Characteristics and Influencing Factors of Storm Surge-Induced Salinity Augmentation in the Pearl River Estuary, South China

Abstract

The Pearl River Estuary (PRE) frequently experiences the impacts of typhoons, storm surges, and saltwater intrusion. While previous research has mainly focused on saltwater intrusion during the dry season, there is limited research on saltwater intrusion caused by storm surges in the PRE. In this study, we systematically investigate the effects of ten typical autumnal typhoons and associated storm surges on saltwater intrusion in the Modaomen Waterway using in situ data of water level, river discharge, and chloride concentrations from 2006 to 2022. We introduce the concept of Storm surge-Induced Salinity Augmentation (SISA) and analyze its characteristics and primary influencing factors. Our findings reveal that SISA primarily occurs in autumn, with reduced upstream river discharge and the dominance of high-salinity water in the estuary. SISA occurs immediately after storm surges and grows rapidly and violently, with a time lag of 2–4 h, but rapidly recedes after the typhoon passage due to heavy rainfall and high freshwater discharge. Typhoons with a westward trajectory have a greater influence, and the southeastern winds outside the estuary during typhoon events are the primary factors determining the intensity of SISA. Pre-typhoon river discharge affects the range and duration of saltwater intrusion. Moreover, the coupling effect of extreme river dryness, spring tide, and storm surges significantly enhances saltwater intrusion. Further research is needed to quantify the spatiotemporal characteristics of SISA accurately.

1. Introduction

Saltwater intrusion, the intrusion of brackish or saline water into river channels or groundwater, leading to increased salinity and chloride levels upstream [1], is a common occurrence in the Pearl River Delta (PRD) during the dry season. Various factors, including river discharge, tidal dynamics, local geomorphology, sea level rise, meteorological conditions, and human activities, influence this phenomenon [2,3,4,5]. The Modaomen Outlet, the largest one among the eight outlets in the PRD, contributes over a quarter of the total runoff into the estuary. Saltwater intrusion during the dry season significantly affects the Modaomen Waterway, which connects the West River and the South China Sea. When the chloride concentrations in the river exceed the drinking water standard of 250 mg/L, the water withdrawal from the Modaomen Waterway is affected and even stopped, leading to inadequate water supply for cities like Zhongshan, Zhuhai, and Macao [6,7,8]. Ecologically, saltwater intrusion in the Pearl River Estuary (PRE) can harm intertidal wetland habitats and disrupt coastal and estuarine species, including plankton [9,10,11,12].

Saltwater intrusion in the Modaomen Waterway exhibits distinct diurnal (small-scale), semimonthly (medium-scale), and seasonal (large-scale) characteristics due to the influence of asymmetric semidiurnal tides, astronomical tides, and seasonal variations in freshwater discharge [13]. On a daily scale, salinity increases during the flood tide and decreases during the ebb tide, with a lag of approximately 1–2 h behind water level changes [14]. On the fortnight scale, saltwater intrusion intensifies during neap tides and weakens during spring tides. It begins to strengthen 1–2 days before neap tides, reaches its peak during the transition from neap tide to spring tide, and weakens during the transition from spring tide to neap tide [15,16]. On a seasonal scale, high river discharge during the flood season suppresses saltwater intrusion, while reduced river discharge during the dry season amplifies it [17]. The interaction between river flow and estuarine exchange flow causes high-salinity bottom water to move landward while surface freshwater flows seaward, resulting in stratification in the Modaomen Outlet [14,15,18]. Wind stress and ocean waves can alter the stratification of saltwater and freshwater in the Modaomen Waterway, affecting the spatiotemporal distribution of salinity [19,20,21].

The PRD is susceptible to storm surges caused by typhoons, which significantly impact saltwater intrusion in the PRE [22,23,24]. Both in situ data and modeling results have shown that Typhoon Hagupit in 2008 caused a substantial increase in peak water levels in the Modaomen Waterway and intensified saltwater intrusion [25]. During typhoon events, strong wave-current interactions enhance vertical mixing in the estuarine region, increasing landward advection and tidal oscillatory salt flux. Storm surges and remote winds elevate the estuarine water level, extending the range and intensity of saltwater intrusion [19,22]. In 2011, during Typhoon Nesat (occurring on 29 September, the 3rd day of September in the Lunar calendar, spring tide), severe saltwater intrusion occurred in the Modaomen Waterway with a river discharge of only 1500 m³/s at Wuzhou, a representative hydrological station for the West River. This was caused by the combination of an extremely dry season, typhoons, and astronomical tides, resulting in a sharp increase in the range, duration, and intensity of saltwater intrusion [26]. Therefore, when typhoons pass through, the combined influence of increased water levels, landward advection, and tidal oscillation leads to a noticeable escalation in the duration, intensity, and range of saltwater intrusion. We refer to this phenomenon as Storm surge-Induced Salinity Augmentation (SISA), which is defined as the difference between the highest chloride concentration in the estuary during typhoon transit and the chloride concentration one or two days before typhoon landfall (when the impact of storm surge is minimal) at similar tidal phases.

