Paul Tol's Notes

Movies of atmospheric CO and methane from TROPOMI data


24 September 2020

The movies here show the CO and methane column mixing ratios retrieved from SWIR data of TROPOMI. They were made in my spare time to see what my calibration work is used for. Both CO and methane mixing ratios are shown on world maps and CO on seven regional maps as well, all in equal-area map projections. The image pixel distance is on average 19 km (the image pixel area is 348 km²) on the world maps, everywhere 7.5 km on the CO map centred at the north pole and everywhere 5 km on the other six regional CO maps. The minimum distance between ground pixels in SWIR data of TROPOMI is about 5 km, so zooming in further would not show more details.

The movies can be used as a survey of the available data without strict filtering. The CO data are shown per orbit and the methane data averaged per day. There are long periods without data before March 2018 due to calibration measurements for commissioning. Data gaps show previous data made whiter, to be able to see trends while making clear which data are new. Think of a submarine sonar screen. In the end, old data that are never replaced become completely white.


2 January 2022

The more remarkable clips of CO data are put together in a showcase movie.

Recent CO data

CO data in the period since 1 January 2021.

CO data selection map
Click or tap the label of one of the 8 regions to get the corresponding movie.

Earlier CO data

CO data in the period from the start of the TROPOMI mission in 2017 to the end of 2020.

CO data selection map
Click or tap the label of one of the 8 regions to get the corresponding movie.

A separate movie shows the Australian fires of November 2019 to January 2020. There is also a version showing the area in that period and the same period the year before, side by side. Zooming out, the Australian CO clouds actually moved around the world.


2 January 2022

Methane data in the period since 1 January 2021 (left) and from the start of the TROPOMI mission in 2017 to the end of 2020 (right). Only movies of the whole world are available.

methane data selection mapmethane data selection map
Click or tap one of the 2 plots to get the corresponding movie.


14 January 2022

On the methane world map, an image pixel is coloured when there are data within 50 km, but a third-order interpolation is used where possible. The result consists of big dots around sparse data and fine details within dense data. The interpolation is applied on the data grid formed by time and across-track viewing angle, not on the ground, because on the used grid the instrument point spread function is almost invariant. In CO maps the same interpolation is used, but image pixels are only coloured when the nearest data grid point has a valid value. No smoothing is applied. The colours used come from one of a set of colour schemes designed by me.

For CO the operational data are used, without applying postfiltering. For methane the data of the latest operational processor (version 2.3.1) and of in-house processing with the same settings are used, corrected by multiplication with (1.02469 − 0.162320×A + 0.213169×A² − 0.0942168×A³) above land and 1.00385 above oceans with retrieved SWIR surface albedo A. Methane data are only retrieved when the cloud fraction according to VIIRS is less than 0.02 and a few other restrictions are fulfilled. Postfiltering deviates from the official version to keep more useful data: methane data are used when the SWIR aerosol optical thickness is smaller than 0.13, the methane precision is smaller than 10 ppb, the reduced chi-square is smaller than 100, the SWIR detector temperature deviates less than 1 K from nominal and the methane column using data at 2310–2315 nm deviates less than 10 % from the column using data at 2363–2373 nm, both according to preliminary nonscattering retrieval. When the VIIRS cirrus reflectance is not available (and hence the VIIRS cloud mask is unreliable), methane data are only used when additionally the aerosol optical thickness according to operational CO retrieval is smaller than 1. The remaining areas of very low methane values should be ignored.

Paul Tol

I am a Dutch instrument scientist with a PhD in atomic physics, working on the TROPOMI calibration and the SPEXone project in the Earth programme of SRON. My email address: