Spherical dome roofs are commonly used in large structures requiring large interior spaces, such as stadiums. They are composed of lightweight materials, such as membranes or metal panels, in a long-span structure. Because of these characteristics, they have high flexibility, small mass, and low natural frequencies, making them sensitive to wind loads. Therefore, it is necessary to consider not only the peak wind pressure acting on the roof but also the fatigue damage caused by wind pressure fluctuations that occur repeatedly over time. The curvature of the dome roof affected the separation point and reattachment location under various conditions, thereby significantly influencing the flow patterns around the structure. Several studies have been conducted on the wind–pressure distribution of spherical dome roofs. Toy et al. [
1] and Taylor [
2] explained that the wind pressure distribution on spherical dome roofs can change according to the approaching flow characteristics and the Reynolds number. Accordingly, various studies have been conducted to clarify the characteristics of wind pressure distribution on dome roofs according to the approaching flow, Reynolds number, and shape. Uematsu et al. [
3] investigated the characteristics of wind pressure distribution according to variations in the rise-span ratio (
f/D) and height-span ratio (
h/
D) to analyze the characteristics of wind pressure owing to shape changes in dome roofs. They confirmed that changes in
f/D significantly affected the wind pressure distribution. Letchford and Sakar [
4] built upon the research of Toy et al. [
1] and Taylor [
2] to investigate the distributions of the average and fluctuating wind pressures on dome roofs at various Reynolds numbers. They observed that the wind pressure distribution remained constant within the Reynolds number range of
to
. In addition, they observed that when the surface roughness increased at these Reynolds numbers, the negative pressure at the apex of the dome roof decreased and the negative pressure in the wake zone increased, leading to a decrease in lift and an increase in drag on the dome roof. Cheng and Fu [
5] investigated wind pressure distribution according to the Reynolds number for various dome roof sizes in smooth and turbulent boundary layers. In smooth boundary layers, they observed the wind pressure distribution to be relatively stable when the Reynolds number was less than or equal to
; however, in turbulent boundary layers, they confirmed that the wind pressure distribution was stable when the Reynolds number was between
and
. Noguchi and Uematsu [
6] conducted a detailed analysis of the wind pressure distribution characteristics according to
f/D and
h/
D and proposed zone-specific wind pressure coefficients for frame and cladding designs. Lee et al. [
7] investigated various
h/
D wind pressure distribution characteristics for elliptical closed and open dome roofs and confirmed that the wind pressure may increase depending on the wind direction, although it shows a similar wind pressure distribution pattern to that of the circular dome roof. Based on the analysis, a peak pressure coefficient was proposed for the cladding design. Rizzo and Avossa [
8] analyzed the aerodynamic characteristics through wind tunnel tests on a large-span project canopy roof for the Pescara football stadium in Italy. According to the results of their research, wind pressure is greatly influenced by wind direction, and, owing to the unusual roof shape, the calculation of the wind load through the wind load code may be partially overestimated compared to the actual wind load. Sun et al. [
9] conducted wind tunnel tests on various wind directions,
f/D, eave heights, and terrain roughness of membrane structures that are widely used in dome roofs. The experimental results confirmed that the mean wind pressure was significantly affected by the wind direction and
f/D, and the terrain roughness had a slight effect on the mean wind pressure; however, the height factor was insignificant. Zhang [
10] analyzed the mean and fluctuating wind pressure distribution characteristics, power spectral density, and coherence of the fluctuating wind pressure between different measurement points through wind tunnel tests on large-span canopy roof structures. Kumar and Stathopoulos [
11] analyzed the characteristics of pressure spectra on flat rectangular roofs and mono-slope roofs and proposed Gaussian and non-Gaussian spectral models based on spectral patterns. Uematsu et al. [
12] investigated the fluctuating wind pressure characteristics on flat circular plane roofs and proposed a spectral model as a function of the geometrical parameters of a building combined with approaching flow turbulence parameters. Lo and Kanda [
13] analyzed the flow patterns around structures by examining the power and cross spectra of the fluctuating wind pressure on spherical domes and proposed a zone division method that can be used for a wind-resistant design based on flow patterns. Sun et al. [
14] analyzed the spectra and patterns under various conditions for 12 dome models with various geometric parameters and proposed zone-specific spectral models divided into three zones. The aforementioned studies by Uematsu et al. [
12], Lo and Kanda [
13], and Sun et al. [
14] focused on flat roofs and spherical domes with
f/D ratios ranging from 0.14 to 0.5, thus limiting their ability to understand the characteristics of the fluctuating pressure spectra for domes with low
f/Ds. Recently, large-space roofs have been predominantly constructed as retractable roofs that are not affected by weather and offer energy-saving effects, such as daylighting and ventilation. These dome roofs have
f/D ratios ranging from 0 to 0.18, with most cases having
f/D ratios below 0.1. Based on the results of research on the general wind pressure distribution, we expect that the flow patterns around the roof may change when
f/D is relatively low [
15,
16,
17,
18,
19,
20]. According to previous studies, the positive pressure is dominant when the
f/D of the dome roof is high, whereas the negative pressure is dominant when the
f/D is low. This is because, as
f/D decreases, the direct effect of the approach flow decreases, and phenomena such as separation occur. Many studies have been conducted on the mean and peak pressures of relatively various
f/Ds values; however, studies on the pressure spectra are limited.
In this study, the characteristics of fluctuating wind pressure were analyzed in a spherical dome with a low f/D. During the wind tunnel experiment, the wind pressure data of spherical domes for various h/D values at an f/D of 0.1 were collected. As mentioned above, the f/D value was targeted at 0.1, which is the average value for an actual dome roof with a low rise-span ratio. Subsequently, the obtained wind pressure coefficient and spectral data were analyzed, and the fluctuating wind pressure characteristics and flow patterns around the roof were discussed in detail. Additionally, the applicability of the spectral model proposed in a previous study was investigated. Based on the analysis results, zone divisions were proposed according to the flow patterns of spherical dome roofs with low f/D values and modified spectral models for each zone.