VERIFYING AN AUTOMATED DECISION ALGORITHM USED FOR FLIGHT DECISIONS IN THE SPARTICUS CAMPAIGN
Open Access
Author:
Hanlon, Christopher James
Area of Honors:
Meteorology
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Johannes Verlinde, Thesis Supervisor Johannes Verlinde, Thesis Supervisor Paul Markowski, Thesis Honors Advisor
Keywords:
automated decision algorithm
Abstract:
The quality of flight decisions produced by an automated decision algorithm for the Atmospheric Radiation Measurement (ARM) Small Particles in Cirrus (SPARTICUS) campaign was verified and compared to the quality of flight decisions made by human forecasters on the SPARTICUS team. The SPARTICUS campaign collected data from cirrus clouds over the ARM Southern Great Plains (SGP) site between January 2010 and June 2010. Self-organizing maps, dynamic programming, and relative humidity forecasts from the Global Forecast System (GFS) were used to generate day-ahead yes/no decisions for research flights over the SGP site throughout the 139-day field campaign. SPARTICUS researchers collected 10 “good” days of data, while the algorithm would have collected 12 “good” days, a 20% improvement. This increase in data yield is similar to the improvement shown by an analogous model used on the 2009 ARM RACORO campaign (Small et al., 2011). In addition to increasing the amount of data collected, the algorithm reduces the amount of time and energy spent by researchers on forecasting and decision-making.
