PEM Tropics-A Nominal Measurement Characteristics of Instruments on the P-3B

NOTES:

(1) Trace gas data are based on measurements of whole air samples collected by D. Blake. The analytical technique was gas chromatography with mass spectrometric (EI and NICI) detection.

(2) The detection limit corresponds to 3 times the background level. Random fluctuation in zero, measure and artifact are all included.

(3) Errors correspond to one sigma values.

(4) Approximately half of uncertainty in precision is due to uncertainty in the artifact.

(5) During flights 4, 6 & 7, an instrument malfunction caused a systematic error of uncertain magnitude. The reported mixing ratios for ozone are lower than the real mixing ratios. Most likely the numbers are 2.5% to 20% low. There is a small chance that the reported mixing ratios are off by more than 20% but they cannot be more than 39% low. During flight 8 a similar problem occurred. Most likely the reported mixing ratios are <1% to 5% low. There is a small chance that they can be off by more than 5% but they cannot be more than 10% low.

(6) Higher uncertainty for molecules with mixing ratios in the 0.001 - 0.1 pptv range

(7) Reported error limits are +/- 42.0% for measurements made at pressures above 940 mBar, and +/- 62.0% for measurements made below 940mBar. The quoted errors are 2 sigma. A few of the later flights (14-17) were plagued with detector problems which greatly enhanced scatter in the measured data. This scatter should not significantly alter the uncertainty of the data except that a larger number of data points must be averaged to reduce the scatter in this data. The reported error for flights 14-17 is +/- a factor of two for measurements made at pressures above 940 mBar and an additional 20% for meassurements made below 940 mBar. The increase in uncertainty at higher altitudes reflects the increased spread in the observed calibration coefficients and the increased uncertainty in the concentration of H2O at higher altitudes.