User Roles & Goals
This page is for describing the types of user that vAirify is likely to have, and for each help us think about:
- Their role / job
- What their goals are
- What their frequent tasks might be
- The tools they use for those tasks
- Their frustrations / pain points
By understanding these we can help maximise the value of vAirfiy.
Works in Production Services within Forecast Department (see the ECMWF How we work page for the organisation structure), and is focussed on air quality forecasts.
- To make sure that forecast data is available, accurate and useful.
- To be able to quickly detect inaccuracies / successes.
- To see a trend of improvement, that there are fewer outliers. For weather forecasting in general there is the concept of a "forecast bust", where the prediction is significantly different to what happens, if something similar can be defined and measured for air pollution forecasts then this could be useful.
To respond to reports of data inaccuracies.
Validation reports of the CAMS global forecast model are available at https://atmosphere.copernicus.eu/eqa-reports-global-services. These are done quarterly, at the time of writing (23rd May) the most recent report covered the period up to end November 2023.
There is also a pair of CAMS forecast evaluation portals. Firstly https://global-eqc-server.atmosphere.copernicus.eu/, the data sources include https://airnow.gov/ which provides current air pollution readings for the US. An example of the data available is shown below, which is for PM10 particles in Vancouver, screenshot taken on 30th May 2024.
Secondly https://global-evaluation.atmosphere.copernicus.eu/ - there is a delay before evaluation data is uploaded, for example for in-situ Carbon Monoxide it is a one month delay. Data is available for Europe & China. An example of the data available is shown below, screenshot taken on 30th May 2024.
The AirNow site includes maps of the air quality in the US:
(The red area in the west was due to an intentional forest burn near Bend, Oregon)
A few other network-specific evaluation activities and international model inter-comparison results are listed at https://atmosphere.copernicus.eu/global-services.
The tools above are provided by ECMWF partners. In addition to these there is also an internal tool that has been developed directly by the ECMWF, plus a number of locally written Python scripts, which are used on-demand when an issue is raised.
There is a significant delay between the occurrence of a large variation between a forecast and an actual reading, and that variation being known to the ECMWF.
Official air quality measurements do not have good coverage of the world, which limits the model evaluation that can be done. At the moment it's not known whether air quality measurements from other parts of the world will be of suitable accuracy to be used as a reference for model evaluation.
A member of the Atmospheric Composition group within the Research Department, working on the Copernicus Atmosphere Monitoring Service. Leads the activities related to monitoring global atmospheric composition and air quality in relation to emissions and transport of pollutants from sources such as forest fires, dust storms and volcanic eruptions. Also works on science communication activities including media and training related to measuring and monitoring atmospheric composition.
- To make sure that forecast data, and other CAMS operational data, is available, accurate and useful.
- To be able to quickly detect inaccuracies / successes.
- To highlight things of interest to the media.
At the beginning of each day checks a variety of charts to identify issues and points of interest. Is interested in doing on-the-fly evaluation using existing tools and observations and analysing further if they don’t look right, and is on the look-out for things that can be pushed out to the media. This activity is also used to support User Support and Engagement in CAMS.
- Global Evaluation
- Regional Evaluation
- European Environment Agency air quality measurements
- Copernicus Charts (e.g. for overall Aerosol Optical Depth, for individual chemistry and for fire activity analyses)
In some cases it can take a long time before ECMWF/CAMS knows about differences between forecast and actual air pollution, the public know first, which means that the ECMWF is on the back foot. Routine monitoring focusses on a limited number of forecasts/products (e.g. aerosol for wildfire, desert dust and volcanic eruptions), sometimes at the expense of other forecast products which would benefit from availability of a simplified and quick overview.
It’s currently possible to check, but it means downloading everything and comparing by hand, which is time consuming.