Climatological depiction of hurricane structure from passive microwave and scatterometer observations: Using the 12-year JPL Tropical Cyclone Information System (TCIS) to create composites and establish reliable statistics

Currently there are still many unanswered questions about the physical processes that determine hurricane genesis and evolution. Furthermore, a significant amount of work remains to be done in validating and improving hurricane models. While the track forecast has improved significantly over the last couple of decades, the accuracy of the intensity forecast is still insufficient. As highlighted by the NOAA-led multi-agency Hurricane Forecast Improvement Project (HFIP), of particular importance is the need to improve our ability to forecast rapid intensification (RI) and weakening (RW) of tropical cyclones (TCs). Recent studies indicate that the hurricane inner core processes and potential oceanic feedback might play a crucial role in determining the storm's intensity and size. Yet the understanding of these processes is still lacking. This brings to the forefront the need to better understand the role of the inner-core convective organization in TC intensity changes. Recent studies have linked RI to intermittent occurrence of deep, strong convective bursts within the inner core (e.g. Hendricks et al., 2004; Montgomery et al., 2006; Rogers, 2010) occupying as little as 5–10% of the area of the hurricane eyewall. However, an alternative hypothesis is that RI follows abundant and well organized but weaker convection in the inner-core region (Gopalakrishnan et al., 2011). A continuous azimuthally symmetric eye wall (i.e., a ring) of precipitation then indicates the imminent onset of RI (Kieper and Jiang, 2012). This occurs when the ring is at least 90% closed and dominated by shallow warm precipitation extending from near the freezing level to the surface. Our investigation employs a long record of satellite observations to address the questions above. Despite the significant amount of satellite observations today, they are still underutilized in hurricane research and operations, due to complexity and volume. To facilitate hurricane research, we developed the JPL Tropical Cyclone Information System (TCIS) of multi-instrument satellite observations pertaining to: i) the thermodynamic and microphysical structure of the storms; ii) the air-sea interaction processes; iii) the larger-scale environment. One of the two main components of the JPL TCIS is an archival database of satellite observations (TCIS-Database- http://tropicalcyclone.jpl.nasa.gov/hurricane/gemain.jsp). It presents the satellite depiction of hurricanes over the globe during the period 1999-2011, offering both data and imagery. It provides a one-stop place to obtain an extensive set of multi-parameter data from multiple observing systems, making the TCIS-Database a unique source to support hurricane research, forecast improvement and algorithm development. In this study we use multi-channel passive microwave observations from a number of different instruments (TMI, AMSR-E, SSMI and SSMIS, all available at TCIS) to investigate the storm structure. In particular, we compute a Rain Index (Hristova-Veleva et al., 2013) that combines the emission and scattering signals from the multi-channel information to present a cohesive depiction of the rain and the graupel above. We create a number of 2D composites by aggregating Rain Index data from multiple storms as a function of their intensity, their intensification rate and their time-offset from the storms' maximum intensity time. We further investigate the storm asymmetry by computing statistics for the four quadrants with respect to storm motion. Figure 1 illustrates some of our preliminary analysis. The goal of these analyses is to understand how the 2D storm structure evolves and how it relates to the storm intensity and intensity changes. Furthermore, we perform similar analyses, looking this time at the surface wind structure as depicted from QuikSCAT and OSCAT scatterometer observations. In particular, we investigate the wind retrievals produced by our new artificial neural network (ANN) technique (Stiles et al., 2013) that provides reliable estimates of the surface winds under the rain and high-wind conditions typical for hurricanes, thus addressing long-standing issues of rain contamination inside the hurricane core. Figure 2 illustrates some early results. We will present the results of our comprehensive analysis of the 2D rain and wind structure and will relate them to the storm intensity and intensification rates with the goal to understand the underlying processes and to provide clues for model validation and improvement, as well as near-real-time predictive capability. By employing a 12-year long record of global satellite observations we will be able to extract statistically-robust signatures of the importance of the storm organization.