Chapter 5 - Deposition and erosion dynamics in Axios and Aliakmonas river deltas (Greece) with the use of Google Earth Engine and geospatial analysis tools

Abstract

River deltas are known as one of the most complex ecosystems, and they are crucial for both agriculture and the environment. The structural changes in the delta fronts can result from any human or natural action that upsets the equilibrium between water and land, making these environments susceptible. Earth observation (EO) offers a major potential for observing the dynamic changes of river deltas due to its ability to produce methodical, cross-temporal, and cost-effective data. The evolution of cloud-based platforms in recent years for EO processing, such as Google Earth Engine (GEE), offers special benefits including quick and low-cost processing of enormous amounts of EO data. This study aims at assessing the added value of GEE in monitoring the coastal surface area of deltaic basins using the complete Landsat archive (TM, ETM+, OLI, L9) and a machine learning (ML) technique. Two river deltas in northern Greece, Axios and Aliakmonas, were chosen as a case study. Their combined delta plains form a lush valley with environmental and agricultural significance. Over the past few decades, this valley's structural characteristics have also shown very strong dynamics. Landsat multispectral data covering the period from April 1986 to the present were merged into the cloud-based system, GEE, where a ML classification approach was implemented allowing to acquire a better understanding of the coastal dynamics of the examined regions. The dynamics of the two rivers' deltas were likewise independently traced utilizing photo-interpretation as a validation dataset. A geographic information system (GIS) was used for the implementation of geospatial analysis approaches to quantify the morphological processes of the studied river deltas. Our findings demonstrated the added value of cloud-based platforms like GEE for operationalizing the researched methods for deltaic morphological alterations like those seen in the analyzed river deltas. The combined use of the GEE cloud-based platform, the ML algorithm for image processing, and exploitation of the presently available Landsat imagery are the adopted herein methodology's distinctive features. The proposed herein method can also be totally automated, applied to other similar regions, and be of great assistance in comprehending the spatiotemporal alterations in coastline surface area over substantial areas.