Ocean Surface Wind Retrieval from ALOS-2/PALSAR-2 and Application Studies Using SAR Winds

Information on tropical cyclone (TC) wind structure [e.g., the radius of maximum wind (RMW), wind radii, and asymmetric structure] is essential for both disaster prevention and deepening scientific understanding of TCs. In the western North Pacific, however, it has been difficult to estimate TC wind structure in the inner core, including the RMW and asymmetric structure, due to the lack of high wind observations. The advent of C-band synthetic aperture radar (SAR) has led to a breakthrough in observing high winds with high spatial resolution in the inner core of TCs.

Japan Aerospace Exploration Agency (JAXA) and Meteorological Research Institute of Japan Meteorological Agency (MRI/JMA) are currently working to develop a new L-band SAR wind product that uses the Phased Array L-band Synthetic Aperture Radar-2 (PALSAR-2) aboard the Advanced Land Observing Satellite-2 (ALOS-2), whose local sun time in a sun-synchronous sub-recurrent orbit is ~12:00 in descending pass and ~00:00 in ascending pass. Because a C-band SAR satellite can observe TCs twice a day in a sun-synchronous sub-recurrent orbit with a local time of ~06:00 in descending pass and ~18:00 in ascending pass, the development of L-band SAR wind products can contribute to increasing the frequency of SAR observations at intervals of up to 6 hours, which is the same time interval at which operational centers issue TC advisories. Specifically, in the L-band SAR development, we collect match-ups between PALSAR-2 HV σ⁰ and the Stepped Frequency Microwave Radiometer (SFMR) surface winds observed by aircraft, develop Geophysical Model Functions (GMFs), retrieve ocean surface winds using the GMFs, and validate retrieved ALOS-2 winds. We confirmed the capability of high wind speed retrieval up to 55 m s^–1. We, however, still struggle with some quality issues caused by Radio Frequency Interference (RFI), Noise Equivalent Sigma Zero (NESZ), Faraday rotation in the ionosphere, and rain attenuation.

We also conduct application studies using SAR wind data. In the latter half of this presentation, we will show the consistency and inconsistency of SAR winds equivalent to a 1-min sustained wind speed with the conventionally estimated 10-min sustained wind best track data (i.e., JMA best track data), methods and results of estimating wind structure parameters using SAR winds as a training dataset, and results of SAR wind assimilation into a numerical model.