Part of the difficulty of forecasting tropical cyclone (TC) track and intensity (where the TC is going and how strong it will be) stems from the lack of frequent, accurate observations over tropical oceans where TCs form and develop. While Hurricane Hunter aircraft can collect vital observations in and near TCs, those observations are limited in time and space. Therefore, satellites are a very important way to fill in the observation gap over sparsely observed areas of the ocean.
One type of satellite observation relies on measuring the reflection of the Global Positioning System’s radio signals off the Earth’s water surfaces. Rough water surface means that less of the radio signal is reflected, and this means that the wind speed is higher than if the water were smooth and more of the signal is reflected. This is the basis of the Cyclone Global Navigation Satellite System (CYGNSS), a group of 8 satellites designed to provide many near-surface wind observations in the Tropics and subtropics, where TCs are most likely to develop and track.
This work estimates the impact the CYGNSS wind observations have on TC track and intensity forecasts around the world. Three sets of two-week forecasts from October 2018 (one forecast per day for two weeks for each set) are produced using NOAA’s global analysis and modeling system. The forecasts differ in that one was created by incorporating only observations already used to make forecasts (the control forecast), and the other two incorporated CYGNSS-derived near-surface wind retrievals in addition to the observations in the control forecast. The two CYGNSS-influenced datasets used either wind speed alone or both wind speed and direction. Since a regional model such as NOAA’s Hurricane Weather and Research Forecast (HWRF) provides more detailed forecasts than the global models, we used the control and scalar wind global forecast datasets as input to the HWRF model. The HWRF forecasts were supplements to the global forecasts to investigate the impact of the CYGNSS observations on forecasts during the developmental phase of Hurricane Michael. This study represents the first study of the impact of CYGNSS data on TC forecasts in a global model with a regional model supplement.
Important Conclusions:
- Both CYGNSS observations of wind speed and velocity (both direction and speed) improved track forecasts around the world by an average of 20-40 km from 36 to 54 h into the forecasts.
- Average intensity forecasts around the world were impacted by around 2 kt or less at all lead times, including sometimes making the forecasts worse.
- Hurricane Michael intensity forecasts from the global model were degraded by 20-40 kt for the 7 October forecast in the CYGNSS wind speed test due to lack of rapid intensification; however, HWRF forecasts using the same GFS forecast as input produced notable improvements in track and especially intensity forecasts. It is important to understand why incorporating the CYGNSS data into one model led to worse forecasts while in another model improved them, so this will be a focus of future research.
The full study can be accessed at https://journals.ametsoc.org/view/journals/mwre/aop/MWR-D-21-0094.1/MWR-D-21-0094.1.xml. For more information, contact aoml.communications@noaa.gov. This research is supported by NOAA’s Climate Adaptation and Mitigation Program (CAMP) and administered by UCAR’s Cooperative Programs for the Advancement of Earth System Science (CPAESS) under awards 567 NA18OAR4310253, NA18OAR4310253B and NA18OAR4310253C.