained models labelled imagery. In future, methodology should be extended other coasts worldwide, well integrate higher spatial temporal resolution

Язык: Английский

Процитировано

0