In the Western North Pacific and the PRE, the peak season for typhoon generation and landfall occurs from July to September, coinciding with the wet season in the Pearl River Basin. During this period, the PRE experiences high freshwater runoff, and the impact of storm surges on saltwater intrusion has received limited attention from both industry and researchers. Over the past two decades, reclamation and waterway dredging in the Modaomen Outlet have altered the underwater topography, extended and deepened the channel, and reduced the capacity of estuarine shoals to store freshwater and counteract saltwater intrusion [4,27,28]. Simultaneously, storage operations of the upstream large reservoirs in the West River and North River have gradually reduced freshwater flow at the end of the flood season, resulting in an earlier occurrence of the dry season and saltwater intrusion overlapping with the typhoon season [26]. Additionally, the rapid development of urban areas in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) has increased the demand for water resources, exacerbating the impact of saltwater intrusion on cities along the Modaomen Waterway, such as Zhongshan, Zhuhai, and Macao [8,29]. Despite extensive research efforts by many scholars on the patterns and dynamic mechanisms of dry-season saltwater intrusion in the PRE, systematic investigations into the effects of storm surge-induced saltwater intrusion are notably lacking [24,25].

This study utilizes in situ salinity and hydrological data spanning from 2006 to 2022 along the Modaomen Waterway and West River. Meanwhile, we select ten representative typhoons that significantly impacted saltwater intrusion in the Modaomen Waterway. By integrating wind field and water level data, we systematically analyze the characteristics and key influencing factors of SISA in the Modaomen Waterway. The results of this study aim to enhance our understanding of the combined effects of typhoons on saltwater intrusion, improve the accuracy of saltwater intrusion forecasting, and optimize water resource management in the industry.

2. Materials and Methods

2.1. Study Area

The Pearl River, situated in south China, ranks second in annual runoff among Chinese rivers, mainly consists of the West River, North River and East River, and forms the Pearl River Delta (PRD) in the estuary plain (Figure 1). The PRD includes the northwestern and eastern river deltas. The northwestern river delta is formed by the West River, North River, Tan River, and their tributaries, while the eastern river delta is formed by the East River and its tributaries. The Pearl River Estuary (PRE) encompasses the Lingdingyang Bay, Huangmaohai Bay, the river network and eight outlets in the PRD, i.e., Yamen, Hutiaomen, Jitimen, Modaomen, Humen, Jiaomen, Hongqimen, and Hengmen.

The Modaomen Waterway connects the West River and pours out to the South China Sea via the Modaomen Outlet, the largest one of the eight major outlets in the PRE (Figure 1). Historical records from 1959–1998 indicated that the mean annual river discharge at the Denglongshan (DLS) Station was 88.4 km³, which constituted 28.2% of the total Pearl River discharge [30]. The average annual runoff at the Wuzhou (WZ) Station, located in the middle of the West River, was 6018 m³/s. During the flood season (April to September), the average flow was 8675 m³/s; in the dry season (the rest of the year), the average flow decreased to 3362 m³/s. The Modaomen Waterway spans approximately 45 km from the upstream Renyi (RY) Water Plant to the downstream Guangchang (GC) Water Pump. It has a width of around 1–2 km and a mean channel depth of 10–15 m. The Dachongkou (DCK) Station is situated 1.5 km north of the GC Station, while the Pinggang (PG) and Zhuzhoutou (ZZT) stations are positioned in the middle section of the Modaomen Waterway, 18 km and 25 km away from the GC Station, respectively. The channel extends approximately 18 km from the GC Station to the southeast corner of Hengqin Island, located at the estuary (Figure 1).

The Modaomen Outlet experiences a mixed semidiurnal tidal cycle, with two high tides and two low tides of different tidal ranges occurring each lunar day. The tidal range decreases landward, with values diminishing from 1.11 m at the Sanzao (SZ) Station to 0.86 m at the DLS Station [15]. Positioned in an area with a subtropical monsoon climate near the Tropic of Cancer, the PRE is frequently threatened by typhoons, with an average of 5.7 typhoons annually [26].

Due to saltwater intrusion, the chloride concentration exceeded 250 mg/L for 182 days at GC and 65 days at PG, significantly affecting the water supply to nearby cities, such as Zhongshan, Zhuhai, and Macao.

2.2. Data

This study compiled in situ data on chlorides, water level and river discharge, upstream watershed precipitation, tropical cyclone tracks, and wind speed data from 2006 to 2022 (

Table 1. Description of relative parameter and data source.

