Preparation of an Objective Analysis Dataset with Akatsuki Horizontal Winds Assimilation for Studying Planetary-Scale Waves in the Venus Atmosphere

At the cloud-top level, around 65–70 km on Venus, there exists a zonal wind with a velocity of about 100 m/s, which is a remarkable phenomenon known as super-rotation. Planetary-scale waves with a zonal wavenumber of 1 and periods of about 4-day and 5-day have also been observed at this altitude and have been interpreted as equatorial Kelvin waves and Rossby waves, respectively (Del Genio and Rossow, 1990 [1]; Kouyama et al., 2015 [2]; Kajiwara et al., 2021 [3]). The intensity of these waves has been shown to vary over time based on long-term observational data of the horizontal wind distribution obtained from cloud tracking by the Ultraviolet Imager (UVI) onboard the Venus orbiter "Akatsuki" (Imai et al., 2019 [4]; Horinouchi et al., 2024 [5]). Numerical model studies suggest that these waves are excited by the Rossby-Kelvin instability and that angular momentum transport by these waves contribute to the long-term variations in super-rotation (Takagi et al., 2022 [6]). 

 

We have previously assimilated horizontal winds obtained from Akatsuki cloud tracking into the world’s first Venus data assimilation system and successfully produced the Venus objective analysis dataset, ALERA-V v1.0 (Fujisawa et al., 2024 [7], 2025 [8]). In addition, in an observation system simulation experiment using the same system, we assimilated synthetic data, including Kelvin waves, and confirmed that Kelvin waves can be reproduced through data assimilation (Sugimoto et al., 2021 [9], 2022 [10]). In this study, we will prepare objective analysis dataset for a period different from that of Fujisawa et al. (2024 [7], 2025 [8]). The goal of this study is to prepare this dataset, which has fewer temporal and spatial restrictions than the observational data, in order to extract a 4-day wave that suggests a Kelvin wave-like structure. This wave was difficult to extract in Imai et al. (2019) [4] due to its weak signal, and it is expected that the structure of this wave can be clarified using the objective analysis dataset.

 

“ALEDAS-V” (AFES-LETKF data assimilation system for the Venus atmosphere; Sugimoto et al., 2017) [11] is used for assimilation, and "AFES-Venus" (Atmospheric General Circulation Model for the Earth Simulator for Venus; Sugimoto et al., 2014) [12] is used for ensemble forecasts. AFES-Venus is a fully nonlinear dynamical general circulation model that assumes hydrostatic equilibrium and is designed for the Venus atmosphere. ALEDAS-V, which uses a local ensemble transform Kalman filter, is the first data assimilation system for the Venus atmosphere. Zonal and meridional winds were assimilated at the cloud top altitude of 70 km using cloud tracking data obtained from the UVI 365-nm images (Horinouchi et al., 2024) [13]. The settings, other than the assimilation period, are the same as those in Fujisawa et al. (2022). The assimilation period is from June 1 to December 31, 2017, during which the existence of Kelvin waves was suggested by Imai et al. (2019) [4]. The objective analysis (assimilated case) is compared with the free run (case without data assimilation).

 

The figure shows the power spectral density of the zonal winds at an altitude of 70 km over a 120-day period starting from July 1, 2017. Panels (a) and (b) represent the free run and objective analysis, respectively. The free run shows signals with periods of 5.8-day and 7.5-day (Panel a), which are comparable to the results from a numerical model by Takagi et al. (2022) [6]. Takagi et al. (2022) [6] indicated that the 5.8-day wave is dominated by a Rossby mode with a peak amplitude at mid-latitudes at an altitude of 70 km. They also noted that the 6.5-7.6-day wave, which is close to the 7.7-day period, is dominated by an antisymmetric structure centered on the equator. The results of the free run in this study show signals at mid-latitudes (50-60 degrees latitude) and high latitudes (70-90 degrees latitude), which differ from the characteristics observed by Takagi et al. (2022) [6]. Future work will need to consider whether these differences are due to variations in the settings of the numerical model.

 

The objective analysis shows signals with periods of 3.7-day and 5.1-day (Panel b). These can be compared with those of Imai et al. (2019) [4], who analyzed the observational data from Akatsuki. Imai et al. (2019) [4] revealed that the 3.5-4.0-day mode shows Kelvin wave characteristics with amplitude confined to the equatorial region, while the 5.0-5.5-day mode shows Rossby wave characteristics with amplitude at latitudes higher than 35 degrees. The 3.7-day and 5.1-day signals in the objective analysis of this study are in good agreement with the characteristics described by Imai et al. (2019) [4]. This indicates that the objective analysis successfully captured the short-period disturbances. In the future, we plan to continue analyzing the structure of these waves.

 

Figure. Power spectral density of zonal winds for high-frequency components: (a) free run, (b) objective analysis. 

 

[1] Del Genio, A. D. & Rossow, W. B. (1990) J. Atmos. Sci., 47(3), 293–318.
[2] Kouyama, T., et al. (2025) Icarus, 248, 560–568.
[3] Kajiwara, N., et al. (2021) J. Geophys. Res. Planets, 126(12), e2021JE007047.
[4] Imai, M., et al. (2019) J. Geophys. Res. Planets, 124(10), 2635–2659. 
[5] Horinouchi, T., et al. (2024) J. Geophys. Res. Planets, 129(3), e2023JE008221.
[6] Takagi, M., et al. (2022) J. Geophys. Res. Planets, 127(4), e2021JE007164.
[7] Fujisawa, Y., et al. (2022) Sci. Rep. 12, 14577.
[8] Fujisawa, Y., et al. (2025) ALERA-V version 1.0: an experimental objective analysis dataset of Venus atmosphere (1.0) [Data set]. Zenodo. 
[9] Sugimoto, N., et al. (2021) Atmosphere, 12, 14.
[10] Sugimoto, N., et al. (2022) Atmosphere, 13, 182. 
[11] Sugimoto, N., et al. (2017) Sci. Rep. 7(1), 9321. 
[12] Sugimoto, N., et al. (2014) J. Geophys. Res. Planets 119, 1950–1968.
[13] Horinouchi, T., et al. (2024) Venus Climate Orbiter Akatsuki Cloud Motion Vector Data Set v1.3, JAXA Data Archives and Transmission System.