Profiling the Moisture Environment of Developing Tropical Storms using Airborne Radio Occultation
An extensive airborne radio occultation (ARO) data set has been collected by the GNSS instrument system for multistatic and occultation sensing (GISMOS) during the PRE–Depression Investigation of Cloud systems (PREDICT) field experiment in 2010. ARO is a promising new technique in which GPS radio occultation measurements are made with an airborne receiver. ARO has the potential to provide spatially and temporally dense data sets over a mesoscale region to complement other observing systems such as dropsondes or meteorological satellites and to improve forecasts after assimilation of the ARO data into numerical weather prediction (NWP) models.^ The PREDICT campaign was the first full scale deployment of GISMOS and a primary goal was to assess the accuracy of the ARO refractivity retrievals near developing tropical storms. Twenty–six research missions were flown which sampled eight storm systems. GISMOS collected data using both geodetic quality GPS receivers, to produce preliminary profiles of the upper troposphere, and a 10 MHz GPS recording system (GRS) used to sample and record the raw GPS signals throughout the lower troposphere. Radio (RF) signals recorded by the GRS are analyzed in post processing mode using a software receiver with an open loop (OL) technique. Both the refractivity results from the geodetic receiver data and GRS data are compared to refractivity profiles calculated from data recorded by nearby radiosondes and dropsondes as well as profiles calculated from numerical weather prediction model output. All ARO retrievals were based on geometric ray optics.^ Twenty-one refractivity profiles of the upper troposphere were obtained from setting occultations measured by the geodetic receivers. The geodetic receiver retrievals agreed within 2% of co-located dropsonde, radiosonde and model output, but on average did not extend below 6 km altitude and only one or two retrievals were obtained per flight. The conventional phase locked loop (PLL) tracking used by the geodetic receivers could not maintain signal lock in the lower troposphere where signal propagation through more complex moisture structure produces rapid fluctuations in amplitude and phase.^ In the middle to lower troposphere, the open loop tracking technique used with the recorded raw GPS signals is more robust for extracting the carrier phase and amplitude measurements. A set of 46 refractivity profiles was obtained from the four research flights that sampled the pre-hurricane Karl system. Typically, 10 to 15 retrievals per flight were possible using open loop tracking, a great improvement over the conventional receivers. Phase measurements were possible with open loop tracking to lower altitudes, 2 km on average. The mean error of the open loop retrievals compared to dropsondes and model output was also about 2% in the 6 – 12 km altitude range. A bias was discovered between rising and setting occultations. Also, a negative bias was found that becomes significant at lower altitudes, likely due to multipath propagation. The full data set was analyzed using geometric optics retrieval methods, in order to examine the dependence of the bias on atmospheric characteristics. Radio holographic techniques based on physical optics are being developed for ARO to retrieve refractivity in the presence of multipath. A test case comparing radio holographic retrieval results to a geometric optics retrieval shows the potential for significantly reducing this type of bias. Further analysis of the extensive geometric optics dataset is carried out, within the context of these biases below 6 km. Several examples of the type of analysis now possible with ARO are given, including radial profiles of thermodynamic variables.^ In the middle to lower troposphere, variations in refractivity are shown to be indicative of moisture variations. ARO refractivity profiles sampling different areas within the tropical wave showed characteristics that were consistent with horizontal moisture gradients present in the NWP model representation of the developing tropical storms. The use of the dense ARO data set for model validation is tested in comparisons with a high resolution Weather and Research Forecasting (WRF) model simulation and the European Center for Medium Range Weather Forecasting Interim Reanalysis (ERAI). Variation in refractivity of ARO profiles preceding the development of the pre-Karl system is consistent with increasing moisture near the storm center. These promising results demonstrate that the ARO refractivity retrievals should be an attractive option for assimilation into NWP models. The ARO data set produced in this thesis have been provided to project collaborators for future work assimilating them into WRF simulations of Hurricane Karl to assess improvements in forecast intensity.^
Jennifer S. Haase, Purdue University, Wen-Yih Sun, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our