Retrieval Coefficient Assessment for TexAQS 2000

MJ Mahoney, July 24, 2002

Background

The TexAQS 2000 campaign was the first time that a Microwave Temperature Profiler (MTP) flew extensively in the planetary boundary layer (PBL). Because of the short time available to prepare for this field deployment, the existing DC-8 MTP was flow "as is." There wasn't time to consider what the optimum instrument configuration might be, such a frequency of operation, scan angles, etc. Since the DC-8 normally flies at an altitude of 10-12 km, much more optically-thick oxygen lines are observed than might be optimal at the typical TexAQS flight altitude of 0.61 km (2000 feet). As it turns out, this is both good and bad.

A particular concern before performing the MTP retrievals for TexAQS 2000 was the impact of ground emission on the observed brightness temperatures. Initially, it was hoped that infrared surface temperatures measured on the NCAR Electra could be used as an additional "observable" in the MTP retrievals to constrain the downward-looking observables. This might have been possible over water, but over land the emissivity is simply too variable. In addition, there is no obvious relationship between the infrared and microwave emissivities, so it is not clear where this approach might have lead. (see Investigation of Heimann Probe Data )

Given the limited resources for this work, the decision was made to weight-down the downward-looking MTP brightness temperature measurements so that they would not impact the retrievals. This was done by requiring five optical-depths (one optical depth is the e-folding distance) between the aircraft and the ground in the viewing direction at the most optically-thin measurement frequency (55.51 GHz). Using this criterium, it was found that for flight at 2000 feet, the rms difference between 206 coastal RAOBs and simulated retrievals was <0.4 K at a distance of 1000 feet (0.305 km) below the aircraft, and this was deemed acceptable. (The rms difference between RAOBs and the simulated retrieved temperature at flight level was 0.25 K.)

Because it was (erroneously) believed that the NCAR Electra never flew when ground inversions existed, the retrieval accuracy assessment did not include any RAOBs with ground inversions. This was an oversight, because ground inversions did exist over the downtown Houston RAOB launch site on August 30, 2000, and was apparent in RAOBs launched at 1400 (50.4 ks) and 1700 (61.2 ks) UT on August 30, 2000. The NCAR Electra took off at ~55 ks on August 30 and flew over downtown Houston at 56.7 ks and 58.4 ks. The first overpass was at 11,000 feet and the second at 2000 feet. Examination of the MTP data shows no evidence for the inversion when at 11,000 feet because it does not have the vertical resolution necessary to do so, but the measurements clearly show the inversion when flying at 2000 feet.

The August 30, 2000, Ozone Event

On August 30, 2000, during the TexAQS 2000 campaign, atmospheric conditions were such that the larger-scale geostrophic flow was offshore and opposed the local sea breeze near Houston, Texas. This situation is especially conducive to pollution events, because air originating from the emission sources near Houston become stagnant over Galveston Bay, leading to very high ozone levels when the sea breeze front finally penetrates inland. A question which we wish to pursue is whether or not the Jet Propulsion Laboratory (JPL) Microwave Temperature Profiler (MTP) data taken aboard the NCAR Electra can be used by Texas A&M University (TAMU) to validate and improve the MM5 model simulations of boundary layer depth, boundary layer variability, and sea breeze structure on August 30 and other days during the TexAQS 2000 field program.

While the MTP measurements clearly map the horizontal temperature gradients over the Houston area, this is not the most important driver for the MM5 model’s ability to forecast a pollution episode. What matters most is the surface forcing as characterized by the temperature field below the NCAR Electra’s flight altitude. Since the NCAR Electra nominally flew at 2000 feet, or near the top of the PBL, relaxing the weight given to the downward-looking MTP measurements could have a negative impact on the usefulness of the MTP data for the MM5 modeling activity.

For the purpose of evaluating the effectiveness of the MTP retrievals to capture the surface forcing, seven Airsonde radiosondes from downtown Houston (HOU) and LaMarque (HSE) were chosen for case studies: from HSE – 2002.08.28 2033UT and 2002.08.31 2300UT, and from HOU - 2002.08.29 2301UT, and 2002.08.30 1400UT, 1700UT, 2000UT, 2259UT. Simulated MTP retrievals were performed to understand the limitations of the downward looking retrievals, and hence the ability of the MTP to characterize the surface forcing.

The figure above shows the accuracy of the simulated retrievals for four differently weighted sets of rectrieval coefficients (RCs) as a function of the distance below the aircraft. Only the five radiosondes without inversions were used; all the radiosondes were used in the next figure. Since the radionsondes used to calculate the retrieval coefficients did not include any inversions, all of the RC sets did quite well, with the most heavily weight RCs being best. The average error has no bias, and its rms error is <0.25 K within 1000 feet of the aircraft.

The figure above shows the results when test radiosondes with inversions are included. As expected, the retrievals deteriorate significantly. However, since the raw MTP measurements clearly show brightness temperatures expected for inversions, it is expected that with proper retrieval coefficients that these errors will drop significantly.

MTP Home Page: http://mtp.jpl.nasa.gov