Improving water body mapping by means of novel EO-data and super resolution

To overcome these challenges, innovative approaches such as the use of Earth Observation (EO) technologies have been widely applied by the scientific community leading to a revolution in how society studies water bodies on both a local and a global scale.

Water bodies mapped within the framework of the Copernicus Land Monitoring Service (CLMS) and its components (e.g. Copernicus Global Land Operations Service or CGLOPS) include a wide range of types, such as natural lakes, dam reservoirs, and marshes, among others. However, their observability by remote sensing is strongly influenced by their size, resulting in datasets being subjected to limitations in terms of spatial and temporal resolution.

The Horizon Europe-funded EvoLand project aims to address these limitations by developing and testing an innovative prototype for water body monitoring, whose objective is to improve the spatial and temporal resolution of current water body mapping methods.

At the moment, the CLMS offering includes the High Resolution Layer Water and Wetness service, which provides a 10-metre resolution across pan-European countries, based on three years of data from both Sentinel-1 and Sentinel-2 satellites. In contrast, CGLOPS delivers a global-scale product at a 100-metre resolution on a monthly basis, utilising only Sentinel-2. EvoLand seeks to progressively enhance the spatial resolution proposed by the existing products, aiming for a final super-resolution of 5m with a monthly temporal resolution. Additionally, the methodology includes evaluating the operational feasibility of  super-resolution approaches.

The finer resolution achieved will enable the monitoring of smaller water bodies with greater precision and regularity. To understand the significance of improved resolution, note the difference between the figures below. Both images were produced as false colours of the same area (Camargues, France) and on the same day. The left image is Sentinel-3 at a 300m resolution and the right is Sentinel-2 at a 10m resolution. It is evident that the clarity of the finer resolution can help in analysing the area at a more precise level.

In addition to improving spatial resolution, EvoLand is exploring the use of novel EO data to further enhance the temporal resolution of the prototype. One noteworthy novel EO data source comes from the Surface Water and Ocean Topography (SWOT) satellite, which was launched in December 2022. This groundbreaking satellite revolutionises hydrology by providing insights into water height, slopes, and surfaces. SWOT’s global coverage and frequent revisits will provide new insights into areas that have never been properly observed in terms of water dynamics. Another example can be seen in using Sentinel-3 products, often overshadowed by the use of Sentinel-1 and 2 for the production of hydrological datasets. The novelty here lies in considering Sentinel-3 products, as they provide daily measurements at a nearly global scale, presenting immense potential for water body mapping when combined with super-resolution techniques. However, it is worth noting that this last approach is not a priority for the prototype’s creation but rather will serve as an additional means to evaluate the accuracy of the new dataset.

To ensure the data reliability and usefulness, the EvoLand project is also in charge of the validation processes. The results of the water body mapping will be compared against existing datasets such as Land Use and Coverage Area frame Survey (LUCAS) and Global Land Survey. However, to assess the quality of the products, especially concerning spatial resolution improvements, high-quality validation data are needed. Very High-Resolution (VHR) data are crucial for this purpose, as they provide detailed information on water body characteristics. EvoLand aims to produce a validation dataset from VHR data to evaluate the accuracy of its products effectively.

Eight prototype sites were pre-selected for tests and validation, including locations both within and outside Europe. Testing is expected to commence in the second quarter of 2024, with specific details to be disclosed in future announcements.

Would you be interested to follow the results of EvoLand and the development of this candidate prototype? Sign up for our newsletter (here below or via this form), follow us on LinkedIn, Mastodon or X (Twitter).

Questions about the article or the project? E-mail us at contact@evo-land.eu.

This article is part of a series providing more details on all 11 EvoLand candidate prototypes.

Previous article in the series: EvoLand candidate prototype to offer Enhanced Forest Monitoring

Image credits: European Union, Copernicus Sentinel-2 imagery