NO.	Parameter	Description	Source
1	Precipitation	Daily precipitation (mm) from GPM IMERG	NASA
2	Water level 1	From GC station (m), hourly recorded	ZWEHG, PRWRC
3	River discharge	From WZ station (m3/s), daily recorded	ZWEHG
4	Chloride 1	From GC, PG, and ZZT stations (mg/L), hourly recorded	ZWEHG
5	Chloride 2 and water level 2	From RY to DCK stations (mg/L), recorded every 5 min	ZWA
6	Typhoon track	CMA Tropical Cyclone Best Track Dataset, recorded every 6 h	CMA
7	Wind speed	ERA5 reanalysis data recorded every 1 h	ECMWF

). Data on water levels and chlorides were sourced from Zhuhai Water Environment Holdings Group LTD. (ZWEHG) and recorded hourly. The river discharge at the WZ Station was obtained from both the Pearl River Water Resources Commission (PRWRC) and ZWEHG, and they represent the daily average discharge. Additional water level and chloride data were acquired from the Zhongshan Water Authority (ZWA) and recorded at a five-minute interval.

Precipitation data were obtained from the Global Precipitation Measurement Mission (GPM). Daily cumulative precipitation (the final run of IMERG) was selected for the analysis. The track data of tropical cyclones were sourced from the China Meteorological Agency (CMA) Tropical Cyclone Best Track Dataset, which recorded information at six-hour intervals, occasionally three-hour intervals [31,32]. Wind speed data were extracted from the ECMWF Reanalysis v5 (ERA5) dataset, providing hourly records.

2.3. Typhoons

Based on in situ data of chloride concentration from various stations, including the GC station in the Modaomen Waterway, and in situ data of river discharge from the WZ station in the West River mainstream, covering the period from 2006 to 2022, a comprehensive analysis was conducted to examine the characteristics of saltwater intrusion in the Modaomen Waterway under the combined influence of river discharge and tidal dynamics. This study specifically focuses on ten autumn typhoons that had a significant impact on saltwater intrusion in the Modaomen Waterway. The selected typhoons are Typhoon Hagupit (2008), Typhoon Mujigae and Typhoon Koppu (2009), Typhoon Nesat and Typhoon Nalgae (2011), Typhoon Kalmaegi (2014), Typhoon Mangkhut (2018), Typhoon Saudel (2020), Typhoon Kompasu (2021), and Typhoon Nesat (2022) (

Table 2. Information and measurements during the landfall/transit of the ten typical typhoons, storm surges, and SISA at the GC station (N—Neap tidal period, S—Spring tidal period, M—Meso tidal period).

Name and Landfall Date of Typhoons	HW 1 (m/s)	LP 2 (hPa)	Storm surge (m)	PWL 3 (m)	River Discharge 4 (m3/s)	Distance 5 (km)	SISA (mg/L)	Tide Phase 6
Hagupit (2008) 24 September 2008	50	940	1.77	3.20	5000	127	9590	N, high tide
Mujigae (2009) 10 September 2009	20	992	0.25	1.83	2817	301	2048	S to M, high tide
Koppu (2009) 15 September 2009	40	960	1.55	3.13	2515	65	10,190	N to M, high tide
Nesat (2011) 29 September 2011	40	960	0.68	2.45	1470	299	6000	S, high tide
Nalgae (2011) 3 October 2011	30	980	0.37	2.50	1690	467	800	S to M, low tide
Kalmaegi (2014) 16 September 2014	40	960	1.15	2.66	4830	315	4600	S to M, high tide
Mangkhut (2018) 16 September 2018	48	950	1.64	2.95	4910	51	6190	M to N, low tide
Saudel (2020) 23 October 2020	30	980	0.11	1.78	6670	460	3400	M to N, high tide
Kompasu (2021) 13 October 2021	33	970	0.62	2.10	3220	325	5700	M to N, flood tide
Nesat (2022) 18 October 2022	42	955	0.39	2.05	1780	381	4834	M to N, flood tide

¹ Highest wind speed (Two-minute mean maximum sustained wind); ² Lowest pressure (Minimum pressure near the tropical cyclone center); ³ Peak water level; ⁴ River discharge at WZ in the 2 days preceding landfall; ⁵ Minimum distance from the Modaomen Outlet; ⁶ Phase of the tide during landfall/transit of the typhoon.

, Figure 2 and Figure 3). All ten typhoons followed a common westward trajectory, with seven of them making landfall not directly on the Guangdong coast but rather in the more southern regions of Hainan and Vietnam (Figure 2). Except for Typhoon Nalgae in 2011, the remaining nine typhoons caused a significant increase in chloride concentration, exceeding 2000 mg/L at the GC station in the Modaomen Waterway, indicating a pronounced salinity augmentation effect (Table 2, Figure 2 and Figure 3).

