A Data Mining Algorithm for Mesoscale Convective Vortices in the Pregenesis Environment for Tropical Cyclone Genesis Probability Prediction
R. MacCracken, P. Houser, K. Mohr 2013-2017: A Data Mining Algorithm for Mesoscale Convective Vortices in the Pregenesis Environment for Tropical Cyclone Genesis Probability Prediction. Office of Naval Research:Long Range BAA for Navy and Marine Corps Science and Technology, $559,005, 1 mo/yr (GMU), 4/01/2013 – 03/31/2015, Sponsor POC: Dr. Ronald Ferek (ron.ferek@navy.mil).
Summary: By using innovative data mining techniques and recently emerging weather reanalysis products, this project will identify, analyze, and develop a predictive capability for the dynamics, structure and evolution mesoscale convective vortices (MCV) originating in West Africa. Naval operations rely heavily on the accuracy of tropical cyclone prediction, since tropical storms, hurricanes and tropical cyclones can affect both land- and sea-based naval operations and facilities. This proposed research project has three specific goals:
(1) To create a statistical model using data mining techniques to predict the probability of the occurrence of tropical cyclone genesis (i.e., tropical cyclogenesis) from the pregenesis environment. Studies have shown that certain West African mesoscale convective vortices (MCV) possess and retain a dynamical structure in the pregenesis environment in West Africa that increases the probability that tropical cyclogenesis could occur downstream. With 27% of the global tropical cyclones forming from the interaction between African easterly waves (AEW; Chu 1999), mesoscale convective systems (MCS), and MCVs which originate in West Africa, this study will focus on West Africa as the pregenesis region. Data mining techniques will uncover relationships within the data that may not have previously been uncovered from cyclogenesis case studies, as well as generating probabilistic predictive models based on these relationships.
(2) To perform a climatological study which categorizes the origin, structure, evolution and geographic distribution of the MCV in West Africa and the Atlantic and Pacific basins. Tracking and analyzing a long-term data record of instances of MCVs will reveal far more information about the MCV than a case study, or short term climatological study would find. Additionally, this climatological database will be necessary for further data mining applications.
(3) Address science questions, such as mechanisms leading to MCV formation, regeneration of secondary convection and the interactions between the MCV and the larger scale, leading to intensification of the AEW, and tropical cyclogenesis in both the Atlantic and Pacific Oceans. These science questions have been raised by multiple studies without finding definitive answers. Providing more insight into these questions will greatly aid in further the understanding of this atmospheric phenomena, which in turn will lead to better representation in numerical weather prediction models.