**1. Introduction**

Typhoons occur frequently in the Western Pacific Ocean (WP) [1], affecting energy exchange at the air–sea boundary layer [2,3] and leading to several secondary hazards, such as extreme waves [4,5], landslides, and heavy rains [6]. Binary typhoons, in which two storms of tropical cyclone intensity or more occur simultaneously, have also been recorded in the WP. Because of the strong wind interactions endemic to binary typhoons, binary typhoons have more complicated effects on the sea surface than single typhoons due to the influence of total heat flux exchange on the upper ocean response [7,8]. At present, moored buoys [9] and satellites [10,11] provide real-time observations of oceanic conditions, particularly winds and waves, during hurricanes and typhoons. However, those devices are unable to generate time-series data with a fine spatial resolution; that is, the resolution of a scatterometer is typically 12.5 km [12], while that of an altimeter

**Citation:** Sun, Z.; Shao, W.; Yu, W.; Li, J. A Study of Wave-Induced Effects on Sea Surface Temperature Simulations during Typhoon Events. *J. Mar. Sci. Eng.* **2021**, *9*, 622. https:// doi.org/10.3390/jmse9060622

Academic Editors: Wei-Bo Chen, Shih-Chun Hsiao and Wen-Son Chiang

Received: 23 April 2021 Accepted: 2 June 2021 Published: 3 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

is 10 km [13]. Thus, the data collected by these devices are not sufficient for long-term distribution analyses.

Over recent decades, as wave theory and computation technologies have matured, several numeric wave models have been proposed. At the beginning of the 1980s, WAMDIG initially proposed the third-generation numeric wave ocean model (WAM) [14], which integrated basic wave propagation effects to describe the evolution of a two-dimensional ocean wave spectrum. Subsequent authors developed the WAVEWATCH-III (WW3) [15] and Simulating Waves Nearshore (SWAN) [16] models based on the principles of the WAM. The main difference between the WW3 and SWAN models is the applicability of the model: SWAN was originally developed as a nearshore model, while WW3 was developed for the oceanic scales. Therefore, the WW3 model is usually employed for wave simulations over large regions, such as global seas [17] or the western Pacific Ocean [18], while the SWAN model is typical used to analyze coastal waters [19]. Model-simulated waves are also commonly used in synthetic aperture radar (SAR) wave monitoring [20,21] and, in particular, as auxiliary data for typhoon analysis [22].

During typhoon events, the sea state is complicated due to strong, synoptic-scale air–sea interactions and turbulent mixing at the sea surface [23]. In particular, sea surface temperatures (SSTs) during typhoons are rarely measured using real-time techniques such as Argos [24]. This air–sea mixing, as well as the Ekman pumping induced by cyclonic vorticity, deepens the mixed layer at the sea surface, decreasing SST [25–27]. Remarkably, observational data have shown a maximum SST cooling of 9 ◦C [28], and short-term satellite data recorded during two typhoon events revealed anomalously cold SST patches that were up to 6 ◦C colder than the SSTs of the surrounding warm tropical sea [29]. SST cooling in turn increases typhoon intensity and movement by modulating energy fluxes and stability at the air–sea boundary layer [30]. Therefore, SST cooling is one of the more noticeable oceanic responses to a moving typhoon due to its significant influence on oceanic and atmospheric dynamics [31]. Due to the limited availability of observational data during typhoons, the coupled atmospheric–oceanic model provides a powerful alternative way to study SST cooling and its impact on typhoon intensity [32].

Recent numerical experiments have aimed to clarify the unique characteristics and underlying mechanisms of SST cooling [33,34]. Two case studies [35,36] have suggested that SST cooling in the inner-core region of the cyclone, which is defined as within a 111-km radius of the cyclone center [37], may weaken typhoon intensity. However, the largest SST reduction often occurs in the right-rear quadrant of the typhoon. Although sea-surface waves themselves act over small scales, ranging from meters to kilometers, wave-induced effects, such as breaking waves, nonbreaking waves, radiation stress, and Stokes drift, affect the air–sea energy exchange at the boundary layer, especially during strong winds. Thus, wave-induced effects should be considered in analyses of SST cooling. The produced cooling is a function of both typhoon forward speed and intensity. Generally, the lower the forward speed, the higher the cooling rate and the higher the intensity, meaning a larger cooling rate is expected [38,39]. The extra cooling and turbulent mixing on the right side of the track in the Northern Hemisphere as a result of the rightward bias can contribute to a larger deepening of the mixed layer [40]. In most cases, binary typhoons are stronger than single typhoons in terms of duration and range, and will cause strong upwelling and mesoscale cyclone vortices in certain areas, which will also have a greater impact on SST [41]. Furthermore, it is important to assess SST cooling during binary typhoons, which include both parallel- and cross-type movements.

The ocean circulation model, which has been termed the Princeton Ocean Model (POM), is commonly used to simulate global marine dynamics [42,43], such as current and SST. An updated version of POM, the Stony Brook Parallel Ocean Model (sbPOM) [44], has been improved and enhanced using the parallel computation technique. The scalability of the POM model is better than that of its predecessors. In this study, we simulated wave fields during certain typhoon events using the WW3 model and calculated four of the effects of strong winds: breaking waves, nonbreaking waves, radiation stress, and

Stokes drift. Subsequently, SST was simulated using an sbPOM that included these four factors. In particular, we focused on fluctuations in SST cooling during various types of typhoon movement.

Indeed, the primary aim of this study was to assess SST cooling during binary typhoons with different paths. The datasets used, which are described in Section 2, include the typhoon events, the forcing wind fields, the open boundary conditions for modeling, and measurements from the Jason-2 altimeter and Argos. The model settings for the WW3 and sbPOM simulations are also given in Section 2. Waves were simulated using the WW3 model and SSTs were simulated using sbPOM, both based on the four effects of the typhoon waves. We assessed the accuracies of these models and the discussions in Section 3. Our conclusions are summarized in Section 4.