Among these typhoons, Hagupit, Koppu, and Mangkhut exhibited the most substantial salinity augmentation. These three typhoons made landfall in the Dianbai District, Taishan City, and Taishan City in Guangdong Province on 24 September 2008, 15 September 2009, and 16 September 2018, respectively. They all had low central pressure, high wind speeds, and a closer track to the Modaomen Outlet (127, 65, and 51 km), resulting in significant storm surges (1.55–1.77 m) and high total water levels (2.95–3.20 m), and were accompanied by rapid augmentations in chloride concentrations (9590, 10,190, and 6190 mg/L). Due to being in the flood season before the passage of the typhoons, the WZ river discharge was 5000, 2515, and 4910 m³/s for Typhoons Hagupit, Koppu, and Mangkhut, respectively. The typhoon-induced rainfall further contributed to an increase in runoff, resulting in short durations and a limited intrusion range of saltwater intrusion.

Five typhoons passed through the study area in the autumn around October. Except for Typhoon Saudel in 2020, the river discharge before the typhoon passage was relatively low (1470–3220 m³/s). Compared to Typhoons Hagupit and Mangkhut, these typhoons had relatively higher central pressures, lower wind speeds, were farther away from the Modaomen Outlet, and had smaller storm surges, but were influenced by smaller river discharge from the West River, the typhoon’s path, and the fortnight tidal cycle. This resulted in a longer duration of saltwater intrusion, which aligns with the saltwater intrusion features of the Modaomen Waterway. Specifically, the salt transport is primarily influenced by estuarine exchange flow, which strengthens during neap tide. Under the influence of similar freshwater flow from the West River, the most severe saltwater intrusion occurs during the transition from neap tide to spring tide, while the relatively weaker phase occurs during the transition from spring tide to neap tide [15].

2.4. Calculation of Typhoon Rainfall-Induced Runoff

During the passage of a typhoon, the impact of typhoon rainfall on the upstream area of the WZ station can be represented by the river discharge at WZ. However, the influence of typhoon rainfall on the downstream area of WZ and the North River basin is more significant. Therefore, relying solely on the river discharge at WZ after the typhoon’s landfall may result in an underestimation of the river discharge in the Modaomen Waterway. To examine the inhibitory effect of typhoon rainfall-induced runoff on saltwater intrusion, this study utilizes daily precipitation of GPM IMERG to calculate daily rainfall-induced runoff from the downstream of WZ to the Modaomen Outlet and the entire North River basin. Meanwhile, the river discharge at WZ is also considered. This combined value represents the river discharge at the Modaomen Outlet after the typhoon’s landfall. The total area of the basin for production and convergence below WZ is 74,660 km², including 17,960 km² for the West River basin below WZ and 56,700 km² for the North River basin (Figure 1a).

The Rainfall–Runoff Coefficient Method is employed to rapidly calculate the single-day rainfall-induced runoff within the basin area, as defined by Equation (1) [33,34]:

$${∆Q}_P=\frac{\sum _{i=1}^m\sum _{j=1}^n{K_{ij}P}_{ij}{S}_{ij}}{T}$$

(1)

where ${∆Q}_P$ represents the daily rainfall-induced runoff within the basin (m³/s). $P_{ij}$ denotes the single-grid rainfall precipitation. $S_{ij}$ represents the single-grid area, depending on the resolution of the precipitation data. $T$ is time, indicating the period of the precipitation accumulation. $K_{ij}$ signifies the grid rainfall–runoff coefficient determined based on land use types. For example, water bodies, urban land, and forest land have rainfall–runoff coefficients of 1, 0.9, and 0.5, respectively.

Based on ${∆Q}_P$, the total runoff of the basin is further calculated using Equation (2):

$$Q={∆Q}_P+Q_{wz}$$

(2)

where $Q_{wz}$ is the river discharge at WZ for the previous day, and $Q$ represents the total river discharge of the basin and approximately one-third of $Q$ discharges through the Modaomen Outlet.

3. Results

3.1. Impact Process Analysis of Storm Surge-Induced Salinity Augmentation

During the passage of typhoons, such as Typhoon Koppu, in 2009, the downstream stations of Modaomen Waterway, such as GC or DCK, were rapidly affected by storm surge-induced augmentations in surface chloride concentration (Figure 4). The trend of chloride concentration change aligned with the water level variations, reaching its peak 2–4 h after the highest water level and rapidly diminished with decreasing water levels. The duration and recession rate of chloride concentration after each typhoon varied due to factors such as the typhoon’s movement path, river discharge, and tidal phase (Figure 4). In the figure, only the impact of Typhoon Koppu was considered for the two consecutive typhoons in 2009, Mujigae and Koppu, as Mujigae was distant from the PRE, exhibited lower maximum wind speed, and caused minimal impact. Similarly, for the two consecutive typhoons, Nesat and Nalgae, in 2011, only the impact of Nesat was assessed in the figure, as it brought substantial rainfall, leading to a significant increase in basin river flow, while Nalgae was distant from the Modaomen River Outlet and had a negligible impact on salinity augmentation.

