Comparison of CAMS reanalysis data with surface methane measurements in the high Arctic

This study examines the success of the surface methane reproduction by the Copernicus Atmosphere Monitoring Service (CAMS) EGG4 global greenhouse gas reanalysis in the Arctic region by comparing it with continuous measurements at the research station "Ice Base Cape Baranova" (79˚16′ N, 101˚45′ E) in 2015–2020. Verification of the CAMS database in this way will be useful before its subsequent application in numerical modeling. 

The station is located on Bolshevik Island of the Severnaya Zemlya archipelago in the central sector of the Russian Arctic between the Kara and Laptev seas. Continuous methane measurements were carried out using a Picarro G2401 gas analyzer using cavity ring-down spectroscopy (CRDS) in a joint collaboration between the Finnish Meteorological Institute (FMI) and the Arctic and Antarctic Research Institute. The reanalysis data are provided on a grid with a resolution of 0.75° × 0.75° (~ 80 km) with a three-hour time interval. The reanalysis data for the coordinates of the considered station were extracted using bilinear interpolation. A total of 14,872 pairs of reanalysis and instrumental observations were considered for the analysis. 

In 60.4% of cases, the CAMS reanalysis values ​​are smaller than the station measurements, while the average deviation of measurements from the reanalysis data and its standard deviation are 16.7±14.5 ppb (0.85%). The excess of measured values ​​over the reanalysis results is also typical for other high-latitude stations, Alert and Barrow, which is noted in the reports on the verification of the CAMS database. The maximum deviation of the observed values ​​from the reanalysis data is 161.1 ppb (7.5%), and was recorded in February 2017. The number of cases when the modulus of the difference between the measured and modeled values ​​is more than 100 ppb does not exceed 0.2%. Such discrepancies are observed mainly in the winter months, as well as in August and September. The majority of the differences between the observed values ​​and the CAMS reanalysis results (43.1%) are in the range from -10 to 10 ppb.

The simulated and measured observation series correlate well with each other: the correlation coefficient R for the entire observation period is 0.82±0.01. However, in some months, worse comparability is observed. The lowest values ​​of the correlation coefficient were noted in March 2016 (R = 0.55), July 2016 (R = 0.52), April 2019 (R = 0.41), March 2020 (R = 0.47) and June 2020 (R = 0.46). 

In 2019 and 2020, the reanalysis results are significantly lower than the measured values ​​compared to other years - thus, the average annual difference between the measurement data and the reanalysis for these years are 14.0 and 11.7 ppb, respectively (versus -4.6 in 2016, 1.8 in 2017, and 1.0 ppb in 2018). The estimate of the interannual methane increase based on the reanalysis data for the period under review is generally poorly comparable with the estimate based on the observational data, and does not allow for unambiguous conclusions to be drawn about the methane variability from year to year.

The seasonal variability of methane is well reproduced by the reanalysis data, with CAMS data being higher than the observed values ​​from December to April inclusive, and lower from May to November. The maximum deviations are observed from August to October, which may indicate that the reanalysis data underestimate the contribution of regional or local-scale sources to the warm period of the year.