VALIDATION OF MIPAS TEMPERATURE DATA WITH THE U. BONN LIDAR AT THE ESRANGEDURING JULY AND AUGUST 2002

摘要

The Bonn University lidar is located at the Esrange (68rnN, 21 E) in northern Sweden, near the city of Kiruna.rnDuring July/August 2002 a measurement campaign forrnthe validation of Mipas, Gomos, and Sciamachy data wasrnperformed. During 41 measurement runs a total of nearlyrn350 hours integration time was accumulated. Most ofrnthe measurements could be used for the calculation ofrntemperature profiles in the aerosol-free part of the atmosphere,rnwhich is above 30 km altitude.rnFor the period July/August 2002 we received 56 Mipasrnfiles containing temperature data, processed with the softwarernMIPAS/4.61 for the Esrange location. The range ofrnvalues encountered in these files are: tangent point distancesrnfrom the lidar range from 60 km to 995 km, altitudesrnrange from 7 km to 70 km, temperatures rangernfrom 207 K to 283 K, and temperature errors range fromrn0.3 K to 2.7 K. These data ranges look quite reasonablernand self-consistent.rnOut of these 56 Mipas temperature profiles we could usern33 profiles for validation. Selection criteria were the simultaneousrnspatial and temporal coincidence of the Mipasrnand lidar measurements. The time window was met,rnwhen the lidar measurements started or ended within onernhour about the Mipas measurement time. We used twornspace windows. The first window comprised all datarnwithin 500 km of the Esrange (in total 11 profiles) and thernsecond window consisted of all profiles within 1000 kmrnof the Esrange (in total 33 data-sets). We interpolated thernlidar data to each Mipas corrected altitude.rnThe comparison of all available Mipas-Lidar temperaturernpairs in the altitude range 30 to 70 km showed mean valuesrn(a measure for the accuracy) for the temperature differencernof 0.3 % (≈ 0.7 K) and 0.7 % (≈ 1.6 K) in thern500 km and 1000 km tangent point range windows, respectively.rnA t-test revealed that these differences cannotrnbe attributed to chance, but must be real. The standard deviationsrnof the distributions (a measure for the precision)rnare of the order of 2 % (≈ 5 K) for both range groups.rnWhile the Mipas temperatures do not agree on averagernwith the lidar temperatures for the available data-set, therndeviation is very small and fits well with the targeted accuracyrnfor the Mipas temperature data.