Prototype GOES-GOES Stereo 3D Winds with a Path into NOAA Operations

Measurements of winds at a global scale are essential for understanding climate and weather dynamics and for operational weather forecasting. NOAA’s current satellite winds capability is based on tracking a cloud or moisture feature in a sequence of images taken from a single satellite. Accurate assignment of height within the vertical column of atmosphere is as important as accurately measuring the horizontal wind velocity. When using a single satellite, wind height assignments must be derived from cloud-top brightness temperatures and an a priori vertical temperature profile. Improving height assignments has been an important area of recent development. With two GOES-R spacecraft now operational, accurate stereo tracking of cloud and moisture features is now possible. This has the advantage of relying only on basic geometry. The wind height assignment is retrieved from the observed parallax between pairs of satellites as each tracks the feature.

We have developed algorithms based on stereo vision for 3D-Wind retrievals using two GOES satellites. These algorithms are based on similar ones that we have previously developed for GEO-LEO satellite combinations and validated using MISR-GOES [1] and MODIS-GOES [2] satellite combinations. Our prototype retrieval capability (referred to as GOES-GOES 3D Winds) implements these algorithms using the FORTRAN programming language that allows for its future transition into NOAA’s operational GOES-R ground system.

We have prototyped the GOES-GOES 3D Winds capability using MATLAB and offline versions of the GOES-16/17 Level-2 cloud and winds product software in disaggregated components and tested it with Band 2 (visible 0.64 microns), Bands 8-10 (WV 6.19, 6.95, and 7.34 microns), and Band 14 (IR 11.2 microns) using images from GOES 16 and GOES 17. We have run cases where GOES 16 was the reference satellite as well as cases where GOES 17 was the reference satellite. We have cross compared the results against NOAA’s operational GOES-16 and GOES-17 wind products and a MISR-GOES 3D-Wind product we have developed for NASA. The MISR-GOES inter-comparison shows good agreement in both wind vectors and retrieved heights, with small height biases ~200m. This capability has been ported to FORTRAN and has been integrated within an offline processing framework used by the GOES-R Algorithm Working Group’s wind science team. It is now possible to run test cases seamlessly and to compare wind height assignments obtained with the IR method and with rawinsonde measurements. Running GOES-GOES 3D-Winds in this manner is the first step in the path to running it as part of NOAA’s GOES operational winds.

References

1. Carr, J.L.; Wu, D.L.; Kelly, M.A.; Gong, J. MISR-GOES 3D Winds: Implications for Future LEO-GEO and LEO-LEO Winds. Sens. 2018, 10, 1885
2. Carr, J.L.; Wu, D.L.; Wolfe, R.E.; Madani, H.; Guoqing, L.; Tan, B. Joint 3D-Wind Retrievals with Stereoscopic Views from MODIS and GOES. Submitted to . Sens. 2019, 11.