romanokhrimchuk/blacksea-coastlines

Mapping Black Sea's dynamic coastline at MSL using Sentinel-2 Imagery

This repository provides a visual representation of the Black Sea coastlinefor the period 2017–2024, derived using remote sensing (Sentinel-2 imagery) and modeled tidal heights at Mean Sea Level (MSL).

The core datasets are: annual coastline layers and a coastline change-rate point layer with a default spacing of 100 m.

I have also computed coastline change-rate datasets with a range of point spacings (in meters):

2, 5, 10, 20, 50, 100 (used in this repo as an example), 200, 250, 500, 750, 1000, 1500, 2000, 2500, 3000, 4000, 5000, 7500, 10000

If you are interested in any of these alternative point spacings, please contact me to access these datasets.

For an interactive overview of the derived datasets, visit:
--> Map Preview

Methodology

This repo presents a tide-aware workflow for extracting subpixel coastline positions from Sentinel-2 imagery archive (2017–2024) across the Black Sea region. Designed for use with Open Data Cube (ODC), the approach enables scalable, multi-year coastal monitoring across different coastal settings (Okhrimchuk, et. al, 2024). For each image, the Modified Normalized Difference Water Index (MNDWI) was computed to enhance the land-water boundary. Tidal heights, modeled at a 1/30° spatial resolution using the TPXO10-atlas-v2 global tide model (release date: 14 August 2024) (Egbert, et. al, 2022) and pyTMD library, were interpolated across the image stack based on time dimension for pixel-level tidal correction. To minimize tidal bias, a filtering procedure excluded observations outside the central 50% of the tidal range, retaining only images captured near median tide levels for consistent coastline detection. Annual composites were created by calculating median reflectance values from the tidally-masked stack, providing an estimate of coastline position at approximately mean sea level. coastlines were then extracted using optimized thresholding and segmentation for subpixel precision (Bishop-Taylor et al., 2021). The final outputs are a series of coastline products normalized to a common tidal reference, enabling unbiased comparisons of coastal change across space and time.
As shown in Figure 1, this method integrates open-access satellite imagery, global tidal models, and cloud-native geospatial tools. The resulting coastline datasets support consistent, scalable, and repeatable coastal analysis, offering a robust foundation for monitoring coastline dynamics along the Black Sea coast and beyond.

Figure 1. Subpixel coastline extraction workflow using ODC and tide modelling

Lexcube is used for 3D visualization of the data cubes shown in Figure 1. I highly recommend it to researchers and professionals working with multi-dimensional spatio-temporal data (Sochting et al., 2024)

The extraction of coastlines is based on a modified implementation of the methodologies developed for Digital Earth Australia & Digital Earth Africa platforms.

References

Bishop-Taylor, R., Nanson, R., Sagar, S., & Lymburner, L. (2021). Mapping Australia’s dynamic coastline at mean sea level using three decades of Landsat imagery. Remote Sensing of Environment, 267(112734), 112734. https://doi.org/10.1016/j.rse.2021.112734.

Egbert, G. D., & Erofeeva, S. Y. (2002). Efficient inverse modeling of barotropic ocean tides. Journal of Atmospheric and Oceanic Technology, 19(2), 183–204. https://doi.org/10.1175/1520-0426(2002)019%3C0183:EIMOBO%3E2.0.CO;2

Sochting, M., Mahecha, M. D., Montero, D., & Scheuermann, G. (2024). Lexcube: Interactive visualization of large Earth system data cubes. IEEE Computer Graphics and Applications, 44(1), 25–37. https://doi.org/10.1109/MCG.2023.3321989

Okhrimchuk, R., Demidov, V., & Sliusar, K. (2024). Innovative approaches to big Earth observation data processing in Earth science. In Proceedings of the International Conference of Young Professionals “GeoTerrace 2024” (pp. 1–5). https://doi.org/10.3997/2214-4609.2024510006.

Acknowledgements

I would like to express my sincere gratitude to the following people for their support and contributions that made this project possible:

1. Dr Robbi Bishop-Taylor
   I greatly appreciate his paper on the coastline mapping methodology and his contributions to the open-source Digital Earth Australia Coastlines codebase. His research, code development, and open tools strongly inspired me to work on this pet project.

2. Dr Svetlana Erofeeva
   I am deeply thankful for her work on the modeling of barotropic ocean tides and for providing me with access to the TPXO10-atlas-v2 model (release date: 14 August 2024).

3. Dr Kenneth Mubea & Technical Manager - Edward Boamah
   Many thanks to both of them for their technical advice and consultations, as well as for clarifying the implementation details of the coastline methodology within the open and free Digital Earth Africa platform.

Contact

If you have any questions, are interested in additional versions of the change-rate or annual coastline datasets, or would like to explore a collaboration, feel free to reach out:

• Email: romanokhrimchuk@gmail.com
• LinkedIn: https://www.linkedin.com/in/roman-okhrimchuk/