GPS/INS navigation precision and its effect on airborne radio occultation retrieval

In February 2008 the GNSS Instrument System for Multistatic and Occultation Sensing (GISMOS), developed at Purdue University, was deployed on a Gulfstream V aircraft to make atmospheric observations in the Gulf of Mexico coastal region. The primary objective of the flight campaign was to test the performance of the profiling system in retrieving atmospheric refractivity, which is related directly to temperature and moisture content. 20 dropsonde profiles and 28 extra radiosonde profiles were collected for comparison with measurements from the GISMOS airborne observing system. The largest instrumental errors in the airborne observing system are associated with velocity errors in the navigation system. We investigate the sensitivity of the refractivity retrievals to navigation accuracy. The state-of-the-art Applanix POS/AV 510 Global Positioning System (GPS)/Inertial Navigation System (INS) uses carrier phase differential measurements integrated with an Inertial Measurement Unit (IMU) to produce a highly accurate high-rate navigation solution. The specified accuracy for this navigation system is 5mm/s. In previous simulations of radio occultation retrievals, this level of accuracy has been demonstrated to result in less than 0.5% refractivity error from the surface to about 9 km for an airplane flying at 10 km altitude [Xie et al., 2008]. 0.5% refractivity corresponds to approximately 1K temperature error or 5% humidity error at 2km altitude, which is a minimum requirement for upper air observing systems for meteorology. We use this dataset to confirm that the navigation system is consistent with the specifications, and to demonstrate some practical operational considerations for conducting radio occultation observations. The degradation of the retrieval quality without the IMU data and with a lower accuracy IMU is examined, in order to determine whether less accurate, but still useful data can be retrieved at significantly reduced cost. Without IMU data the east and north components of the velocity errors are estimated to be less than 2 cm/s. It is found that a large bias exists in the retrieved refractivity, which prevents such a system from providing reliable observations. On the other hand, a less accurate IMU, in the absence of large biases, would still provide acceptable accuracy measurements from the surface to 3.5 km altitude. The accuracy of the real-time autonomous GPS/INS navigation solution is investigated to evaluate the potential for on-board real-time processing of radio occultation profiles. The possible use of this airborne observing system for studying the diurnal cycle of moisture is addressed. Observations that accurately characterize this daily cycle can be critical for improving and tuning climate model representations of clouds and convection, especially over oceans. It is found that typical diurnal variations in moisture from the surface to 7.5 km altitude are much larger than the limiting errors in the radio occultation retrievals. Therefore we anticipate that the new airborne RO technique could be used to tackle this problem effectively