Challenge 12 - ObsFinder: An interactive geo-selector for C3S observation data

[DiegoDomenici]

Challenge 12- ObsFinder: An interactive geo-selector for C3S observation data

Stream 1 - Data Visualization for Earth Sciences Applications

Goal

ObsFinder aims to provide an interactive web mapping application for finding (and downloading) in situ observation data available via the CDS.

We would like a Single Page Application (SPA) which allows users to explore the locations of the in-situ observations based on simple selection criteria (time-period, observable). The application should be to create CDSAPI requests based on the geographic interactive selections.

Mentors

Edward Comyn-Platt, Gionata Biavati, Paul Poli (all ECMWF)

Skills Required

• Experience with web frontend languages, e.g. JSreact, TypeScript
• Experience with web mapping software
• Experience with python
• Desirable: Experience with building applications as docker images and deploying with Kubernetes and Helm

---

Description

The Copernicus Climate Change Service (C3S) has made available a wealth of in situ observation data via the Climate Data Store. Observations from large number of stations (tens of thousands), cover multiple observables (e.g. temperature, wind) and a wide range of sources and platform types (e.g. ground-based weather stations, atmospheric soundings). However, visualizing available data that are irregularly distributed across observables and in space and time is a non-trivial challenge. The current CDS catalogue entries provide users with a simple area/time-range selection option, but this gives no indication of the data quantity. Users must first download the data before they can identify which station(s) may be of particular interest for a given task/downstream application.

This project will use in situ observation data available from the CDS, along with some inventory files (stations’ locations, variables and time period observed). The single page application will be built and deployed following examples of other similar applications produced under C3S.

A proposed solution is an interactive map which displays the locations of the in-situ observations. Users could select a single station, a list of stations or a rectangular region and the application should produce a cdsapi request which can be used to download the data. If time permits, the application could also include some form of visualisation of the returned data, e.g. time-series for point data.

Some ideas/thoughts for implementation:

• An interactive map which displays the station locations, and allows users to select a single station, or a grouping of stations.
• Displaying all the stations (thousands) on a single map will make for a poor user experience. Besides cluttering the interface, it will also result in performance issues. Methods for grouping station locations (server-side, and client-side, e.g. based on zoom level) will be needed, so that individual pins are only shown when relevant.
• Able to choose the dataset of interest, each of which will have their own array of station locations.
• The app should produce a cdsapi request which users can run locally
• Desirable, the application could submit the request and the user can collect data when it is ready for download.
• Desirable, the application could provide some visualisation of the requested data.

---

Evaluation Criteria

• Feasibility
• Transferability
• Easy to maintain / Future-proof approach
• Matching requirements

[irmariad]

Hi, we are a development team interested in this challenge, and we have some questions regarding the data:

• The CDS catalog for in-situ observations includes different data types, some are point-like (station-based observations), while others are gridded, derived from the interpolation of observations. The challenge description mentions the visualization of stations, so we assume that only point-like datasets are relevant. Could you confirm this?

• Additionally, the proposed solution mentions a rectangular region for selection. Could you clarify what this refers to? Does it mean retrieving the data for all stations within that region, or is it related to gridded data?

• Finally, for writing our proposal, it's important for us to know where the inventory files are stored. Is there a way to query an endpoint to retrieve the stations metadata, or should we rely on a CSV file or another format?

Thank you in advance for your help.

Irma Riádigos

[EddyCMWF]

Hi @irmariad ,

Thank you for your interest in the project.

1. For this project we are interested in the point data from the in situ stations, for example the GNSS dataset:
   https://cds.climate.copernicus.eu/datasets/insitu-observations-gnss
2. for the rectangular area, we would like to retrieve all the station data within that area
3. Potentially... At the moment we have the list of stations at the end of an API endpoint, but this does not contain the associated latitude/longitude data. We may be able to add this to the API, but failing that (and to get started) the lat/lon data (and other station metadata) will be provided in a CSV/JSON file.

I look forward to reviewing your proposal.

Eddy