Statistical Analysis of Meteorological Observations from NOAA's Hurricane Hunter Aircraft for the Development of Quality Control Algorithms

Tropical cyclones typically develop and evolve in oceanic data sparse regions, thereby making them challenging meteorological phenomena to predict, analyze, and study. Forecasters and researchers rely on in situ data collected from the National Oceanic and Atmospheric Administration's (NOAA) Hurricane Hunter aircraft, two WP-3D Orions, operated out of MacDill AFB, FL by the National Marine and Aviation Operation's Aircraft Operation Center (AOC). These aircraft provide vertical atmospheric profile data from GPS dropsondes, precipitation profiles from lower fuselage and tail radars, specialized probes that collect and display cloud and precipitation particles, a Step Frequency Microwave Radiometer (SFMR) instrument to estimate surface oceanic wind speed, and flight level data from instrumentation mounted to the airframe. The flight level data are averaged into 30-second observations known as High Density Observations (HDOBs). It is crucial that HDOBs accurately represent in-flight environmental conditions because hurricane forecasters utilize these observations to analyze the structure, movement and evolution of tropical cyclones. Presently, there is no automated methodology to quality control HDOBs in real-time before they are transmitted from the aircraft.

AOC conducted statistical analysis as the starting point for development of automated quality control routines. The initial data set used in the statistical analysis was all HDOBs from tasked tropical research or reconnaissance missions for NOAA WP-3D flights into tropical waves up to Category 5 hurricanes from 2005 through July 2010. Analysis was performed on a representative number of flight-level parameters across the full range of tropical systems treated as a single universal data set. The initial results confirmed that statistically flagging outlying parameter values plus or minus three standard deviations beyond a normal statistical density distribution may be feasible and useful. The data were then stratified with respect to the strength of tropical system (tropical wave, tropical storm, Category 1 hurricane, etc.) for further analysis.

The operational objective of this effort is to develop and implement quality control algorithms that flag anomalous data before the HDOBs are transmitted from the aircraft. Flagging, rather than removal of outlier data, allows the Flight Meteorologist the discretion to identify instances where there is a valid meteorological reason for extreme values, e. g., very warm 700 millibar flight level ambient temperatures in the eye of an intense hurricane, and allow them to be included in the HDOB. While all AOC flight level data are carefully examined in post-flight analysis before finalized data sets are forwarded to researchers, this real-time quality control methodology will improve HDOB data validity for hurricane forecasters and researchers utilizing these observations in an operational mode.