During the passage of Typhoon Hagupit on 23 September 2008, Typhoon Kalmaegi on 16 September 2014, and Typhoon Mangkhut on 16 September 2018, which coincided with the end of the flood season, the river flow at WZ in the 2 days before the typhoon had relatively high values (4830 to 5000 m³/s). The chloride concentration at GC was below 250 mg/L, the minimum threshold for saltwater intrusion detection. During the passage of these three typhoons, the storm surge at GC reached as high as 1.15 to 1.77 m, resulting in a maximum water level of 2.66 to 3.20 m and salinity augmentations of 4600 to 9590 mg/L (Table 2). In a similar flood season scenario, Typhoon Koppu, on 15 September 2009, had a lower upstream river discharge at WZ of 2515 m³/s but still experienced a maximum storm surge of 1.55 m and a significant salinity augmentation by 10,190 mg/L. The temporal shape of the chloride concentration curve was similar to the water level curve during the storm surge process and lagged by 2–4 h. After the storm surge receded, coupled with the sharp increase in freshwater runoff from the West River due to typhoon rainfall, the chloride concentration at GC rapidly decreased and returned to below 250 mg/L within 1–2 days (Figure 4a,b,d,e). Compared to Hagupit and Kalmaegi, Typhoon Mangkhut resulted in a more prolonged duration of high chloride concentration at GC (Figure 4a,d,e). This was primarily because the Mangkhut was closer to the Modaomen Outlet, leading to a greater wind and storm surge impact on saltwater intrusion (Figure 2 and Figure 3g). Typhoon Mujigae passing through the PRE five days before Typhoon Koppu in 2009, due to its greater distance from the PRE and lower wind speed, resulted in weaker storm surge and saltwater intrusion effects compared to the subsequent passage of Typhoon Koppu (Table 2).

In the case of Typhoons Nesat and Nalgae from 29 September to 3 October 2011, both passed by the PRE and subsequently landed in the northeast and southwest of Hainan, with the shortest distance to the Modaomen Outlet by 299 km and 467 km, respectively (Figure 2). Two days before the landfall of Typhoon Nesat, the river flow at WZ was only 1470 m³/s, categorized as an extremely low-flow period. During the spring tide period when the typhoon passed, severe saltwater intrusion had already occurred in the PRE, with the chloride concentration at GC reaching 6500 mg/L. The storm surge mixing with the tidal flow and the storm surge-induced water level rise (0.68 m, peak water level 2.45 m) significantly contributed to high saltwater intrusion in the Modaomen Waterway. The peak chlorinity increase at GC was 6000 mg/L, resulting in a peak chloride concentration of 12,300 mg/L. The convergence of extremely low river discharge, spring tide, and storm surge led to a prolonged and severe saltwater intrusion in the PRE, which only retreated after the typhoon rainfall-induced high runoff (Figure 4c). Following Typhoon Nesat, Typhoon Nalgae, which was farther from the Modaomen Outlet by 467 km but close to a high tide, reached a peak water level of 2.5 m at GC, even though it produced only 0.37 m of storm surge (Table 2). The chloride concentration increase was only 800 mg/L. This was due to the storm-induced rainfall from the previous typhoon Nesat, which pushed the river flow at WZ above 6000 m³/s. This increase rapidly suppressed saltwater intrusion at GC, putting an end to the severe upstream saltwater intrusion process (Figure 4c).

Around 23 October 2020, Typhoon Saudel passed through the PRE and then landed in central Vietnam (Figure 2). The closest distance from the typhoon track to the Modaomen Outlet was 460 km. The GC station experienced only a 0.11 m storm surge and a peak water level of 1.78 m. Although the river flow at WZ reached a high value of 6670 m³/s two days before the landfall of Typhoon Saudel, the Modaomen Outlet was in the transition from neap tide to spring tide during the passage of the typhoon. The storm surge prevented the subsiding of accumulated saltwater in the Modaomen Waterway into the South China Sea, resulting in a peak chloride concentration increase of 3400 mg/L. However, the SISA was short-lived, rapidly receding after the typhoon’s passage and exhibiting characteristics of saltwater intrusion driven by astronomical tides (Figure 4f).

