Sentinel 5 Precursor (S5P) was launched on the 13th of October 2017 carrying the TROPOspheric Monitoring Instrument (TROPOMI). TROPOMI on the Sentinel 5 Precursor (S5P) satellite observes the CO global abundance exploiting clear-sky and cloudy-sky Earth radiance measurements in the 2.3 µm spectral range of the shortwave infrared (SWIR) part of the solar spectrum. TROPOMI clear sky observations provide CO total columns with sensitivity to the tropospheric boundary layer. For cloudy atmospheres, the column sensitivity changes according to the light path. Carbon monoxide (CO) is an important atmospheric trace gas for our understanding of tropospheric chemistry. In certain urban areas, it is a major atmospheric pollutant. The main sources of CO are the combustion of fossil fuels, biomass burning, and atmospheric oxidation of methane and other hydrocarbons. Whereas fossil fuel combustion is the main source of CO at Northern mid-latitudes, the oxidation of isoprene and biomass burning play an important role in the tropics.

This dataset represents a collation of surface measurements of nitric oxide (NO), nitrogen dioxide (NO2), carbon monoxide (CO) and isoprene (C5H8) from publicly available data sets (of hourly, daily and monthly resolutions), for the aim of improved evaluation of surface ozone in global atmospheric chemistry models. Measurements begin in 1980 running through to 2015. The data comes in a range of formats, with a plethora of associated data quality issues, requiring substantial cleaning before being able to be utilised for model assessment. 1,033,463,750 measurements from 16,996 sites are processed through numerous data quality checks, resulting in 76,413,458 observations from 1607 sites of appropriate quality (with the majority of excluded observations due to urban influence). Observations are heavily weighted towards North America and Europe, with generally sparse coverage over the rest of the globe (with the exception of CO). See documentation for more details. Data is provided as multiple globally gridded output files, each consisting of a series of metrics designed to reflect the distributions of the observed ozone precursor species, allowing fair and easy comparison with global models. Metrics include the moments of the distribution (i.e. mean, temporal standard deviation, skewness and kurtosis) and percentiles. A total of 80 different netCDF-4 files are produced, with metrics calculated in multiple temporal (monthly and annual) and spatial configurations (8 different resolutions), for each different species. The format of the output netCDF-4 files is designed to be consistent with the related dataset which compiled surface ozone observations (v2.7).