Effects of ocean grid resolution on tropical cyclone‐induced upper ocean responses using a global ocean general circulation model

Abstract

Tropical cyclones (TCs) have the potential to influence regional and global climate through interactions with the upper ocean. Here we present results from a suite of ocean‐only model experiments featuring the Community Earth System Model, in which we analyze the effect of tropical cyclone wind forcing on the global ocean using three different horizontal ocean grid resolutions (3°, 1°, and 0.1°). The ocean simulations are forced with identical atmospheric inputs from the Coordinated Ocean‐Ice Reference Experiments version 2 (COREv2) normal year forcing conditions, featuring global blended TC winds from a fully coupled CESM simulation with a 25 km atmosphere. The simulated TC climatology shows good agreement with observational estimates of annual TC statistics, including annual frequency, intensity distributions, and geographic distributions. Each ocean simulation is composed of a 5 year spin‐up with COREv2 normal year forcing, followed by 18 months with blended TC winds. In addition, we conduct corresponding control simulations for each grid resolution configuration without blended TC winds. We find that ocean horizontal and vertical grid resolutions affect TC‐induced heat and momentum fluxes, poststorm cold wake features, and ocean subsurface temperature profiles. The responses are amplified for smaller grid spacing. Moreover, analyses show that the annually accumulated TC‐induced ocean heat uptake is also sensitive to ocean grid resolution, which may have important implications for modeled ocean heat budgets and variability.

Tropical Cyclone Activity in the High‐Resolution Community Earth System Model and the Impact of Ocean Coupling

“…The spatial correlation coefficient between the observation and each of the model simulations is shown at the upper right corner of the respective figure plot. All models generally produce realistic geographical distributions of TCs, consistent with previous modeling studies with 25 km CAM (Bacmeister et al, ; Li & Sriver, ; Small et al, ; Wehner et al, ). Track density in CPL and SLAB have higher spatial correlation with the observations than in ATM.…”

“…Therefore, we adjust the drag law in the model using the scheme proposed by Moon et al () to achieve a more realistic TC‐ocean interaction. Similar corrections have been used in several TC‐ocean studies, including Li and Sriver () and Bueti et al ().…”

“…Using a higher‐resolution (e.g., 0.1°), ocean model in a fully coupled framework would require extensive model calibration and testing to achieve a steady climate state, which is beyond the scope of this study. In addition, Li and Sriver () examined the sensitivity of TC‐ocean interactions to ocean model horizontal resolution (3°, 1°, and 0.1°) in the POP ocean model. That study found that while the higher‐resolution eddy‐resolving 0.1° ocean simulates the air‐sea interactions under TCs with more detail, the nominal 1° ocean model generally captures both local effects (e.g., surface cooling along tracks) and basin‐scale quantities (e.g., contribution to ocean tropical ocean heat convergence).…”

“…Ocean‐atmosphere coupling is also important for analyzing TC‐induced feedbacks that can influence climate variability. An increasing number of studies have found TC‐induced ocean vertical mixing can lead to ocean heat uptake and transport (Bueti et al, ; Cheng et al, ; Emanuel, ; Jansen et al, ; Jullien et al, ; Li & Sriver, ; Sriver & Huber, ), potentially influencing global SST patterns and affecting large‐scale circulations in the ocean and the atmosphere (Fedorov et al, ; Sriver & Huber, ). These potential feedbacks can affect climate mean‐state and variability on seasonal to interannual time scales.…”

Regional variations in the ocean response to tropical cyclones: Ocean mixing versus low cloud suppression

“…We introduce several key changes to CESM to improve the representation of TC effects, including increased ocean‐atmosphere coupling frequency to capture more realistic representation of near‐inertial internal waves [ Jochum et al, ] and updated surface wind drag coefficient parameterization suitable for high wind regimes [ Moon et al, ]. The model generally simulates realistic TC climatologies (including number, seasonality and intensity distributions), as well as the transient upper ocean responses to TC forcing [ Li et al, ; Li and Sriver , ; McClean et al, ]. The model exhibits some biases, such as reduced TC activity in the North Atlantic and relatively weak storms in the East Pacific (Supporting Information Figures S3 and S4).…”