1 Introduction

Marine dynamic disasters, such as storm surges, enormous waves, and tsunamis, are the most harmful natural disasters to coastal countries worldwide (NRC, 1999; Trenberth et al., 2015; Hu et al., 2019). In the context of global warming and sea-level rise (SLR), some changes in spatiotemporal features have emerged, affecting the scope and frequency of marine dynamic disasters.

Storm surges have been the most severe marine disasters, affecting the lives and properties of people due to their high frequency and intensity (Ye, 2017; Jin et al., 2018). The water level significantly rises when the storm surge coincides with the high tide of the astronomical spring tide. Driven by the strong power generated by the nonlinear superposition of typhoons, waves, and tides, the ultrahigh water level will invade the inland near the coast, destroy the dike, submerge the land, and destroy the houses near the coast and other facilities, which will introduce disasters to the people (Dawson et al., 2009; Neumann et al., 2015).

The SLR caused by climate change raises the basic water level of storm surges and tides, directly leading to an increase in extreme water level (EWL) during the storm surge. In periods of high tide and intense storms, EWL events, such as flooding, will occur frequently and intensely (Woodworth et al., 2011; Wahl & Chambers, 2015). In earlier research, Yang et al. (1995) found that the return period of the storm surges along Tianjin and Shanghai will change from a 100- to a 10-year return period when the sea level rises by 0.50 m. However, the current rising rate of SLR is twice as fast as in the twentieth century, approximately 3.6 mm per year (Church et al., 2013). SLR, due to climate change, will significantly shorten the return period of EWL with its acceleration. By 2050, the 100-year return period of EWL will become a 10- or 30-year period, while the 1000-year return period of EWL will be shortened to a 10-year return period by 2100 (Wu et al., 2017). These phenomena may lead to severe losses for countries prone to storm surge disasters (Bengtsson et al., 2007; Zhang & Li, 2019).

The influence of the SLR on storm surges comes from two aspects. On the one hand, the increased offshore water depth caused by SLR will affect the astronomical tidal range, thus influencing the risk of storm surges. On the other hand, the surge caused by typhoons of the same intensity will also be affected. Moreover, studies in Qingdao, China, show that both effects strongly impact storm surges (Feng et al., 2014; Ma et al., 2022).

The rising sea surface temperature (SST) is another factor influencing the storm surge, which provides additional energy for storms formed at sea. The current scientific consensus is that this factor will decrease the number and increase the intensity of tropical cyclones globally (Emanuel, 1987; Holland, 1997; Karl et al., 2009). However, regional differences are observed. Feng et al. (2014) found that increased SST may indirectly impact the intensity of the surges in Qingdao.

Overall, climate change will significantly increase the severity of future storm surge disasters. Therefore, strengthening the extreme risk research on storm surge disasters under climate change is significant to protecting coastal zones.

Qingdao (Fig. 1) is located on the southeast coast of the Shandong peninsula, facing the Yellow Sea. The storm surge is one of the leading marine disasters affecting Qingdao. As a coastal city, Qingdao occasionally suffers from tropical storm surges. Yuan et al. (2012) indicated that the frequency of typhoons affecting Qingdao is approximately once a year, with the average occurrence of light storm surge disasters of approximately 1.2 years, heavy disasters of approximately 5 years, and particularly severe disasters of approximately 10 years. Historically, two typhoons have directly hit Qingdao. During the first typhoon ever recorded that hit Jiaozhou Bay in August 1939, 17 people were killed, and $4 million (c. 1939) in damages occurred in the city area alone (Cai et al., 2010). Another severe storm surge disaster occurred on August 19, 1985. The typhoon 'Mamie' (8509) killed 29 people, injured 368 people, destroyed more than 8,000 m of sea dikes, and caused approximately 500 million RMB in economic losses (Lin et al., 1998). Considering global climate change, typhoon intensity and frequency will change, introducing new storm surge characteristics and affecting the storm surge risk in Qingdao.

Fig. 1
figure 1

Potential floodplain defined from 6 m contour of Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) in Qingdao. The gray color denotes the area lower than 6 m elevation

Full size image

Therefore, predicting and understanding storm surge trends affected by climate change in Qingdao is essential for developing climate adaptation measures. However, research on future risks related to storm surges and inundations in Qingdao is limited. Previous studies can be divided into two categories in terms of the storm surge in Qingdao. One type of analysis uses observational data, such as long-term water levels, wind speeds, and other coastal site data, to examine the occurrence pattern of storm surges and their characteristics (Dong et al., 2005, 2015; Zhang et al., 2006). The second type mainly uses numerical simulation to study the impact of storm surges of different intensities and paths on Qingdao. Cao et al. (2013) established a storm surge inundation model in Jiaozhou Bay to study the impact of runoff and dampproof dikes on the inundation. Feng et al. (2014) studied the influence of SLR and SST changes on storm surges in the Qingdao area. Climate change directly affects coastal areas via storm surges, which cause disastrous flooding. Studies on the impact of future changes in SLR and typhoon intensity on the inundations induced by typhoon storm surges in Qingdao under the influence of climate change are currently unavailable.

