**2. Materials and Methods**

Four typhoons that passed through the China Sea were analyzed in this study: Tembin, Bolaven, Chan-hom, and Linfa. The tracks of these typhoons were identified as either parallel- or cross-type. Superimposition of the tracks of these four typhoons over the water depths obtained from the General Bathymetry Chart of the Oceans (GEBCO) data showed that the paths of Tembin and Bolaven (19–30 August 2012) were almost parallel, while the path of Chan-hom intersected that of Linfa (7–12 July 2015; Figure 1).

**Figure 1.** The tracks of the typhoons analyzed in this study superimposed on the water depth data from the General Bathymetry Chart of the Oceans (GEBCO).

Two numeric models, a WW3 model and a sbPOM, were implemented for the China Seas. The European Centre for Medium-Range Weather Forecast (ECMWF) has provided operational products for the global seas since 1958; ECMWF data have a high resolution (up to 0.125◦) at intervals of 1 h. Using the ECMWF wind data, we hindcasted the long-term wave distributions previously simulated using SWAN and WW3 models in two Chinese seas: the Bobai Sea [45] and the South China Sea [46]. The results indicated that the modelsimulated waves were consistent with buoy data and altimeter measurements. However, simulated measurements were systematically underestimated as compared to observational

measurements [47]. In a recent study, we reconstructed "H-E winds," composited of ECMWF winds and a parametric Holland model [15], for typhoons. The model was trained by fitting the shape parameter to buoy-measured observations. We then compared simulated wind speeds with those measured by moored buoys, and found that the root mean square error (RMSE) of wind speed was less than 3 m/s for the shape parameter equivalent to 0.4 [15]. A representative example of H-E wind fields, for Typhoons Chanhom and Linfa at 18:00 UTC on 5 July 2015, is shown in Figure 2. In this figure, two typhoon centers are clearly apparent in the H-E wind fields. Importantly, the underestimation of ECMWF winds was improved by reanalysis using H-E winds.

**Figure 2.** The wind map for Typhoons Chan-hom and Linfa at 18:00 UTC on 5 July 2015. Wind maps were composited using the European Centre for Medium-Range Weather Forecasts (ECMWF) wind data and a parametric Holland model (H-E).

The initial fields for the SST simulation of sbPOM were the monthly average SST and the sea surface salinity from the Simple Ocean Data Assimilation (SODA) data, which have a spatial grid resolution of 0.5◦. A representative example from August 2012 is shown in Figure 3. The upper boundary forcing fields were obtained based on the National Centers for Environmental Prediction (NCEP) total heat flux parameters (latent heat flux, sensible heat flux, long-wave radiation, and short-wave radiation) at 6 h intervals and a spatial resolution of 1.875◦ × 1.905◦ (longitude × latitude); in contrast, ECMWP provides flux data twice per day. Argos is an international cooperative project begun in 2000. The Argos project, which aims to profile ocean temperature and salinity, is a major component of many ocean observation systems [48]. In this study, we used the high-quality SST measurements from Argo project to validate the simulations of the sbPOM. As an example, the map of NCEP total heat flux at 18:00 UTC on 5 July 2015 is shown in Figure 4; in this figure, triangles represent the geographic locations of the available Argos stations (>20) used in this study.

**Figure 3.** The Simple Ocean Data Assimilation (SODA) data used in the Stony Brook Parallel Ocean Model (sbPOM) on August 2012. (**a**) Monthly average sea surface temperature. (**b**) Monthly average sea surface salinity.

**Figure 4.** The National Centers for Environmental Prediction (NCEP) total heat flux at 18:00 UTC on 5 July 2015, overlaid with the geographic locations of the Argos stations used in this study.

In addition, SWH measurements were obtained from the altimeter Jason-2, which is a follow-on satellite to that of the Jason-1 oceanography mission of the National Aeronautics and Space Administration (NASA) and is a valuable source of global wave distributions [49]. These SWH data were used to validate the WW3-simulated SWH data. The WW3-simulated SWH map was overlaid the data from the Jason-2 altimeter at 18:00 UTC on 5 July 2015 (Figure 5). In this figure, the two cyclone-induced wave patterns are clearly visible. The basic settings of the WW3 model and the sbPOM (e.g., forcing fields, open boundary conditions, and output resolution) are summarized in Table 1. The details of the WW3 model and the sbPOM are given in Appendices A and B, respectively.

**Figure 5.** The significant wave height (SWH) data obtained from the WAVEWATCH-III (WW3) model at 18:00 UTC on 5 July 2015, overlaid with the footprints of the Jason-2 altimeter.


