toar2_category has been made available

In the first assessment of TOAR, stations were classified in a globally uniform way based on a set of Earth Observation (EO) datasets that have been processed at the station locations. The method to determine the toar1_category used manually selected threshold values of station altitude, population density, nighttime light intensity, NO2 column density, and NOx emissions to characterize stations as urban, rural, or unclassified. While this approach provided a useful distinction between 'clearly urban' and 'clearly rural' locations, it could not classify all locations and was criticised for lacking an objective definition of thresholds.

Therefore in TOAR-II, a new machine learning (ML) approach has been developed to obtain a more advanced and unbiased classifier using similar objective metadata from the TOAR-II database.
The results of this approach need to be added to the TOAR-II database as part of the stations's global metadata as the toar2_category.

The method itself will also be added to the GeoPEAS ("Geodata Point Extraction and Aggregation Service").

The findings of this ML method will soon be published in a paper by
Ramiyou Karim Mache, Sabine Schröder, Michael Langguth, Ankit Patnala, and Martin G. Schultz.