Therefore, this study focuses on investigating the impact of climate change on storm surges and inundations induced by different typhoon tracks (Fig. 2) in Qingdao. The year 2010 is taken as the present-day scenario. Notably, rising SLR and SST in 2035 and 2055 are emphasized when considering future climate change scenarios. The typhoon intensity change is influenced by the rising SST. Meanwhile, three different levels of adaptation are defined in accordance with the Qingdao future development plan to adapt to the changes in the storm surges and the actual situation in Qingdao.

Fig. 2
figure 2

Eight typical typhoon tracks

Full size image

The details of the rest of the paper are presented as follows. The introduction to the ADvanced CIRCulation (ADCIRC) model and model configuration are given in Sect. 2. After the validation of the model in Sect. 3, simulation results are then provided in Sect. 4. The discussion is presented in Sect. 5. Finally, conclusions are drawn in Sect. 6.

2 Method

The ADCIRC model was used to establish the storm surge numerical model in the Qingdao coastal area to study the risk of floodplains under different typhoon scenarios.

2.1 Model description

The ADCIRC model for shelves, coasts, and estuaries is used to simulate the typical storm surges and inundations in the Qingdao area. The ADCIRC model is a finite element ocean model developed by the University of North Carolina. This model uses unstructured triangular mesh and can be applied to the hydrodynamic calculation model of the ocean, coast, and estuary. The model can comprehensively consider the influence of various complex boundary conditions and complex dynamics, reasonably describe the nearshore flow field, and effectively simulate the nearshore complex shoreline (Westerink et al., 1992), which is widely used in storm surge risk assessment systems at home and abroad.

The governing equations of the model are two-dimensional shallow water equations comprising continuous and momentum equations. The ADCIRC basic equations in the Cartesian coordinates are

$$\frac{\partial \zeta }{\partial t}+\frac{\partial UH}{\partial x}+\frac{\partial VH}{\partial y}=0,$$
(1)
$$\frac{\partial U}{\partial t}+U\frac{\partial U}{\partial x}+V\frac{\partial U}{\partial y}-fV=-\frac{\partial }{\partial x}\left[\frac{{P}_{s}}{{\rho }_{0}}+g\zeta \right]+\frac{1}{H}\left[{M}_{x}+\frac{{\tau }_{sx}}{{\rho }_{0}}-\frac{{\tau }_{bx}}{{\rho }_{0}}\right],$$
(2)
$$\frac{\partial V}{\partial t}+U\frac{\partial V}{\partial x}+V\frac{\partial V}{\partial y}+fU=-\frac{\partial }{\partial y}\left[\frac{{P}_{s}}{{\rho }_{0}}+g\zeta \right]+\frac{1}{H}\left[{M}_{y}+\frac{{\tau }_{sy}}{{\rho }_{0}}-\frac{{\tau }_{by}}{{\rho }_{0}}\right],$$
(3)

where \(\left(x,y\right)\) is the Cartesian coordinate; \(\zeta\) is the water surface elevation; \(U\) and \(V\) are the depth-averaged velocities in the \(x\) and \(y\) directions, respectively; \(H=h+\zeta\) represents total water depth, \(h\) is the water depth relative to datum; \(f\) is the Coriolis parameter; \({P}_{s}\) is the sea surface pressure; \({\rho }_{0}\) is seawater density; \(g\) is the gravitational acceleration; \({M}_{x}\) and \({M}_{y}\) are horizontal momentum diffusions in the \(x\) and \(y\) directions, respectively; \({\tau }_{sx}\) and \({\tau }_{sy}\) are the wind stress at the sea surface; \({\tau }_{bx}\) and \({\tau }_{by}\) are the bottom stress.

At the bottom, a nonlinear mixed bottom friction formula is used,

$${\tau }^{*}=\frac{{C}_{f}{\left({U}^{2}+{V}^{2}\right)}^{1/2}}{H}.$$
(4)

For the bottom friction coefficient,

$${C}_{f}={C}_{fmin}{\left[1+{\left(\frac{{H}_{break}}{H}\right)}^{\theta }\right]}^{\frac{\lambda }{\theta }},$$
(5)

where \({H}_{break}\) is the critical depth, \(\theta\) determines the rapid approach of the hybrid bottom friction relationship, and \(\lambda\) determines the speed of the bottom friction growth from deep to shallow water. According to the formula, \({C}_{f}\) will be close to \({C}_{fmin}\) in deep water (\(H\)>\({H}_{break}\)) and to \({C}_{fmin}{\left({H}_{break}/H\right)}^{\lambda }\) in shallow water (\(H\) < \({H}_{break}\)). \({C}_{fmin}\) is set in this study as 0.0011.

Overflows and inundations occur when the water level is above the land elevation, and the ADCIRC model uses Henderson’s classic hydraulic formula to calculate spillway discharge.

$$Q={C}_{m}\frac{2}{3}\sqrt{2g}{h}_{1}^\frac{2}{3},$$
(6)

where Q is the spillway discharge, Cm is the drainage factor, g is the gravitational acceleration, and h1 is the height above the seawall.

$${C}_{m}=0.611\left[{\left(1+\frac{{v}_{1}^{2}}{2g{h}_{1}}\right)}^\frac{3}{2}-{\left(\frac{{v}_{1}^{2}}{2g{h}_{1}}\right)}^\frac{3}{2}\right],$$
(7)

where V1 represents the current speed.

