Observational Study on Wind Profiles and Wind Characteristic Parameters of Typhoons over Coastland and Coastal Sea of China

Tropical cyclones (TCs) are one of the most destructive natural disasters that cause severe economic losses and human casualties in many parts of the world. Understanding TC extreme winds and related wind characteristics is of great importance for both wind engineering and meteorology. In the field of wind engineering, the strong wind profile serves as prerequisite input for wind-resistant design of wind-sensitive structures such as high-rise buildings, long-span bridges, and large wind turbines (Song et al., 2016). In meteorology, parameterization of wind characteristic parameters such as the drag coefficient (C_d) and aerodynamic roughness length (Z₀) is crucial for prediction of TC track and intensity in atmospheric numerical models (Bell et al., 2012; Feng et al., 2021; Moon et al., 2004). However, accurate representation of the structure of wind profiles and the wind characteristics throughout the boundary layer, and associated physical processes such as air–sea/air–land exchanges of momentum, heat, and moisture, remains difficult for the case of landfalling typhoons because of lack of observational data (Duan et al., 2019).

Former studies show that the wind characteristics of typhoons over coastland have been studied separately to those over the coastal sea; however, research on the wind characteristics of landfalling typhoons in the sea–land transition area is still lacking. Relatively little is known about the spatiotemporal evolution of the near-surface wind field during typhoon landfall, and appropriate representation of the wind characteristics in such cases presents a notable challenge to researchers wishing to analyze typhoon wind characteristics in coastal areas. The main objective of the present work was to elucidate the variation in wind characteristic parameters with wind speed in both coastland and coastal sea areas during typhoon landfall.

This study comparatively investigated the near-surface wind profiles and wind characteristic parameters of landfalling typhoons over coastland and coastal sea areas of China. The typhoon observational data were obtained at 17 observation points using Doppler wind lidars and wind tower anemometers during the passage of 15 typhoons that made landfall over the southeastern coast of China during 2009–2020. Specifically, the three wind characteristic parameters of the aerodynamic roughness length, friction velocity (U_*), and drag coefficient, which were obtained via the logarithmic law wind profile method, were investigated at coastal areas of China. Moreover, this study paid special attention to the variation in the drag coefficient with wind speed under the onshore wind condition.

As shown by the typhoon near-surface wind profiles over land and over the sea, although wind speed does not increase monotonically with height, the normalized wind speed does generally increase with increasing height. With increase in wind speed, the near-surface wind profiles become much closer to the power law wind profile. The power exponent (α) reaches its maximum value (0.1062) when U(10) is 18–22 m s⁻¹.

In this study, the wind characteristic parameters (i.e., Z₀, U_*, and C_d) were acquired using the least square fitting of the logarithmic law wind profile under neutral stratification. It was found that Z₀ is substantially greater over land than over the sea at low wind speeds. When U(10) is >14 m s⁻¹, the value of Z₀ over the sea starts to exceed that over land owing to drag arising from sea surface waves. The reduction in Z₀ with increase in wind speed at low wind speeds over land is presumably due to the diminishing role of viscous effects and increase in the turbulent dissipation rate. Friction velocity over land was found to be smaller than that over the sea at high wind speeds (>18 m s⁻¹), which contrasts the observation-derived results of Zhang et al. (2011). Because observational data at extreme wind speeds are lacking in this study and in earlier studies, clarity over whether U_* levels off or decreases slightly with increase in wind speed at high wind speeds over land remains elusive and needs to be verified in further studies. The critical wind speed at which C_d peaks over the coastal sea was found to be 24 m s⁻¹, which is approximately 6–16 m s⁻¹ lower than that over deep water.

Notably, the variation in C_d with U(10) shows a double-peak distribution under the condition of an onshore wind over the sea, which has never been reported in earlier observational studies conducted over either open ocean or shallow sea areas. Two possible reasons that included water depth change and the relative direction between wind and waves were proposed to explain the double-peak distribution in C_d. Finally, a formula was proposed by means of the linear least square fitting method to describe the variation in C_d with U(10) under the condition of an onshore wind over land during typhoon landfall. This formula is expected to be applied in the surface layer scheme to improve numerical simulation results of landfalling typhoons.

This study examined the wind characteristic parameters of typhoons over coastland and coastal sea areas of China. The findings will provide reference for further scientific research of the typhoon boundary layer and the development of the typhoon boundary layer scheme in numerical models. Although abundant wind profile data of typhoons have been used, observational studies of typhoons in the northwestern Pacific are demanded in further research because observational data at extreme wind speeds remain lacking.

Figure caption: Variations in (a) roughness length (Z₀), (b) friction velocity (U_*), and (c) drag coefficient (C_d) with 10-m mean wind speed (U(10)) under four conditions: offshore wind over land, onshore wind over land, offshore wind over the sea, and onshore wind over the sea. Panel (d) is an enlarged drawing of the variation in C_d with U(10) under the condition of an onshore wind over the sea. The values between the two whisker ends represent the 95% confidence interval. The numbers above the abscissa denote the number of samples at each wind speed interval.