Comparison of GOES-17 Atmospheric Motion Vectors With AR Recon Dropsonde Data and Assessment of Wind Fields in the Global Forecast System During Atmospheric River Events

October 9, 2025

A paper titled “Comparison of GOES-17 Atmospheric Motion Vectors with AR Recon Dropsonde Data and Assessment of Wind Fields in the Global Forecast System during Atmospheric River Events” was recently published in the AGU’s Journal of Geophysical Research: Atmospheres (JGR-A). This study was led by Minghua Zheng (CW3E), with contributions from F. Martin Ralph (CW3E), Xingren Wu (NOAA/NCEP), Bin Guan (UCLA/JPL), Duane Waliser (NASA JPL/UCLA), Iliana Genkova (NOAA/NCEP), Luca Delle Monache (CW3E), Vijay Tallapragada (NOAA/NCEP), Zhenhai Zhang (CW3E), David Santek (UW–Madison/CMISS), Zhenglong Li (UW–Madison/CMISS), and Scot Rafkin (SwRI).

This study presents an innovative comparative analysis of Atmospheric River Reconnaissance (AR Recon) dropsonde wind observations, GOES-17 atmospheric motion vectors (AMVs), and NCEP Global Forecast System (GFS) wind fields during AR events over the Northeast Pacific Ocean. This collaborative effort aligns with priorities identified in the CW3E 2025-2029 Strategic Plan including Novel Observations, and Atmospheric Rivers and Extreme Precipitation Research, Prediction, and Applications, and was funded by a NASA ROSES GOES-R project in collaboration between UCSD and UCLA/JPL.

AMVs provide horizontal wind estimates by tracking cloud or water vapor features on successive satellite images. The AMV products can augment wind data in data-sparse oceanic areas like those frequented by ARs. However, AMVs exhibit biases and uncertainties, especially due to height assignment issues, and there are limited conventional data (e.g., radiosondes) to assess GOES-17 (i.e., GOES-West) AMVs over oceans. AR Recon missions, which sample ARs to improve forecast skill over the U.S. West, offer a unique opportunity to compare GOES-17 AMVs. This study aims to estimate biases and uncertainties in GOES-17 AMVs in the Northeast Pacific using dropsondes from AR Recon, and assess GFS model wind analyses and background fields during AR events.

The study uses AMV and AR Recon dropsonde wind data incorporated into the operational Global Data Assimilation System (GDAS) (Figure 1). Results show that GOES-R AMVs improved wind data distribution compared to the pre-GOES-R period, particularly in the upper and lower troposphere.

A collocation method was developed to compare AMVs with dropsondes (Figure 2). The comparison reveals that AMVs exhibit a small vector wind speed bias of -0.7 m/s. The uncertainty for AMVs is estimated at 5–6 m/s (Figure 3). Comparison of collocated GFS model background wind fields shows small biases. Data assimilation reduces root-mean-squared differences, but the small biases in operational AMVs warrant further attention, as AMVs are a predominant wind data source in the GFS over oceanic regions. Notably, a substantial portion of the AMVs is rejected during the assimilation. A comparison between these data and dropsondes shows lower biases compared to the assimilated AMVs, although larger uncertainties. Revisiting the assimilation process may enable better use of these data in oceanic regions.

This study demonstrates the critical role of AR Recon data in cross-validating satellite-derived and model wind data over the ocean, underscoring the value of a diverse observing network for understanding atmospheric dynamics and improving model initial conditions. It also highlights the need for high-vertical-resolution wind products and profiles to fill remaining data gaps, particularly in regions prone to extreme weather events such as landfalling ARs originating from the North Pacific Ocean. These findings support the expansion of current AR Recon missions into a global program and the integration of other new promising observational techniques.

Figure 1. An example of GOES AMVs (wind barbs, [knot]) available for GDAS for an impactful AR. AMVs plotted are for IR long-wave (a, report code 245) and water vapor cloud top (b, report code 246), valid for the 6-h assimilation window centered at 0000 UTC 4 February 2020 (AR Recon 2020 IOP 4). Yellow-orange-red shades are the integrated water vapor transport (IVT) amplitude (unit: kg m^-1 s^-1) based on the GFS final analysis valid at 0000 UTC 4 February 2020. Grey contours are sea level pressure from final analyses. Cyan dots are locations of dropsondes collected during 2020 IOP 4. This figure is modified from Fig. 1 of Zheng et al. (2025, JGR-A).

Figure 2. The collocation method for AMVs and dropsondes. An illustration (a) of the collocation criteria to collocate a dropsonde wind observation with surrounding GOES-17 AMV observations; (b) an example of a matched dropsonde (black wind barb and diamond marker) and upper-level AMV (red wind barb and circle marker) wind pair. This figure is modified from Fig. 2 of Zheng et al. (2025, JGR-A).

Figure 3. Scatter plots comparing zonal and meridional winds between assimilated GOES-17 AMVs and collocated dropsonde data. (a) Zonal wind (u-wind) and (b) meridional wind (v-wind) correlations are shown between GOES AMVs and 1540 dropsonde profiles collected during the 6-hour analysis windows of 42 IOPs in AR Recon 2020 and 2021. The color of each point represents the observed pressure (hPa) for the dropsonde data. The black dotted line indicates the 1:1 ratio. (c) Bias and root-mean-squared difference (RMSD) for u-wind, v-wind, vector difference, and the wind speed, with units in m s^-1. Each pair of numbers in square brackets represents the 95% confidence interval for the bias or RMSD values, based on 1,000 bootstrap resamplings. Numbers in green shaded boxes indicate means or biases that differ significantly from zero at the 95% confidence level, based on the bootstrap test. Both correlations are Pearson’s correlations and are significant at the 95% confidence level based on bootstrap testing. This figure is modified from Fig. 5 of Zheng et al. (2025, JGR-A).

Zheng, M., Ralph, F. M., Wu, X., Guan, B., Waliser, D., Genkova, I., Delle Monache, L., Tallapragada, V., Zhang, Z., Santek, D., Li, Z., & Rafkin, S. (2025). Comparison of GOES-17 atmospheric motion vectors with AR Recon dropsonde data and assessment of wind fields in the Global Forecast System during atmospheric river events. Journal of Geophysical Research: Atmospheres 130(8), e2024JD043267. https://doi.org/10.1029/2024JD043267