| MTP Home Page |

Notes on CRYSTAL-FACE ER-2 and WB-57 MTP Retrievals

MJ Mahoney
Last Updated: 2003.02.18

These notes discuss the calibration of the MTP temperature profiles aboard the ER-2 and WB-57 aircraft for the CRYSTAL-FACE campaign. They also discuss the nature of the MTP tropopause solutions, and the effectiveness of the retrieval coefficients in retrieving temperature profiles.

Contents:

Temperature Calibration

MTP temperatures at flight level for the CRYSTAL-FACE campaign were calibrated against radiosondes launched near the ER-2 and WB-57 flight tracks. Based on these comparisons, corrections were applied to the Mission Meteorological System (MMS) temperatures on the ER-2 and the PTW temperatures on theWB-57. These corrected flight level temperatures were then used as the reference temperature on the ER-2 and WB-57, respectively, for calibration of the MTP retrieved temperature profiles. Based on 25 radiosonde comparisons, we found that on the ER-2 the outside air temperature (OATer ) from the preliminary MM files was given by:

OATer = OATmms + 0.30 +/- 0.21 K (preliminary MM files)

That is, the MMS preliminary temperatures were too cold by 0.30 K. However, on 2003.07.15 final MM files were submitted, and examination of two flights showed that the final MMS temperatures were increased by 0.36 K, making MMS, MTP and radiosondes all agree for the final MM files!

On the WB-57 the outside air temperature (OATwb) based on 27 radionsonde comparisons was given by:

OATwb =  OATptw - 0.32 +/- 0.22 K, (final PT files)
OATwb = OATmms -0.51 +/- 0.22 K (preliminary MM files).

where OATmms and OATptw are the MMS and PTW outside air temperatures (OATs), respectively. This is consistent the average difference for OATmms - OATptw = -0.14 to -0.20 K that I find by directly comparing the final MM and PT files. However, the peak-to-peak difference between MMS and PTW on the WB-57 sometimes exceeds 0.7 K and is clearly pressure altitude dependent. The final MMS and PTW pressure altitudes do not agree, and I am trying to get to the bottom of this. Initially, since the PTW pressure altitudes were lower, I suspected a leak in the static air line (which would increase the measured pressure and thus reduce the altitude), but that does not appear to be the case.

As has been the case for several years now, the retrieval coefficients used for the CRYSTAL-FACE (C-F) MTP temperature profile retrievals were calculated using radiosondes launched during the field campaign. For C-F these included radiosondes primarily from southern Florida launch sites: Key West (EYW), Tampa Bay (TBW), Miami (MFL), Andros Island (MYNN) and Cape Kennedy (XMR). In fact extra sondes were generally launched from the first three sites (EYW, MFL and TBW) on flight days. The use of these sondes to calculate retrieval coefficients (RCs) ensures that  the RCs represent the actual conditions during the campaign rather than a climatology for the campaign location based on previous years' radiosondes.

Radiosonde Tropopauses

Examination of the radiosondes launched during the C-F campaigh showed that while many had sharp tropopause inversions typical of the deep tropics, many others had "dull" tropopauses. That is to say, there is a region which is 1 km or so thick where the temperature is nearly isothermal, or increases slightly with altitude. This is

SD 1
(1) Key West (EYW) 20020729 0000UT and 20020729 1200UT
 SD 2
(2)  Miami (MFL) 20020710 1200UT and 20020210 1800 UT
SD 3
(3) Tampa Bay (TBW) 20020728 2100 and 20020729 0000UT
 SD 4
(4) Miami (MFL) 20020718 1800UT and 20020719 0000UT

illustrated in the above figure for radiosondes launched from Key West, Tampa Bay and Miami, with the time between each launch varying from 3-12 hours. For profiles like these, the retrieved MTP tropopause altitude will be at the  top or bottom of this "tropopause layer" depending on how far above or below the tropopause the  aircraft actually is. This occurs because the MTP vertical resolution degrades from 100 meters near flight level in proportion to the distance.

ER-2 and WB-57 Retrievals

Because of the presence of sharp and dull tropopauses during the CRYSTAL-FACE campaign, it was decided to select radiosondes for retrieval coefficient (RC) calculations based on whether the radiosondes had sharp or dull tropopauses.  Several sets of retrieval coefficients were calculated for these two circumstances using radiosondes collected during the campaign. When retrievals were performed using these RCs, it was found that a few of the ER-2 flights and most of the WB-57 flights had tropopauses which moved up and down by ~600 meters, depending on the proximity of the aircraft to the tropopause. The reason for this is that the MTP retrievals use an information content metric to decide which set of retrieval coefficients best matches the actual MTP brightness temperature measurements. In addition, if the lapse rate of this tropopause layer is close to -2 K/km, radiometric noise will also cause the tropopause solution to move up and down a little. Examples of this behaviour are given below.

ER2 CTC
(a) ER2 CTC for 2002.07.23 (click to enlarge)
 WB57 CTC
(b) WB-57 CTC for 2002.07.21 (click to enlarge)

On the left is the color-coded temperature curtain (CTC) for the ER-2 flight of July 23, 2002. Early in this flight at 65 ks UT, retrieval coefficients based on a dull tropopause were found to best match the MTP measurements. As a result a lower tropopause was found and the tropopause temperature was warmer than, say, after 71 ks when retrieval coefficients based on sharp (and colder) tropopauses were found to better match the MTP measurements. Two important points also need to be made: first, during most of this flight, the ER-2 was more than 6 km above the tropopause. At this distance, the MTP cannot produce an accurate tropopause measurement. The retrieved tropopause altitude will simply reflect the average tropopause altitude of the radiosondes used to calculate a particular set of retrieval coefficients. However, since the radiosondes used to calculate the RCs were from the deployment period, the tropopause solutions should be "reasonable." The second point is that we could have easily avoided this "jumping" tropopause situation by allowing retrievals with only the sharp or dull retrieval coefficients. While this would produce a much better behaved tropopause solution, it would not be "fair" to measurements since, as shown above, both sharp and dull tropopauses can be present on short time periods (or similarily, over short distances, although this case can't be made using radiosonde measurements.)

On the right (above) is the CTC for the WB-57 flight of July 21, 2002. Since the WB-57 generally flew above, near, and below the tropopause during CRYSTAL-FACE, the situation is very different than for the ER-2. Becauses the WB-57 flew close to the tropopause, it should be able to make accurate tropopause height measurements, but whether sharp or dull RCs are used for the retrievals will depend sensitively on just how close the WB-57 was to the  tropopause. The MTP will certainly be able to tell whether the tropopause -- sharp or dull -- is above or below the WB-57, but just where the measured tropopause lies will depend not just on the proximity of the WB-57, but also on the actual tropopause shape. As for the ER-2, a "better-behaved" tropopause solution could be found by using only sharp or dull RCs, but there is no justification for doing this. In fact, it is certainly reasonable to expect that very active thunderheads might change the ambient tropopause height and the temperature there. We prefer to let the measurements speak for themselves rather than make unjustified choices. As was the case for the ER-2, the  metric that evaluates the quality of the temperature profile retrievals was excellent.