2.2 Model configurations

The data source, study area, and atmospheric forcing field of the constructed storm surge model are comprehensively introduced in this section.

2.2.1 Data sources

The model bathymetry is extracted from the SKKU bathymetry data (Choi et al., 2002) with 1′ resolution. Additionally, the coastline data comes from Choi et al. (2002). The model coastline of Qingdao is manually modified in accordance with the satellite image from Google Earth due to the coastal and estuarine topographies, which play an important role in storm surges (Ding & Wei, 2016). The Shuttle Radar Topography Mission (SRTM) data (Jarvis et al., 2008) with a resolution of 90 m is used as the altitude data. A deviation of 35.7 cm exists between the 1985 national elevation reference point and the quasi-geoid defined by WGS84 (Guo et al., 2004). Therefore, the SRTM DEM and topography are corrected to establish a foundation for calculating storm surges and inundations.

Seventy-one typhoons that affected Qingdao from 1949 to 2014 can be divided into eight types according to their tracks (Wang et al., 2020). Among them, 18 were Type A, 16 were Type B, 9 were Type C, 8 were Type D, 7 were Type E, 5 were Type F and G each, and 3 were Type H. The eight typical typhoon tracks (Fig. 2) were chosen from 71 tropical cyclones affecting Qingdao according to the typhoon track and intensity, and these tracks are denoted as Tracks A to H (Wang et al., 2020). The typical typhoon case of each type is listed in Table 1.

Table 1 Typical typhoon case of each type in this study
Full size table

Information on typhoons, such as the moving path, the central pressure, and the maximum wind speed, can be downloaded from the CMA-STI Best Track Database (Ying et al., 2014).

The temperature increase will directly and indirectly influence the typhoon intensity. The scenarios of typhoon intensity in 2035 and 2055 are from Wang et al. (2020).

The SLR in the future will be obtained from the Ministry of Natural Resources of China (2023). In that bulletin, the sea level along the Chinese coast in 2050 is estimated to rise by 0.21, 0.22, and 0.24 m under three different scenarios compared with 1995–2014. The selected typical value is 0.22 m, and 2005 is taken as the reference year. Therefore, the sea level will rise by 15 and 25 cm in 2035 and 2055, respectively, compared with 2010 if the linear relation is assumed. The rising sea level will increase the water level of the coastal zone and lead to additional inundation areas.

2.2.2 Study area

The ADCIRC model uses an unstructured triangular mesh of varying resolutions to calculate storm surges to improve the accuracy of the simulation and computational efficiency. The domain of the model covers the entire Bohai Sea, the Yellow Sea, and the East China Sea (Fig. 3). The model has a horizontal resolution varying from 18.5 km south of Taiwan Island to 3 km along the coast of the Yellow Sea. The grid is refined in Qingdao and its surrounding areas, with a resolution of 900 and 300 m in the floodplain area. The floodplain grid on land with a 6 m elevation is mapped (Fig. 1) because the maximum EWL in Qingdao in the past year was 551 cm (Li et al., 2009). In addition, the resolution at the junction of the shoreline and the open boundary is linearly transitioned to increase calculation stability.

Fig. 3
figure 3

Model grid of the storm surge model in Qingdao. Left: the grid of the entire research area; Right: the refined grid for the vicinity of Qingdao

Full size image

The normal velocity Vn = 0 is taken at the boundary of the continent and the island, which satisfies the condition that the coastal boundary cannot be penetrated. At the open boundary, the tidal level of eight constituents (M2, S2, K1, O1, N2, K2, P1, and Q1) is provided by Oregon State University Tidal Prediction Software (Egbert & Erofeeva, 2002).

The mode is set to a cold start. At the initial time t = 0, the water level of each grid point ζ = 0, and the horizontal component of the flow field u = 0, v = 0.

2.2.3 Typhoon diagnostic model

The circular typhoon model of Jelesnianski is used (Jelesnianski, 1965) for typhoon events in this study.

When \(0 \le r \le R\),

$$\left\{\begin{array}{c}W_x=V_x\frac r{r+R}-W_R\left(\frac rR\right)^\frac32\left[\left(x-x_0\right)\sin\;\theta+\left(y-y_0\right)\cos\;\theta\right]\\W_y={\mathrm V}_{\mathrm y}\frac{\mathrm r}{\mathrm r+\mathrm R}+W_R\left(\frac rR\right)^\frac32\left[\left(x-x_0\right)\cos\;\theta-\left(y-y_0\right)\sin\;\theta\right].\\P_a=P_0+\frac14\left(P_\infty-P_0\right)\;\left(\frac rR\right)^3\end{array}\right.$$
(8)

