Atmospheric Data Error Analysis for the 1994 CTAS 
Descent Advisor Preliminary Field Test

 M. R. Jardin, Steven M. Green
 Ames Research Center


                        Summary

This report describes the analysis of the errors in the 
atmospheric data that were collected during the Preliminary Field 
Test of the Center-TRACON Automation System (CTAS) Descent Advisor. 
The field test was conducted during September, 1994, to determine 
how accurately the DA could predict aircraft trajectories during 
typical descents into Denver's Stapleton Airport. During the field 
test, atmospheric data (temperature and winds) were measured by the 
NASA Transport Systems Research Vehicle (TSRV) Boeing 737. These 
data were to be compared with discrete (3 hr. update rate) real-
time atmospheric prediction (3 hr. forecast) data coming from the 
Mesoscale Analysis and Prediction System (MAPS) created by the NOAA 
Forecast Systems Laboratory (FSL). In addition to the errors in the 
MAPS data, a small but measurable amount of error was contributed 
by the CTAS software processing of the atmospheric data in some 
cases. The purpose of this report is to quantify the total error 
between the TSRV-measured winds and the winds as utilized by CTAS, 
and also to determine how much of the total error came from the 
MAPS model and how much was due to CTAS software processing of the 
MAPS data. Knowing the total error will facilitate analyses of CTAS 
trajectory prediction accuracy as a function of atmospheric data 
errors while analysis of the CTAS contribution to the total error 
will provide a measure of MAPS accuracy during the field test. The 
results of the analysis are that the RMS wind speed difference 
between the TSRV data and the CTAS-processed MAPS data was about 
21 knots of which 18 knots were from the MAPS model, and the 
remaining 3 knots were due to CTAS processing errors. The maximum 
wind speed error measured during the flight tests was about 50 
knots, which occurred at cruise altitude and relatively high wind 
conditions.

In addition to the basic data analyses, a novel approach to 
analyzing the atmospheric data through the use of similarity 
parameters is also presented. The goal of the similarity parameter 
analysis is to determine whether or not it will be possible and 
practical to find a functional relationship among a small set of 
dimensionless groups for describing the accuracy of atmospheric 
models. If a practical set of dimensionless groups can be found and 
an empirical or theoretical functional relationship determined, 
then this information could be used by trajectory-prediction 
software (such as CTAS) to estimate the accuracy of the atmospheric 
model in real-time. The results presented here are of a preliminary 
nature and are not conclusive because of the incomplete and 
relatively small amount of data collected during the field test. 
However, the results do warrant further exploration of the use of 
similarity parameter analysis and application of this analysis to 
the trajectory prediction problem.