The river discharge was relatively low on the 2 days prior to the passage of Typhoon Kompasu in 2021 and Typhoon Nesat in 2022, with river discharge of 3220 m³/s and 1780 m³/s at WZ, and there was already obvious saltwater intrusion. The maximum chloride concentration at GC was up to 2500 mg/L and 5700 mg/L (Figure 4g,h). During the typhoon’s transit and landfall, both were in the flood tide of the spring tide, with the closest distance from the typhoon’s path to the Modaomen Outlet by 325 km and 381 km. The GC station experienced storm surges of 0.62 m and 0.39 m, reaching peak water levels of 2.10 m and 2.05 m, resulting in chloride augmentations of 5700 mg/L and 4834 mg/L, and peak chloride concentration of 8200 mg/L and 9685 mg/L, respectively (Table 2). Notably, after the storm surge receded from Typhoon Kompasu, the chloride concentration at GC rapidly decreased to around 2000 mg/L (Figure 4g). In contrast, during the passage of Typhoon Nesat in 2022, the storm surge peak water level and peak chloride concentration at GC maintained around 2.0 m and 8000 mg/L, respectively, for three consecutive days, causing the chloride concentration to remain above 5000 mg/L for over a week (Figure 4h).

3.2. Spatiotemporal Variations of Storm Surge-Induced Salinity Augmentation

Utilizing in situ data of chloride concentration from various stations along the Modaomen Waterway, we conducted an analysis of the influence of storm surge on the spatial (Figure 5 and Figure 6) and temporal (Figure 7) variations of saltwater intrusion. Since chloride concentration data in DCK, DLS, LSW, MJ, XH, QL, and RY were only available for 2021 and 2022, we solely analyzed these data during Typhoon Kompasu in 2021 and Typhoon Nesat in 2022 (Figure 6 and Figure 7). Referring to drinking water quality standards, two chloride concentration thresholds were employed in the statistical analysis of saltwater intrusion duration: 250 mg/L and 2500 mg/L, corresponding to exceeding the standard and severe exceedance, respectively. The statistical timeframe encompasses only the 5-day period around the transit of typhoons, including about two to three days of both pre- and post-landfall periods.

Apart from Typhoon Nalgae in 2011, all nine typhoons resulted in severe SISA, characterized by an increase in chloride concentration exceeding 2000 mg/L at the GC station (Table 2). With the exception of Typhoon Saudel in 2020, the range of saltwater intrusion extended to at least the PG or Xihe (XH) station (Figure 7). Notably, Typhoon Nesat in 2022 caused saltwater intrusion to reach upstream as far as the RY station, resulting in chloride concentration exceeding the standard for a consecutive period of 20 h (Figure 6 and Figure 7).

During the passage of Typhoons Hagupit on 23 September 2008, Koppu on 15 September 2009, Kalmaegi on 16 September 2014, and Mangkhut on 16 September 2018, despite relatively high river flow observed at the WZ station in the two days preceding the typhoons, the GC station experienced severe chloride concentration exceedance (>2500 mg/L) for 19, 28, 5, and 15 h, respectively (Figure 7). Similarly, the PG station recorded chloride concentration exceedance (>250 mg/L) for 10, 12, 0, and 12 h, respectively (Figure 5 and Figure 7). The exceedance of the chloride concentration in GC significantly affected the water supply at downstream stations. On 24 October 2020, two days prior to the landfall of Typhoon Saudel, despite a higher river discharge of 6670 m³/s at the WZ station, the DCK station near GC experienced continuous chloride concentration exceedance for 86 h, primarily occurring during the combined effects of high tide and storm surge. Despite the continued impact of saltwater intrusion on water withdrawals at the GC station, water supplies at the upstream PG and ZZT stations have been much less affected by saltwater intrusion.

From 29 September to 3 October 2011, during the consecutive passage of Typhoons Nesat and Nalgae, two days preceding the landfall of Typhoon Nesat, the river discharge at WZ was only 1470 m³/s. This led to a significant spatial and prolonged influence of the storm surge on saltwater intrusion in the Modaomen Waterway. The duration of severe chloride concentration exceedance reached 102 h at GC, with a continuous exceedance period of 85 h and a peak chloride concentration as high as 12,300 mg/L. Additionally, the PG and ZZT stations, situated in the middle section of the Modaomen Waterway, experienced chloride concentration exceedance for 102 and 74 h, respectively, primarily during the landfall of Typhoon Nesat (Figure 5c and Figure 7c). The continuous saltwater intrusion that occurred at the GC, PG, and ZZT stations resulted in wider and sustained water supply impacts during the passage of Typhoons Nesat and Nalgae in 2011.