When \(r > R\),

$$\left\{\begin{array}{c}{W}_{x}={V}_{x}\frac{R}{r+R}-{W}_{R}{\left(\frac{r}{R}\right)}^{\beta }\left[\left(x-{x}_{0}\right)\mathrm{sin}\theta +\left(y-{y}_{0}\right)\mathrm{cos}\theta \right]\\ {W}_{y}={\mathrm{V}}_{\mathrm{y}}\frac{\mathrm{R}}{\mathrm{r}+\mathrm{R}}+{W}_{R}{\left(\frac{r}{R}\right)}^{\beta }\left[\left(x-{x}_{0}\right)\mathrm{cos}\theta -\left(y-{y}_{0}\right)\mathrm{sin}\theta \right].\\ {P}_{a}={P}_{\infty }-\frac{3}{4}\left({P}_{\infty }-{P}_{0}\right)\left(\frac{R}{r}\right)\end{array}\right.$$
(9)

where \(\left({x}_{0},{y}_{0}\right)\) and \(\left(x,y\right)\) represent the position of the typhoon center and grid point, respectively; \({W}_{x}\) and \({W}_{y}\) are wind speed components in the \(x\) and \(y\) directions, respectively; \({V}_{x}\) and \({V}_{y}\) are the moving speed components of the typhoon center; r is the distance between the typhoon center and grid point; R is the maximum wind speed radius; \(\theta\) is the flow angle and is equal to 20°; \(\beta\) is the attenuation coefficient; \({P}_{0}\) is typhoon central pressure; \({P}_{a}\) is the selected point pressure; and \({P}_{\infty }\) is the peripheral pressure.

Typhoon information on the moving path \(\left({x}_{0},{y}_{0}\right)\) and center pressure \({P}_{0}\) can be downloaded from the CMA-STI Best Track Database (Ying et al., 2014).

The maximum wind speed \({W}_{R}\) is calculated by this formulation proposed by Atkinson and Holliday (1977):

$${W}_{R}=3.029\times {\left({P}_{\infty }-{P}_{0}\right)}^{0.644}.$$
(10)

The maximum wind speed radius \(R\) uses the maximum wind speed radius empirical formula (Pan, 2011):

$$R=28.52th\left[0.0873\left(\varphi -28\right)\right]+12.22exp\left(\frac{{P}_{0}-{P}_{\infty }}{33.86}\right)+0.2V+M,$$
(11)

where \(\varphi\) is the latitude of the typhoon center, \(V\) denotes the moving speed of the typhoon center, and \(M\) represents the starting radius.

3 Model validations

The accuracy of the model determines the accuracy of the simulation results. Thus, the astronomical tides and storm surges are verified.

3.1 Astronomical tide validation

Accurately simulating the astronomical tide is the basis for simulating storm surges. Therefore, water levels in the China Sea were simulated from August 1 to 30, 2008, to evaluate their performance.

Cotidal charts are obtained for each tide constituent (results of M2 and K1 are displayed in Fig. 4) using harmonic analysis. A comparison with the Ocean Atlas (Chen, 1992) shows that model results match the observed amplitude and phase characteristics.

Fig. 4
figure 4

Calculated cotidal charts in the East China Seas. Left: M2; right: K1. The black dotted lines indicate amplitude, and the solid lines indicate phase

Full size image

A harmonic analysis of the observed water level at Wuhaomatou and Xiaomaidao stations from Chen (2011) was performed to examine the error of the amplitude and phase of tide constituents, and the results are compared with those of the proposed model (Table 2). A comparison of the amplitude and phase of simulated water levels with measured data reveals that the absolute error of the simulated tidal amplitude was 0.39 and 7.50 cm. The absolute error of the simulated tidal phase was between 0.29° and 6.11°. The tidal simulations of M2, S2, K1, and O1 agree with the observations, and the accuracy of the simulation is acceptable. This finding demonstrates that the ADCIRC model can accurately simulate the tidal movement in Qingdao.

Table 2 Comparison of the tidal constituents of the model and observation
Full size table

3.2 Storm surge validation

Three severe typhoon storm surge events, namely, typhoon 8509 Mamie, typhoon 1109 Muifa, and typhoon 1509 Chan-Hom, are simulated to evaluate the performance of ADCIRC in simulating the storm surge in the Qingdao area. The simulated and observed storm surge elevations are presented in Fig. 5, and the absolute and relative errors associated with each simulated maximum surge are listed in Table 3.

Fig. 5
figure 5

Simulated and measured surge heights. Upper: Wuhaomatou (a) and Lianyungang (b) during Mamie (8509, 1985–08-17). Middle: Wuhaomatou (c) and Chengshantou (d) during Muifa (1109, 2011–08-06). Lower: Xiaomaidao (e) and Rizhao (f) during Chan-hom (1509, 2015–09-18)

Full size image
Table 3 Absolute and relative errors of simulated maximum surges at gauge stations
Full size table

Typhoon Mamie (8509) caused severe damage to properties and casualties (Lin et al., 1996; Feng et al., 2014). Figures 5 (a) and (b) show that variations in water level are consistent for Wuhaomatou and Lianyungang stations. The relative error of the simulated maximum surge is 14% at Wuhaomatou, and the phase is also consistent with that observed. The relative error is 2%, which is approximately 1 h earlier than observed for the Lianyungang station. Figures 5 (c) and (d) show that the simulated results for Typhoon Muifa are consistent with the observed surge height. The relative error of the simulated maximum surge water is 17% and 9% at the Wuhaomatou and Chengshantou stations, respectively, with the maximum surge occurring approximately 1 h later than observed at the latter station. In the case of Typhoon Chan-hom, the relative error of maximum storm surge elevation is 10% at Xiaomaidao station (Fig. 5e). In Rizhao station (Fig. 5f), the relative error of maximum surge water is 28%, but the surge height is only 0.40 m, which is insignificant.

