Storm surge prediction with cygnss winds

The NASA Earth Venture Cyclone Global Navigation Satellite System (CYGNSS) is a constellation of eight observatories in a 35° inclination, ~530 km altitude Earth orbit. Each observatory carries a 4-channel bistatic wind scatterometer receiver. Measurements of the ocean surface scattering cross section are converted to 10 meter-referenced wind speed. The mission improves the temporal sampling of winds in tropical cyclones (TCs) with a revisit time of 2.8 hours (median) and 7.2 hours (mean) at all locations between 38 deg North and 38 deg South latitude. Operation at the 1575 MHz GPS L1 frequency permits wind measurements in the TC inner core that are often obscured from other spaceborne remote sensing instruments by intense precipitation in the eye wall and inner rain bands. The potential for improved storm surge forecast skill is examined using simulated CYGNSS science data products for Hurricane Irene. We present and compare ADCIRC 2DDI storm surge hindcasting results of Hurricane Irene using four meteorological forcing scenarios: 1) “True” meteorological data obtained from HWRF reanalysis runs; 2) “Worst-case forecast” using low-resolution NOGAPS forecast wind and pressures; 3) “Best-case forecast” using high-resolution HWRF forecast winds and pressures; and 4) a simulated “CYGNSS forecast” with wind field given by a parameterized model trained using CYGNSS-derived values for the maximum wind speed and radius of maximum winds. The results suggest that the improved temporal resolution of the CYGNSS-derived winds has a positive impact on storm surge modeling predictions.

1. Introduction

Hurricane storm surge is a problem of growing importance due to increasing population density along the coasts, rising sea levels, and the possibility of greater frequency and severity of storm occurrence [1]–[3]. Storm surge modeling can provide guidance for warning and evacuation systems. However, modeling accuracy, particularly with respect to the intensity forecast (as opposed to storm track) is limited by a lack of high resolution meteorological data. Currently, methods of data collection within a tropical cyclone's eyewall are limited to ‘hurricane hunter’ type aircraft and dropsondes. Neither of these methods can provide the needed temporal and spatial data sampling necessary for accurate storm surge forecasting. The objective of this work is to examine and assess the potential for CYGNSS data products to assist with storm surge forecasts. To this end, we utilize simulated CYGNSS data for Hurricane Irene to hindcast storm surge and compare the results to alternate cases that use currently available meteorological data products; specifically, HWRF forecast winds, with imposed forecast errors.