Challenge 32 - Wildfire Emission Explorer

[jwagemann]

Challenge 32 - Wildfire Emission Explorer

Stream 3 - Applied data science for weather, climate and atmosphere

Goal

Develop an application that will be capable of creating wildfire emission and activity plots on-demand.

Mentors and skills

• Mentors: Mark Parrington, Sebastien Garrigues, James Varndell, Miha Razinger
• Skills required:
  • GIS, PostreSQL (PostGIS)
  • Python (geopandas, matplotlib, cartopy)
  • Jupyter (ipywidgets, dashboard)

Note: Challenge is funded by Copernicus. Only nationals from European Union (EU) Member States and countries associated with EU’s Space Programme (currently Iceland and Norway) are eligible to participate (see Terms and Conditions).

---

Challenge description

The CAMS Global Fire Assimilation System (GFAS) assimilates fire radiative power (FRP) observations from satellite-based sensors to produce daily and hourly estimates of biomass burning emissions.

To analyse the significance of a particular wildfire episode, CAMS scientists prepare various graphical products which put the emissions in historical context. The CAMS data and graphical products on fire emissions are regularly shared with users via social media and training activities, and with the global media when reporting on fires.

Examples of current fire plots

Fire Radiative Power, a measure of fire activity, in 2020 (red) and 2019 (yellow) compared to the 2003–2018 average (grey)

Summer Arctic CO2 emissions due to wildfires from 2003 until 2020

The aim of this project is to create an application that would simplify and speed-up the creation of various wildfire emission plots based on a subset of the dataset.

We have some ideas on how to build such an application (see Skills required) but we are inviting candidates to propose their own ideas on the technical implementation details.

Current procedure
The current manual procedure for a new country/continent/region looks like this:

• download gridded daily GFAS emissions data from MARS
• convert gridded emission fluxes (kg/m2/s) to gridded daily total emission (kg) - or daily total FRP
• extract all the pixel values within a specific geographical boundary, compute totals and store them as text files
• calculate climatology
• create a fire emission or activity anomaly plot
  The problem is that one needs to recalculate the background statistics/climatology for every new geographical domain and new time period.

Input data
Sample data:

Expected outputs
Along with the types of plots that we already produce, we would like to be able to extend our capabilities to create plots showing anomalies of fire activity or wildfire emissions similar to these:

User interface
A user should be able to select plot type, date period for the reference period, date period of the specific episode and geographical domain, i.e. bounding box, a country from a drop-down list, a specific region from free text search (using Nominatim?) by using an interactive user interface.

[Giov-P]

Hello!
I am quite interested in this challenge but I realized I have a few last minute doubts before submitting the proposal, so here's a few technical questions:

• In this challenge you are using two variables in the plots, Radiative Power and Flux of CO2: will there be additional variables to handle apart from those (in the CAMS GFAS there are around 50 different variables)? Do you see any value in using multiple variables to create additional plots (correlation plots)?

• From the description, climatology is probably the bottleneck of the process: do you want a solution to this problem from this challenge?

• What are exactly the statistics needed from the reference period in order to produce anomalies (mean and std)?

• You mentioned the use of PostGIS, will the data be stored in an online PostGRES database? from what I understood the data at the moment are queried from GFAS and downloaded locally for the analysis, so no database is present.

Thanks in advance for the support!

[miha-at-ecmwf]

Thanks for your interest in this challenge, @Giov-P.

• The main goal of this challenge is to build a simple example using just the two variables. If the application is successful, it should be possible to extend the scope later on
• Indeed, computing climatology over user-defined regions on-the-fly in an efficient manner is the most difficult part of this challenge. We've done some preliminary tests, and we feel that this step should be feasible by using PostGIS spatial queries
• Again, we want to start simple (average, stdev, min, max, median) and have an option to expand the capabilities after the ESoWC project has been completed
• Creating a pipeline which would fetch the data from a data server, process it and insert it into a database should be part of the activity. If you have any alternative suggestion how to achieve the project goals, we would be happy to hear about it