Nevertheless, the phase error of the maximum surge was approximately zero. Overall, the simulated storm surge by the ADCIRC model agrees with actual observations, not only in the storm surge elevation trend but also in the characteristics of maximum surge elevation. Despite the presence of deviations in the occurring phase, the simulations are considered acceptable, and the parameter settings and calculation methods are reasonable and reliable.

4 Impact of different intensity defense measures on storm surge under climate change scenarios

Climate change and its impacts are inevitable. Climate change will intensify storm surges by affecting the frequency and intensity of storms, which causes the accumulation of seawater near the shore, thereby increasing the likelihood of storm surge flooding. The impact of future climate change on storm surge floodplains along the coast of Qingdao is discussed through the comparison of different dam defense levels, which provides a reference for preventing storm surge disasters in the city.

4.1 Scenario setting

The floodplain change during a storm surge in Qingdao under the current scenario and three future scenarios is investigated to explore the impact of future climate change on the Qingdao storm surge floodplain. The three future scenarios are mainly CLA, CA, and EA. CLA refers to the current level of adaptation strategy, CA refers to the continuation level of adaptation strategy, and EA refers to the enhanced level of adaptation strategy. Notably, the future scenario includes changes in typhoon intensity, sea-level height, and the construction intensity of dams along the Qingdao coast.

For CLA: Coastal protection includes the management level of coastal engineering and flooding protection measures. In this scenario, the management and control capabilities of coastal protection should be maintained as in recent years, such as storm surge protection standards (50- or 100-year return period). New dams will not be built, but the existing ones will be kept in good condition.

For CA: The continuation level focuses on conducting adaptation measures over a period of time. The coastal protection measures should be strengthened or promoted continuously according to the Urban Master Plan in Qingdao (2018–2035). The CA adaptation measures have a high adaptation scenario level to respond to coastal flooding caused by future climate change in Qingdao. New dams are established on other coastlines based on the existing dams. All areas planned for construction land in the coastal zone (except the shoreline from the Yang River to the Dagu River) should build dams to defend against coastal flooding according to the direction of urban planning development before 2035.

For EA: Compared with the CA strategy, the enhanced level of adaptation strategy will continuously improve, especially in coastal protection. The EA measures aim to further reduce coastal flooding areas by strengthening the protection and management level of the coastal zone. The flood control and storm surge dam prevention standards should be high. Therefore, the construction standards of the dam are raised, and new dams are built on the north bank of the Dagu River according to the direction of the urban plan and flooding area distribution.

4.2 Typhoon storm surge in the present day

Under the current situation, this section mainly introduces the differences between storm surge elevation and inundation caused by typhoons with different paths along the coast of Qingdao. The sea level and intensity of typhoons are actual scenarios in the current scenario.

4.2.1 Surge in the present day

The ADCIRC model is driven by the eight typical present-day typhoons and sea-level conditions in this section. Figure 6 shows the distribution of the maximum storm surge caused by the eight types of typhoons.

Fig. 6
figure 6

Distribution of the maximum storm surge elevations in the present day. The black line refers to the coastline of Qingdao, and the white area refers to the land

Full size image

Figure 6 shows that the eight types of typhoons have different effects on the storm surge in Qingdao. The distribution of surges in the coastal areas of Qingdao is significantly different. The offshore surge is small, the nearshore surge is large, and the distribution of surge in Jiaozhou Bay markedly varies under different typhoons. The results indicate that the storm surge elevation caused by typhoon G is the largest among the eight typhoons, and the storm surge elevation caused by typhoon D is the smallest.

The average value of the maximum surge caused by eight different types of typhoons along the coast of Qingdao is illustrated in Fig. 7 to further explore the impact of the typhoon surge on the Qingdao area. The west surge elevation of Jiaozhou Bay is generally higher than its east surge elevation, with an average maximum water increase of over 1 m.

Fig. 7
figure 7

Average distribution of the maximum surge elevations caused by eight typhoons along the Qingdao coast in the present day

Full size image

Storm surge intensity caused by eight different typhoon types along the Qingdao coast is effectively represented by introducing a variable of surge duration, where the minimum storm surge class (greater than 0.5 m) is used as the criterion. The surge duration under this criterion is calculated for each storm surge type. The mean surge duration (MSD) and the mean maximum surge (MMS) along the coast are shown in Fig. 8.

