ISAMS Version 8 Product Quality Summary In the following, `precision' is taken to be the random error resulting from the ISAMS error analysis, and `accuracy' is taken to be the root-sum-square of the random and systematic errors. The random error includes both radiance measurement error (detector noise) and retrieval smoothing error, which properly should be included in the accuracy rather than the precision. For full details see the ISAMS Version 8 validation report. The column labeled `bias' has been derived mainly on the basis of comarison with preliminary version 9 retrievals. The data is initially retrieved at level 2 on a grid with vertical spacing of about 2.5 km. This number in the `resolution' column indicates that the originally retrieved data are independent of each other. However the data supplied at level 3 is interpolated onto the UARS pressure grid, degrades the resolution in an unpredictable way. * To be supplied Temperature ----------- `Altitude' bias accuracy precision resolution mb K K K km 100 -1 2 2 2.5 10 -1 3 2 2.5 1 -2 4 3 2.5 0.1 +1 12 12 7.5 0.01 +3 17 17 10.0 The stratospheric values (100-1mb) are based on comparisons with NMC data (fig 2.6 of the ISAMS V0008 validation report), the mesospheric values (0.1-0.01mb) are based on the error analysis (table 2.4). Caveats The V0008 retrievals are only useful up to 0.1mb - at higher altitudes the profiles are largely climatology. The error bars ('Quality' values) associated with the ISAMS profiles are overestimates of the random error in the stratosphere, mesospheric value are probably realistic. A warm bias of 2-3 K is observed at the peak of the Pinatubo aerosol distribution (20mb at low latitudes). The temperatures in the presence of PSCs are probably unreliable since these are treated as having the same spectral characteristics as volcanic aerosol. On some days thermal-snap effects are shown by spikes in the T(p) contours at the same two latitudes each orbit. Aerosol ------- `Altitude' bias accuracy precision resolution mb % % % km 68 3 26 23 2.5 46 -1 23 15 2.5 32 -1 21 18 2.5 21 -4 37 35 2.5 15 -4 89 88 2.5 Caveats There is an anomaly in the radiance in the height range 19-25km due to a problem with the field of view correction algorithm. The symptom is that the retrieved aerosol in this range is higher before 11 April 92 than after that date by 20-30%. It is likely that the earlier data is overestimated. This problem is not included in the table above, and will be corrected in version 10. Ozone ----- `Altitude' bias accuracy precision resolution mb % % % km 100 -15 41 27 30 -15 12 10 10 -15 13 10 3 0 17 13 1 +5 18 13 0.3 0 19 16 0.2 0 21 19 Caveats Even though ISAMS ozone has been retrieved between 100 and 0.2 mb in version 8, the data outside the range 10 to 1 mb is suspect and should be used only with caution, Below 10 mb residual aerosol effects, acting indirectly via the temperature retrieval cause perturbations in the Pinatubo layer. Above 1 mb, daytime data are influenced by nonLTE processes, and night time data are biassed by the use of a daytime a priori profile, thus the day/night variation will be poorly represented. Water ----- `Altitude' bias accuracy precision resolution mb % % % km 10 -22 24 23 7.0 3 -8 19 18 6.0 1 +5 22 21 4.5 0.3 +7 52 52 8.0 0.1 +1 70 70 10.0 Caveats The accuracy and precision at 0.3 and 0.1 mb are probably overestimates. The data is retrieved for night time only, due to nonLTE effects in the daytime which have not yet been accounted for. Aerosol contamination leads to overestimated H2O at 10mb near the equator. Due to a change of PMC pressure on day 33 V0008 H2O is 5-10% too low for days 17-33. Methane ------- `Altitude' bias accuracy precision resolution mb % % % km 10 -15 20 19 3 +4 20 19 1 +8 15 14 0.3 +4 36 36 Caveats Nitrous Oxide ------------- `Altitude' bias accuracy precision resolution mb % % % km 10 -21 22 22 3 -17 23 22 1 -24 22 22 Caveats The version 8 nitrous oxide values are clearly too high by a factor of about two. The error analysis does not reflect this. The reason has not yet been finally identified, but recent indications are that the telemetry may be indicating the amount of gas in the cell incorrectly. Nitrogen Dioxide ---------------- NIGHTTIME `Altitude' bias accuracy precision resolution mb % % % km 100 30 10 3 1 0.3 DAYTIME `Altitude' bias accuracy precision resolution mb % % % km 100 30 10 3 1 0.3 Caveats Nitrogen Pentoxide ------------------ The error analysis indicates that N2O5 errors are greater than 100% at all levels. This data should not be used except with great caution, and consultation with the ISAMS team. Carbon Monoxide -------------- NIGHTTIME `Altitude' bias accuracy precision resolution mb % % % km 10 -16 23 19 3 -12 35 31 1 +1 61 59 0.3 -2 31 31 0.1 -1 37 37 0.03 0 40 40 DAYTIME `Altitude' bias accuracy precision resolution mb % % % km 10 - 28 24 3 - 31 27 1 - 32 28 0.3 - 24 23 0.1 - 32 32 0.03 - 26 22 Caveats Nitric Oxide ------------ The nitric oxide channel has proved difficult to calibrate due to emission from thermospheric NO in the region of the limb scan used for a space view in the other channels. A calibration scheme had not been developed in time for the version 8 processing, but will be implemented in version 10. It is not expected that validated NO will be available until version 11. Nitric Acid ----------- The nitric acid channel is heavily contaminated with aerosol emission To date, the retrieval in the presence of this contamination has proved difficult to the accuracy required for public release of the data. Consequently nitric acid is not available in version 8 or 9. It is expected that it will become available in a future version of the data processing.