An Investigation of Dust Aerosols and Atmospheric Profiles Associated with North Atlantic Hurricanes Using Multi-Sensor Measurements

MODerate resolution Imaging Spectroradiometer (MODIS) Aerosol Optical Thickness (AOT) and the Atmospheric Infrared Sounder (AIRS) atmospheric profiles (temperature and moisture) Level 2 products are studied to investigate dust aerosol impacts on hurricanes through spatial analysis for four hurricanes Isabel (2003), Frances (2004), Katrina (2005) and Helene (2006). Spatial regions were selected based on two concentric circles (an annulus) and segmented by left and right regions of 180° each around the hurricane's eye. Statistical relationships between MODIS AOT, AIRS temperature and percent relative humidity (RH) among these four hurricanes is presented. The variation of temperature and RH values represented in static-left, static-right, dynamic-left and dynamic-right for the 500 mb, 700 mb and 850 mb pressure levels were analyzed to determine the impacts of dust aerosols on temperature and RH. For the four hurricanes' average, AOT was found highest in Helene as opposed to (0.27) Isabel (0.23), Frances (0.18) and Katrina (0.16). Dust aerosols showed negative impacts on the atmospheric moisture levels and positive on the temperature.;Additionally, a systematic multivariate analysis of MODIS aerosol retrievals over the North Atlantic spanning 7 hurricane seasons combined with Statistical Hurricane Intensity Prediction Scheme (SHIPS) parameters is presented. My focus is on investigating the effects of 5 categories that include aerosol, wind, relative humidity, shear, and temperature on hurricane intensities. The primary goal is to be able to explain the same physical phenomena equally well by significantly reducing the number of parameters employed in the statistical analysis. Five categories which include, wind, aerosol, shear, relative humidity, and temperature components were established by reducing 56 variables to 20. Aerosol, wind, humidity, shear, and temperature were all contributing factors in the regression equation with the ranking for the contribution found as (1) wind, (2) aerosol, (3) shear, (4) relative humidity, and (5) temperature components. A very important discovery that the rank of Aerosol has been preceded the humidity in three cases with higher Adjusted R 2. This demonstrated the fact that by removing redundant variables from the predictor variable set improves the performance of the models by refining the effect of the curse of dimensionality as well as enhancing the interpretability of the model.