Fig. 8
figure 8

Mean surge duration and maximum surge along Qingdao coastline induced by eight typical typhoon processes in the present day

Full size image

Figure 8a shows the MSD of eight typical tracks along the Qingdao coast. The figure reveals that the MSD of eight typical tracks in Qingdao is 10 h on average; among which, the largest MSD is Track G, approximately 25 h, and the relatively small MSD is Tracks D and H, approximately 1 h. Figure 8b shows the MMS of the typhoon storm surge for eight typical tracks along the Qingdao coast. The MMS for the eight typical tracks is 0.84 m on average, demonstrating the highest MMS caused by Track G of approximately 1.35 m and the lowest MMS caused by Tracks D and H of approximately 0.42 and 0.47 m, respectively. Overall, Track G, which has the longest MSD and the largest MMS, is the most severe storm surge under the present-day scenario. Meanwhile, Tracks D and H generate low hazards in all eight types of typhoons.

4.2.2 Inundation in the present day

Furthermore, the floodplain under the influence of eight typhoons is analyzed to investigate the influence of the storm surge floodplain on Qingdao.

The distribution of the inundation caused by eight typical typhoon tracks in the real scenarios is displayed in Fig. 9. Notably, the tidal flat is defined as land due to the setting of the coastline of Qingdao. In tidal flat areas, under the influence of tides, these areas will be submerged due to tidal fluctuations in the absence of a storm surge. The tidal phase also affects the depth of inundation and the inundation area. When calculating the inundation area of storm surge, the area flooded by tide is excluded, and the remaining area is considered the inundation area of typhoons to accurately reflect the flooding caused by storm surge.

Fig. 9
figure 9

Distribution of the maximum water level elevations. The black line refers to the coastline of Qingdao, and the white area refers to the land

Full size image

Figure 9 shows a difference between the inundation area and inundated depth caused by typhoons with different intensities and tracks. The inundated areas are mainly concentrated in Jiaozhou and Haixi Bays, among which the water depth in the area at the top of Jiaozhou Bay is the largest. The maximum flooding depth in the western part of Jiaozhou Bay is generally greater than that in the eastern part of the inundation caused by eight typhoons. Figure 9 shows that the water flooding the eastern part of Jiaozhou Bay is from the northwestern coast. Seawater will overflow into the eastern region when the flooded water reaches a certain height in the northwestern region of Jiaozhou Bay. The figure reveals that the inundation site often concentrates at the head of the bay, and then the inundated area is expanded from the coastline to the inland. The range of inundation caused by eight typical typhoon tracks is from 3.4 km2 to 58.6 km2, among which Track G is the most significant to Qingdao.

4.3 Typhoon storm surge in the future

Under future scenarios, this section mainly discusses the differences in surges and inundations caused by typhoons with different tracks along the coast of Qingdao. For brevity, only the results of the 2055 scenario are comprehensively discussed.

4.3.1 Surge in the future

The maximum value of the storm surge of eight typical future typhoons in the 2055 scenario is shown in Fig. 10.

Fig. 10
figure 10

Distribution of the maximum storm surge elevations in 2055. The black line refers to the coastline of Qingdao, and the white area refers to the land

Full size image

These results indicate that the eight types of typhoons have different effects on the storm surge in Qingdao. In the future scenario, the storm surge distribution in the coastal areas of Qingdao will be significantly different. The nearshore surge is typically larger than the offshore surge, which is consistent with the current scenario. Jiaozhou Bay is particularly complicated in terms of surge distribution, with strong local effects. The maximum storm surge value of eight typical future scenario typhoons in the bay is mainly distributed within a 0.5–2.0 m range. The maximum value is mainly distributed near the coastal area of Xihaianxinqu and north of Jiaozhou Bay.

The relative value of MSD and MMS for the storm surge of eight typical future scenario typhoons is shown in Fig. 11. The MSD and MMS of eight typical future scenario typhoons are gradually increasing from the present to 2055.

Fig. 11
figure 11

Mean surge duration (MSD) and mean maximum surge (MMS) along Qingdao coastline induced by eight typical typhoon processes in the future

Full size image

In the 2055 scenario, the longest MSD occurs in Track E, which takes approximately 28 h, and the shortest in Track C, which takes approximately 6 h; the highest MMS occurs in Track H at approximately 1.42 m, and the smallest occurs in G at approximately 0.76 m. A comparison of the present and 2055 data reveals that the MSD increases by 4.74 times on average, and the MMS rises along the coast by 47%. For surge duration, Track H is the fastest-growing process, demonstrating a growth of 20 times, and Track G is the slowest, with a growth of − 56%. For MMS, Track H is the fastest-growing process, with a growth of 2 times, and Track G is the slowest, with a growth rate of − 44%. Under the future scenario, the MSD and the MMS increase 4 times and 37% in 2035 and 5 times and 47% in 2055 compared with the present day, respectively.

4.3.2 Inundation in the future

The distribution of the inundation caused by eight typical typhoon tracks in the future scenarios is displayed in Fig. 12. As previously indicated, the inundation area under the influence of tide is excluded when calculating the inundation area caused by typhoon storm surge to reflect the flooding due to storm surge accurately.