Prior to the landfall of Typhoons Kompasu on 13 October 2021, and Nesat on 18 October 2022, when the river discharge at WZ was relatively low, saltwater intrusion predominantly occurred at the DCK or GC station. The typhoon-induced storm surge resulted in saltwater intrusion reaching the XH or PG station, located 19 km away from GC. The peak chloride concentration at the XH station lagged behind the peak time at the DCK station by 12 h, with a continuous exceedance period of 15 h (Figure 6a and Figure 7g). However, two days before the landfall of Typhoon Nesat, saltwater intrusion had already reached the upstream MJ station, situated 11 km upstream from the GC station, with a peak chloride concentration of 4100 mg/L. During and after the passage of the typhoon, saltwater intrusion extended upstream to the RY station, located 45 km from GC, with a cumulative exceedance period of 20 h. In addition to the RY station, various downstream stations also experienced chloride concentration exceedance for different durations, and the XH or PG station experienced continuous exceedance and severe exceedance for 57 and 83 h, respectively. The GC station consistently recorded severe exceedance throughout the entire period (Figure 6b and Figure 7h). The passage of Typhoon Nesat exacerbated the range and duration of saltwater intrusion throughout the entire Modaomen Waterway, resulting in severe impacts on water supply for the Zhuhai, Macau, and Zhongshan water plants.

3.3. Changes in Water Level Gradients and Their Impact on Salinity Augmentation

During the passage of typhoons, storm surges in the estuary give rise to a sequential increase in water levels from the estuary to the upstream river channels, resulting in a sustained and significant landward positive pressure gradient. This phenomenon enhances the mixing of estuarine water and drives high-salinity water into the river channel landward, thereby intensifying saltwater intrusion. To investigate this phenomenon, this study focused on two typhoon events, namely Typhoon Mangkhut in 2018 and Typhoon Nesat in 2022. These typhoons varied in intensity during the transition from spring to neap tide. Four monitoring stations, namely DCK, MJ, XH, and Shouwei (SW), were selected for analysis. This study examined the instantaneous water levels during the period of 2 days prior to the typhoon’s landfall, the landfall day, and 2 days after the typhoon, considering flood tide, high tide, and ebb tide. A comparative analysis was conducted to evaluate the impact of storm surges on the along-river water level gradient and saltwater intrusion (Figure 8).

From 14 to 18 September, 2018, as Typhoon Mangkhut passed, notable variations in water level gradients were recorded in the Modaomen Waterway. During the flood tide and high tide (Figure 8a,b), the water level at DCK was 0.45 m higher than at SW, which is located 24 km upstream of DCK. Moreover, the water level at DCK was 1.5–2.0 m higher than the water level recorded 2 days prior to the typhoon’s landfall. The water levels along the river exhibited a significant landward decrease, leading to a strong positive pressure gradient that facilitated the landward transport of high-salinity water beyond the estuary. Two days before the typhoon, a landward positive pressure gradient was also observed along the Modaomen Waterway during flood tide. However, the water levels and gradient were much smaller than the peak of the storm surge (Figure 8a). At the high tide, there was a similar water level along the river 2 days before and after the typhoon’s landfall, indicating the absence of an apparent water level gradient. During the ebb tide, the water level at DCK during the storm surge was 0.67 m lower than at SW, and the water levels along the river exhibited a more pronounced seaward decrease, promoting the transport of saltwater toward the sea and the recession of saltwater intrusion. Two days after the typhoon’s landfall, water levels at all monitoring stations quickly returned to their pre-storm surge levels (Figure 8a–c), demonstrating a seaward decreasing water level gradient that aligned with the “single peak” characteristic of the chloride concentration variation curve recorded during Typhoon Mangkhut (Figure 4e and Figure 5e).

Around 18 October 2022, during the passage of Typhoon Nesat, the trend of water level gradient changes in the Modaomen Waterway at various monitoring stations resembled that observed before and after the landfall of Typhoon Mangkhut. However, the storm surge intensity was weaker, resulting in a smaller water level gradient. During flood tide, high tide, and ebb tide, the water level at DCK during the storm surge was only 0.11 m, 0.07 m, and −0.21 m higher than at SW, respectively (Figure 8d–f).

4. Discussions

4.1. Influence of Storm Surge and Discharge on Salinity Augmentation

In this study, we investigate the influence of typhoons on saltwater intrusion and introduce the concept of SISA. SISA is characterized by the drastic rise in water levels in the estuary due to storm surges, creating a significant and prolonged positive pressure gradient that facilitates the landward transport of high-salinity water. This phenomenon leads to a noticeable augmentation of saltwater intrusion in the estuary during typhoon passages. The SISA is equivalent to the maximum increase in chloride concentration (Table 2). Theoretically, similar to the definition of storm surge, the SISA can be calculated as the difference between the maximum chloride concentration during typhoon transit and the predicted chloride concentration, assuming no typhoon influence. For ease of calculation, we temporarily consider the chloride concentration at tidal moments corresponding to one to two days before typhoon landfall, when minimal typhoon impact is observed, as a substitute for the chloride concentration in the absence of a typhoon. Precise calculation of SISA can be achieved by establishing a high-precision estuarine salinity transport model.

