Utilising the power of the Copernicus Data Space Ecosystem to visualise EvoLand project results in no time

Utilising the power of the Copernicus Data Space Ecosystem to visualise EvoLand project results in no time

9 Jul 2024

Projects dealing with Earth Observation data often involve data visualisation. This does not only allow data producers (in our case, prototype developers) to gain different insights into their results but is also beneficial for dissemination to the public and reporting to project officers and reviewers. A picture is worth a thousand words, and an interactive visualisation even more: it also provides a more concrete and comprehensible representation of the work done that can be explored and compared to other sources of information.

Visualisation of the C05 prototype (Grassland Cropland GPP) and analysis of one field (1-year gross primary production) using Copernicus Browser

GPP_visualization_statistics

At the halfway point of EvoLand, we present the 11 prototypes after the first development cycle (still limited in temporal and spatial availability) and aim to visualise them in an interactive and user-friendly way. The solution should enable the reviewers and key stakeholders to interactively view the results while being easy to implement, so that the main focus and time could be on improving the prototype algorithms further.

The prototypes are hosted on a dedicated EvoLand S3 bucket provided by CloudFerro, which allows for easy and collaborative access. We explored solutions that could directly access and use this type of data storage, avoiding the need to duplicate the data elsewhere. Luckily, we didn’t have to look far as the Copernicus Data Space Ecosystem (CDSE) provided everything we were looking for:

  • The ability to directly access the data from an S3 bucket,
  • A straightforward way to reference/prepare the data for visualisation,
  • And an interactive way to visualise and explore the data using Copernicus Browser.

To get the data into Copernicus Browser we utilised the “Bring your own COG” (BYOC) functionality offered by Sentinel Hub within the CDSE. Since the prototype results were prepared in the GeoTIFF format, all we had to do was:

  • Transform the GeoTIFFs into Cloud Optimised GeoTIFFs (COGs),
  • Set the correct access rights to the S3 bucket so the Ecosystem could access/read the data
  • Ingest the tiles (reference them in the CDSE catalog).

We followed the instructions as described here.

C01 Continuous Forest Monitoring collection as ingested into the Copernicus Data Space Ecosystem (dashboard visualisation)

CDSE_dashboard

Once the data was ingested into Sentinel Hub, the final step was visualising it through Copernicus Browser. For this we created a dedicated test account (to restrict the data access to this account) and prepared nice visualisation options utilising Sentinel Hub’s Evalscripts, including custom legends for each of the prototypes. Adding legends is natively supported by the Browser, we simply had to define the color per label/class. A guide on how to set up custom visualisations in Copernicus Browser can be found here.

With all the data available within the CDSE, we could start with the fun part: closely examining and analysing the data using the Browser’s built-in functionalities.

With just a few clicks, we were able to:

  • Navigate the data on a map and compare it with different background layers (e.g., OpenStreetMap, Sentinel-2 mosaics)
  • Create nice timelapse animations (C10 Land surface characteristics example below),
  • Use the compare functionality to compare different dates and collections in the same view (e.g., gross primary grassland production with the corresponding Sentinel-2 image)
  • Calculate data statistics (time series, histograms)
  • Download the data (visualisation or raw data) for further use.

a 3-year timelapse animation of land surface prediction (C10 Land Surface Characteristics)

Using the CDSE as a service within our project, we were also able to give some valuable feedback to the developers, some of which was immediately implemented. This not only improved our data presentation but also contributed to a better experience for similar use cases in the future.

We believe this use case is a great example of how leveraging existing tools and functionality can create synergies, and we hope to see more of such cross-European project efforts in the future.

Utilising Copernicus Browser for visualising the EvoLand intermediate results not only provided the CDSE with another use case and valuable feedback but also enabled the EvoLand project to focus on value creation (improving algorithms, testing, etc.) rather than reinventing the wheel.

Cover image credit: European Union, Copernicus Sentinel-2 imagery: mosaic of the river Elbe entering the North Sea

Newsletter

Stay current! Subscribe to our EvoLand newsletter!