Fig. 12
figure 12

Distribution of the maximum water level elevations in 2055. The black line refers to the coastline of Qingdao, and the white area refers to the land

Full size image

As shown in Fig. 12, typhoons with different intensities and tracks cause differences in the submerged area and depth. Similar to the current scenario (Fig. 8), the submerged areas are mainly concentrated in Jiaozhou and Haixi Bays, with the deepest water at the top of Jiaozhou Bay. This phenomenon is the same dynamic process of submerging Jiaozhou Bay in the present-day scenario, and the seawater flows into the eastern area after the western part is submerged. Future scenarios show a significant increase in inundated areas in comparison with the present-day scenario. The most significant process is B, which has expanded the flooded area by 40.7 km2.

4.4 Effects of the adaptation measures

In the future, SLR and typhoon intensity will increase the damage from storm surges. Improving the defense level of coastal dams can effectively reduce storm surge disasters, casualties, and losses. Thus, considering the development and changes in storm surge disasters in future scenarios and the future coastal defense level planning of Qingdao, the response of the Qingdao area to different defense measures is examined.

The hydrodynamic model ADCIRC is used to calculate the inundation process of typhoon storm surges based on a future scenario. Notably, this section focuses on the response of Qingdao to different defense measures in the future scenario. The existing dam level (Sect.  4.4.1) is the control group, while that under the CA scenario (Sect.  4.4.2) and the EA scenario (Sect.  4.4.3) is the experimental group.

4.4.1 Inundation analysis with the CLA strategy

This section first studies the overall change in Qingdao storm surge inundation and then analyzes the distribution of the inundation range from typical cases to demonstrate the impact of storm surge on Qingdao under CLA conditions.

Figure 13 shows the inundation area in the future. The figure reveals that the inundation area of eight typical typhoons gradually increases in future scenarios. However, the extent of the increase in the area varies depending on the typhoon type. The inundation range caused by eight typical typhoon tracks is 16.2 km2 to 81.9 km2 from 2035 to 2055. On average, the maximum value of inundation will increase by 132% and 174% in 2035 and 2055, respectively, compared with the present-day results.

Fig. 13
figure 13

Inundation area during eight future typhoons in Qingdao with CLA adaptation

Full size image

Furthermore, the increase in inundation area in 2055 is − 35% to 643% compared with the present, and the increase in the area ranges from − 20.5 km2 to 40.7 km2 for different types of typhoons. The result shows that the inundation induced by storm surges will have an increasing trend in the future. Overall, under the CLA adaptation, the risk of storm surge inundation continues to increase as sea level rises and typhoon intensity changes.

The distribution of Qingdao floodplain inundation is studied from a typical perspective. This distribution primarily selects the changes in the Qingdao floodplain range caused by two typical typhoon tracks in 2035 and 2055. Figure 14 shows the distribution of the inundation area, with the most significant increase in the submerged area induced by Track B and the least significant increase in the submerged area induced by Track F from 2035 to 2055.

Fig. 14
figure 14

Inundation distribution during two future typhoons in Qingdao with CLA adaptation. Left: 2035, right: 2055. Upper: Track B, lower: Track F

Full size image

Figure 14 shows the distribution of the inundation area, with the most significant increase in the submerged area induced by Track B and the least significant increase in the submerged area induced by Track F from 2035 to 2055. The inundation area of Tracks B and F in 2035 is 51.2 and 20.0 km2, respectively. Meanwhile, the inundation area of Track B increases by approximately 30.7 km2, and that of Track F increases by approximately 0.3 km2 in 2055. Jiaozhou, Chengyang, and Huangdao are the main coastal flooding areas in Qingdao. The Jimo area, which is also in the northern part of Aoshan Bay, is inundated. The inundation flooding distribution caused by typhoons on different tracks is the same in Huangdao but remarkably different in Jiaozhou and Chengyang. In 2055, for the Track B typhoon, the impact on Chengyang and Jiaozhou was far greater than other typhoons. The area of Chengyang inundation could reach the central part of Chengyang, and the impact on Jiaozhou continued to reach inland Jiaozhou. Meanwhile, the inundation area of the construction land and the water contributed to the most economic loss in each case.

4.4.2 Inundation analysis with CA strategy

Previous results show that inundation occurred in the coastal areas of Qingdao with CLA adaptation. The inundation areas in Qingdao are small, but flooding will still cause disaster losses. According to the 2035 Plan from the Qingdao Municipal Government, dams are established in other flooded areas in Qingdao other than the top of Jiaozhou Bay. The hydrodynamic model ADCIRC is used to calculate the inundation process of typhoon storm surges based on a future scenario. Figure 15 shows the future storm surge inundation area statistical distribution with CA adaptation during eight typical typhoon tracks of Qingdao caused by SLR in the years 2035 and 2055.

Fig. 15
figure 15

Inundation area during eight typical typhoon tracks of Qingdao with CA adaptation

Full size image

Figure 15 shows the future inundation area with CA adaptation. The figure reveals that the inundation area of eight typical typhoons gradually increases in future scenarios. However, the extent of the increase in the area varies depending on the typhoon type. The inundation distribution under CA adaptation is similar to that under CLA adaptation, but the inundation area is less than the latter. In the CA scenario, the range of inundation area caused by eight typical typhoon tracks is 15.2 to 77.8 km2. Compared with CLA adaptation, the inundation area of CA adaptation is reduced by 2%–22% in the future.

