5:15 PM - 7:15 PM
[PPS05-P08] Preliminary Steps Toward Data Assimilation Using the Radiation-Coupled AFES-Venus
Keywords:General Circulation Model, Radiation scheme, Internal Variability, Data Assimilation
Recently, a more reliable depiction of the Venus atmospheric general circulation has been achieved by assimilating observational data from the Venus Climate Orbiter AKATSUKI (Fujisawa et al., 2022) into the general circulation model AFES-Venus (Sugimoto et al., 2017) using the ALEDAS-V system (Sugimoto et al., 2017). Nevertheless, in earlier implementations of AFES-Venus, the radiative processes critical for simulating the Venus thermal structure were simplified by prescribing solar heating and applying Newtonian cooling. To achieve a more physically consistent estimation of the general circulation, we have integrated a radiation scheme for planetary atmospheres based on the correlated k-distribution method (Takahashi et al., 2023) into AFES-Venus and evaluated its performance (e.g., at the JpGU 2024 meeting).
In this study, we explore optimal parameters for data assimilation in preparation for upcoming cycle experiments using wind data retrieved from camera images. Our model assumes an ideal gas with a constant specific heat at constant pressure (Cp = 900 J/(kg·K)), which is a reasonable approximation for the cloud layer above 40 km altitude. However, due to the limitations of the ideal gas approximation with a constant Cp in the lower atmosphere and the uncertain distribution of radiatively active gases, the model exhibits a cold bias even in the cloud layer. Consequently, our model is 'imperfect,' and data assimilation must be tailored to account for these shortcomings. Furthermore, since the observed variable is a camera image containing vertically integrated information about the cloud field, it is important to incorporate non-local vertical information to achieve accurate state estimation.
In this presentation, as an initial step toward developing a data assimilation product with the radiation-coupled model, we investigate the probability distribution of internal variability and estimate the corresponding covariance matrices. We also examine methods for generating initial perturbations within a broad state space characterized by large uncertainties, and report on a parameter study regarding the localization scale and the number of ensemble members.
In this study, we explore optimal parameters for data assimilation in preparation for upcoming cycle experiments using wind data retrieved from camera images. Our model assumes an ideal gas with a constant specific heat at constant pressure (Cp = 900 J/(kg·K)), which is a reasonable approximation for the cloud layer above 40 km altitude. However, due to the limitations of the ideal gas approximation with a constant Cp in the lower atmosphere and the uncertain distribution of radiatively active gases, the model exhibits a cold bias even in the cloud layer. Consequently, our model is 'imperfect,' and data assimilation must be tailored to account for these shortcomings. Furthermore, since the observed variable is a camera image containing vertically integrated information about the cloud field, it is important to incorporate non-local vertical information to achieve accurate state estimation.
In this presentation, as an initial step toward developing a data assimilation product with the radiation-coupled model, we investigate the probability distribution of internal variability and estimate the corresponding covariance matrices. We also examine methods for generating initial perturbations within a broad state space characterized by large uncertainties, and report on a parameter study regarding the localization scale and the number of ensemble members.