Data Archiving

An important feature of the 12-pack sampling system is that every 12-pack carries a log of sampling time, position and, when available, temperature and relative humidity. This means that once the flask has been analyzed for greenhouse gases, carbon isotopes, halocarbons and hydrocarbons the measurement value can be immediately recorded in the database and compared with surrounding sites and previous profiles.

For a good portion of the aircraft dataset a global positioning system (GPS) was not available. Without GPS the sampling time was derived from the individual 12-pack. To ensure that the 12-pack time closely matched the Greenwich Meridian Time (GMT) each 12-pack clock was reset before being shipped from Boulder to the site. Because the internal battery of the data logger was sensitive to being stored at low temperatures (<0°C), times have been corrected manually based on the date and time of flight according to pilot records. In April 2006 a procedure was initiated to record the difference in time between the 12-pack clock and GMT when the 12-pack was returned to Boulder after a sampling flight. For more than 99% of the 12-packs returned to Boulder the offset has not been greater than 6 minutes. For the remaining 1% of samples collected, the offset after being manually corrected is likely to be less than 2 hours between the 12-pack clock and GMT.

For flights where the GPS was not available, the ground position and altitudes have been inferred from the original sample plan specific to each site (site list). Unless a pilot reported having been unable to reach a sampling location or altitude for a specified region, it is assumed that the original sampling plan has been followed. Because a typical profile is done by flying to the maximum altitude at the latitude and longitude of the sampling site followed by a slow downward spiral, leveling off at each sampling altitude, it is expected that the samples are taken within 0.1 degrees of the site location. GPS data collected more recently confirm this to be the case. It is also assumed that the pilots have used aircraft altimeters or pressure altitude to guide their sampling. Above 18,000 feet (5486 m) the pressure altitude is based on the American Standard Atmosphere (ASA, Sissenwine et al.), which assumes a constant sea level pressure of 101.325 kPa and constant lapse rate of 6.5 C/km up to 11,000 m. Below 5486 m the pressure altitude is determined using a sea level pressure, which varies with time and space and a constant lapse rate of 6.5 C/km. The local sea level pressure is estimated by ground control based on a local pressure measurement and an altitude correction, which assumes a lapse rate of 6.5 C/km down to sea level.

Although all sites now have GPS, much of the dataset described assumes that the pilot was guided by the sample plan and altitude measured by altimeter. To correct pressure altitude to geopotential height, we have utilized the NAM (Eta) Data Assimilation System (EDAS) ( for sites in the continental U.S., and the global NCAR/NCEP reanalysis ( for sites outside the continental U.S. (PFA, RTA, HAA, ULB) to derive the geopotential height. For each sample the ambient pressure is derived based on the target altitude and standard FAA protocol for altimeters. Using the EDAS or global dataset, the ambient pressure is converted into geopotential height. Below 5486 m, the local sea level pressure is also derived from local ground pressure from EDAS and local ground elevation. This method was tested using two years of data at the Carr, CO site (CAR). As expected the pressure altitude was as much as 600 m below the GPS measured altitude during the hotter summer days; the offset increased with altitude. Below 5486 m the average offset between the pressure altitude and GPS was -80±80 m (Figure 1a). After the correction was made the offset was 8±27 m suggesting that the derived geopotential height was not significantly different from the measured (GPS) height (Figure 1b). Above 5486 m, the mean offset between the pressure altitude and GPS was 220±150 m. By using the meteorological re-analysis the offset between estimated and measured altitude was 7±21 m. Although there is less GPS data at other sites, similar trends have been documented when the GPS data can be compared with estimated altitude derived from pressure altitude and reanalysis data. This altitude correction has only been made to data with a sample date on or after January 1, 2004.

Figures 1a & 1b. Comparison between sampling plan pressure altitudes and GPS position. a) illustrates difference before altimeter heights were corrected by re-analysis data and b) figure illustrates variability (colored by month) in offset after corrections made by re-analysis data. Text illustrates the average offset between 2000masl and 5400masl and between 5400masl and 8500masl.