Figure 16 shows two types of typhoons that cause the largest increase in flooded areas (Track B) at approximately 77.8 km2 and the least increase in flooded areas (Track F) at approximately 17.6 km2. This finding indicates the presence of additional spatial distribution characteristics in different land-use types. In the CA adaptation strategies, new dams should be built in accordance with the direction of urban planning development before 2035. The inundation areas of different land-use types in various scenarios are shown in Fig. 16. The main coastal flooding areas (construction land and water) are concentrated in Jiaozhou Bay, and Jiaozhou and Chengyang may have the largest economic losses. That is to say, if new dams are built based on existing dams, then most of the coastal zone areas will not be flooded. Therefore, reducing the coastal risks in Qingdao is an effective measure.

Fig. 16
figure 16

Inundation distribution during two typical typhoon tracks of Qingdao caused by sea-level rise in the years 2035 and 2055 with CA adaptation

Full size image

4.5 Inundation analysis with an EA strategy

Previous results show that the top of Jiaozhou Bay was flooded by a storm surge with CA adaptation. The risk of coastal flooding should have a high reduction level for effective coastal protection and risk management. Dams are added around the northwest of Jiaozhou Bay in the model based on the CA adaptation. The ADCIRC hydrodynamic model was used to simulate the storm surge inundation to provide suggestions for the Qingdao government.

Figure 17 shows the future statistical distribution of the inundation area with EA adaptation. The figure reveals that the inundation area of eight typical typhoons also gradually increases in future scenarios. However, the extent of the increase in the area varies depending on the typhoon type. The inundation distribution under EA adaptation is similar to that under CA adaptation, but the inundation area is less than the latter. In the EA scenario, the range of inundation area caused by eight typical typhoon tracks is from 0.5 to 44.9 km2. Compared with CA adaptation, the inundation area of EA adaptation is reduced by 28%–97% in future scenarios. Meanwhile, compared with CLA adaptation, the inundation area of EA adaptation is reduced by 32%–98%.

Fig. 17
figure 17

Inundation area during eight typical typhoon tracks of Qingdao with EA adaptation in the future

Full size image

Figure 18 shows that the Jiaozhou area has not been flooded in the EA adaptation, but the inundation in the Chengyang area has not changed compared to that in the CA and CLA adaptations. Adding the dams in the northwest part of Jiaozhou Bay has effectively reduced the storm surge flooding in Jiaozhou. After an enhanced adaptation measure, only Chengyang will have a few construction lands and water areas flooded in the future. Among eight typical typhoon tracks, Track B (Figs. 18a,b) causes the most serious inundation of approximately 44.9 km2, and Track D (Figs. 18c,d) causes the least inundation of approximately 0.7 km2 in future scenarios. EA adaptation demonstrates the most significant inundation reduction, which can minimize inundation on average by more than 50%.

Fig. 18
figure 18

Inundation distribution during two typical typhoon tracks of Qingdao in the future scenario with EA adaptation

Full size image

5 Conclusions

An unstructured grid hydrodynamic model was developed in this paper for Qingdao, China, using the ADCIRC model to study the impact of climate change on storm surges and inundations induced by eight typical typhoon tracks in current and future scenarios. The eight typical typhoon tracks were chosen from 71 tropical cyclones affecting Qingdao according to the typhoon track and intensity, and they are denoted as Tracks A to H. The model surge results induced by typhoons were validated with the available observations.

For storm surges under present-day and future scenarios, the storm surge caused by typhoons with different intensities and tracks has various effects on the Qingdao area. The overall feature is a small offshore surge and a large nearshore increase. Meanwhile, the surge distribution along Jiaozhou Bay is particularly complicated, with strong local effects. For most storm surge processes, the surge height in western Jiaozhou Bay is greater than in eastern Jiaozhou Bay. Track G, which has the longest surge duration (25 h) and the largest average maximum surge (1.35 m), is the most severe storm surge under the present-day scenario. Under the future scenario, the MSD and the MMS will increase 4 and 5 times and 38% and 47% in 2035 and 2055, respectively.

Under present-day scenarios, the inundation area and depth induced by typhoons with various intensities and tracks differ. Jiaozhou and Haixi Bays are the most heavily submerged areas around the Qingdao coastline, among which the inundation depth in the area at the top of Jiaozhou Bay is the largest. The maximum inundation depth in the western part of Jiaozhou Bay is generally greater than that in the eastern part of Jiaozhou Bay due to eight typhoons. More importantly, the inundation site often concentrates at the head of the bay, and then the submerged area is expanded from the coastline to the inland.

The following three adaptation measures are set for inundation under future scenarios: CLA, CA, and EA. In the CLA adaptation, eight typical typhoon tracks induced an inundation range of 16.2 to 81.9 km2. On average, the maximum value of inundation will increase by 132% and 174% in 2035 and 2055, respectively. In the CA adaptation, the dam was established except for the coastline from the Yang River to the Dagu River, which blocked part of the seawater intrusion. Overall, CA adaptation will reduce the inundation area by more than 12% compared with CLA adaptation. The EA adaptation adds dikes on the north bank of the Dagu River based on the CA scenario. Hence, the EA scenario shows the most significant inundation reduction, which can minimize inundation by more than 50%.