Comparison of Potential Temperature Gradient Estimates from Various Temperature Profile Data Sources

From July through September 2015 near Los Angeles International Airport (LAX), measurements of temperature profiles from two passive radiometers and a RADAR Radio Acoustic Sounding System (RADAR-RASS) were conducted. All three instruments were located near the ocean just to the west of the airport.  In addition, temperature profiles from NOAA’s Rapid Refresh (RAP) hourly-updated assimilation/modeling system and the WMO/NOAA Aircraft Meteorological Data Relay (AMDAR) data base were utilized.  For all temperature profiles a standard algorithm was used to convert from temperature profiles to dry potential temperature (PT) profiles. Data acquired at times when there was precipitation or fog were discarded. Humidity effects were not considered.

The goal of this effort was to use intercomparisons of the data obtained from the various ground-based remote temperature profiling instruments and the RAP data to characterize the variability of derived potential temperature gradient (PTG) values at out-of-ground-effect altitudes where the effect of trailing vortices on separations between landing aircraft would need to be considered. This information is important because PTG (along with atmospheric turbulence, wind and various aircraft parameters) is identified in the open literature as a key parameter in determining the lifetime and descent of these vortices which may adversely affect trailing aircraft if not properly accounted for.

The altitude of landing aircraft in the final approach phase is approximately 1000 ft (305m) and trailing vortices generated at this altitude often descend on the order of 350ft (107m). Thus these vortices may inhabit the altitude range of 650-1000ft (198-305m) where they could impact trailing aircraft. For this reason, the current effort is focused on PTG at an altitude in the middle of this range which is about 250m. An estimate of this quantity may be obtained from the difference between PT measured at approximately 200m and 300m.

The observed scatter of the derived PTG data was employed as a practical way of quantifying the measurability and repeatability of the values being compared. This information is important for determining the variability and practical measurability of PTG, the stratification parameter that supports aircraft wake vortex data analyses, as well as for providing bounds on the variability of the environmental data used for wake modeling. This effort is an integral part of the Federal Aviation Administration’s ongoing effort to examine and prototype additional weather based dynamic wake turbulence separation concepts.

The radiometers were a Model  MP-3000A from Radiometrics Corporation and a Model MTP-5 PE supplied by Kipp & Zonen USA, Inc. The MP-3000A uses 12 channels in the ranges 22-30 GHz and 51-59 GHz and acquires data along three different sight lines: vertical, 15deg N of vertical and 15deg S of vertical. For this study, the data from the 15deg north-of-vertical sky view was used because the portion of the sky seen in this way was close to the portion scanned by the MTP-5 PE. The vertical resolution of the MP-3000A profiles is 50m between altitudes of 0m and 500m and 100m between 500m and 1000m. Profiles were produced every ten minutes, but only hourly profiles were used in the comparisons for consistency with the other data sources.

The MTP-5 PE uses a single channel operating at a frequency of 60 GHz (5mm wavelength) scanning from 0deg to 90deg towards the northwest to produce vertical temperature profiles by measuring atmospheric thermal radiation in the center of the oxygen absorption band. An inverse method is employed to convert raw brightness temperature data to a vertical profile of temperature. The profile resolution is 10m from altitudes of 0m to 100m, 25m from 100m to 200m, and 50m from 200m to 1000m. Profiles were produced every five minutes, but as with the MP-3000A, only hourly profiles were used in the comparisons for consistency with the other data sources.

Note that the MP-3000A and MTP-5 PE temperature profiles are inherently smoothed as a result of the inversion process that produces the profiles from the raw radiometer data.

The RADAR-RASS device, Model LAP3000 from Scintec, operating at 915 MHz with a narrowband acoustic source frequency close to 2KHz, produced virtual temperature profiles with a resolution of about 62m between altitudes of 174m and 985m. The RASS data were smoothed with a 100m boxcar filter to eliminate higher wavenumber noise. Ten minute averaged profiles were produced once per hour. Since the effort focuses on the vertical gradients of PT, the virtual temperatures were treated as if they were non-virtual temperatures. Resulting errors in the gradients were estimated to be on the order of 1-2%.

Temperature profiles from RAP data were used as if they were measured by an actual sensor. In fact, RAP profiles were shown to agree reasonably well with profiles from AMDAR data which are measured by actual temperature sensors aboard departing and arriving aircraft. This agreement was visually demonstrated for morning and afternoon on two days in each of the three months of the deployment. As a result, it is felt that hourly RAP profiles can be used as a reasonable surrogate for AMDAR data which would be impractical to utilize on an hourly basis.

Comparisons of the various data sources were focused on (i) linear fits to scatter plots of the derived PTG data from pairs of sensors and (ii) evaluation of RMS differences between data from pairs of sensors. It was determined that potential temperature gradients at 250m altitude could be estimated to within 0.5 to 1.0 degC/100m. Fast-time model runs for potential temperature gradients of 0, 1, 2 and 3 degC/100m will be shown to illustrate the effect of a gradient difference of 1degC/100m.