Challenge #13 - ORIGAMI (global river gauges mapping)

[EsperanzaCuartero]

Challenge 13 - ORIGAMI (global river gauges mapping)

Stream 1 - Software development for weather, climate and atmosphere

Goal

Develop a software to facilitate the mapping of river gauges on the drainage network of a distributed hydrological model.

Mentors and skills

• Mentors: @ecCinziaMazzetti @zservin32 @colonesej
• Skills required:
  • Image analysis
  • Pattern recognition
  • Python

---

Challenge description

Data and software
We plan to use stations coordinates and drainage networks from CEMS (EFAS and GloFAS) and HTESSEL at different spatial resolutions. We plan to use OpenStreetMap to identify rivers and derive metadata.

What is the current problem?
River gauges are mapped on the hydrological/land surface model to the pixel that better represents the location of the station in the model’s drainage network. This procedure currently uses river gauges coordinates and drained area information from the station and the model. It tries to identify the model pixel with the closest drained area in the proximity of the station’s coordinates. This approach can lead to large mistakes because of uncertain/missing information on the station’s drained area or issues with the model drainage network, sometimes occurring hundreds of kilometres away from the location of the station. Difficulties in stations mapping increase at coarser model resolutions and with the complexity of the drainage network. An accurate mapping is often only possible through a manual procedure, but this can cause long delays in the adoption of a station for either model calibration or verification.

What could be the solution?
We are looking for a solution that won’t make use of the drained area information but will instead mimic the human mapping procedure. We would like to use image analysis and/or pattern recognition techniques to match the real river to the one in the model and then map the station on the correct model pixel, also exploiting additional metadata such as the station name or the river name. The mapping of each station should include a quality flag showing a confidence level in the mapping result.

Ideas for the implementation
We envisage that implementation will include the following steps:

• extract the river map for the area surrounding the station and the available metadata, such as rivers names, from OpenStreetMap or any other open dataset.
• match the river image to the image of the model drainage network
• map the station using coordinates and metadata (like the name of the river or the name of a nearby location).

[jwagemann]

Hi,
join us for the ECMWF Summer of Weather Code Ask Me Anything session and learn all things ESoWC.

When: Wednesday, 24 March 2021 at 4 pm GMT

What: learn everything about ESoWC - how it works, the challenges this year, some tips for your proposal and listen to ESoWC experiences from previous participants

How: register here.

[tusharmanekar]

Dear mentors,
I had some queries about this challenge (Challenge #13 - ORIGAMI (global river gauges mapping)). My queries are as follows-
From what I understand this challenge can be broken into two parts- the pattern recognition from image analysis to mimic the human mapping procedure and the software development part.

1. Are Machine Learning and Deep Learning methods suitable for the pattern recognition and image analysis part? (Can you suggest some reading on it?)
2. What is the level of sophistication and complexity of the software expected? (Can you give an example of a relatable and satisfactory software?)

Thank you in advance.

[EsperanzaCuartero]

Dear Tushar,
Thanks for your interest. The mentors of the challenge will be in contact as soon as possible. Best, Esperanza

[colonesej]

Dear @tusharmanekar ,
Thanks for your interest. About your questions:

1. With this challenge we are looking for someone with some expertise in the machine learning field to help us. So, choosing the techniques is part of the work;
2. We do not expect a high level of complexity in the resulting software. A python package with a command line tool would be good enough.

[tusharmanekar]

Dear mentors,
Is this challenge funded by copernicus and thereby adhere to their rules and regulations (eg. At least one participant of the group must be an EU resident)?

Thank you.

[EsperanzaCuartero]

Dear Tushar, this challenge is not funded by Copernicus. Therefore, the regulations you have just mentioned don't apply to this challenge. You can submit your proposal regardless of your nationality.