The results of this study demonstrate a significant linear positive correlation (95% confidence) between SISA and storm surges at the downstream GC station in the Modaomen Waterway (Figure 9). Additionally, the total water levels, which combine storm surge and tide, exhibit a stronger correlation with SISA. Furthermore, there is a significant logarithmic negative correlation between the river discharge at the WZ station two days before typhoon landfall and the daily average chloride concentration and the duration of severe chloride concentration exceedance within five days at GC (Figure 9c,d). Previous studies by Pan et al. have also shown a significant negative correlation between the range of saltwater intrusion induced by typhoons and river discharge [26]. This reflects the inhibitory effect of river discharge on saltwater intrusion and SISA in the Modaomen Waterway. A higher river discharge leads to a stronger inhibitory effect, and when the river discharge at WZ exceeds 4000 m³/s, the daily average chloride concentration at GC is below 250 mg/L.

4.2. Impact of Typhoon Rainfall on Storm Surge-Induced Salinity Augmentation

During the transit or landfall of a typhoon, widespread rainfall occurs, increasing freshwater discharge in the basin and resulting in a rapid retreat of saltwater intrusion. This retreat has been observed during Typhoon Hagupit in 2008, Typhoon Nesat in 2011, and Typhoon Mangkhut in 2018 (Figure 4). In Figure 4, we primarily use the river discharge at WZ to represent the inhibitory effect of basin freshwater discharge on saltwater intrusion in the Modaomen Waterway before the typhoon’s transit or landfall. The river discharge at WZ shows a significant negative correlation with chloride concentration in the downstream Modaomen Waterway (Figure 9c). Thus, the WZ station, located in the middle-lower reaches of the West River, serves as a crucial hydrological station for the flow of the West River to the PRD. During the dry season, the river discharge at WZ is minimally influenced by tides, providing an accurate representation of basin freshwater conditions. With a distance of 344 km from WZ to the GC station and an average river discharge velocity of approximately 8.5 km/h at a river discharge of 4000 m³/s, it takes about 40 h for freshwater flowing from WZ to reach the Modaomen Outlet. Therefore, it is more reasonable to use the river discharge one to two days before the typhoon’s landfall to characterize the freshwater discharge of the Modaomen Waterway, as post-typhoon river discharge from WZ does. The actual discharge at the Modaomen Outlet is about one-third of the basin freshwater flow during the dry season [30].

Figure 10 illustrates the influences of six typhoon rainfall-induced runoffs on the retreat of SISA. These six typhoons brought substantial heavy rainfall-induced runoff to the downstream WZ and the North River basins, accelerating the retreat of saltwater intrusion. Since the impact of basin river discharge on saltwater intrusion is prolonged, this study uses daily average river discharge and chloride concentration for measurement.

For three typhoons of Hagupit in 2008, Kalmaegi in 2014, and Mangkhut in 2018, the river discharge at WZ exceeded 4000 m³/s one to two days before landfall. Correspondingly, the daily average chloride concentration at the GC station was below 250 mg/L or close to zero. Storm surges for these typhoons ranged from 1.15 to 1.77 m, resulting in daily average SISA ranging from 1300 to 6000 mg/L. Under the influence of typhoon rainfall-induced runoff, the basin freshwater discharge reached from 7100 to 27,000 m³/s, causing high chloride concentrations at GC station to last only one to two days before returning to pre-typhoon levels. For instance, during Typhoon Hagupit in 2008 and Kalmaegi in 2014, the high chloride concentration values at the GC station lasted only one day. Additionally, the three typhoons close to the PRE generated extremely high storm surges and salinity augmentation at the GC station (Table 2). Fortunately, this occurred during neap tide periods and was accompanied by high runoff from WZ, followed by substantial typhoon rainfall-induced runoff, significantly reducing the duration and range of storm surge-induced saltwater intrusion (Figure 4a,d,e and Figure 10a,d,e).

On 29 September 2011 (the 3rd day of September in the Lunar calendar, spring tide), when Typhoon Nesat transited, the river discharge at WZ was only 1470 m³/s, and the daily average chloride concentration was as high as 4761 mg/L. On 29 September, SISA increased the daily average chloride concentration to 8838 mg/L at GC. This was the first observation of the coupling effects of extremely low-flow conditions, spring tide, and storm surges, resulting in the ultimate duration and range of saltwater intrusion in the PRE. This had a severe impact on the water supply at several water plants along the Modaomen Waterway. After the typhoon transit, typhoon-induced rainfall increased the total discharge in the basin to 6409 m³/s on 29 September. By 1 October, the daily average chloride concentration at GC had decreased to 2476 mg/L (Figure 10c). Subsequently, Typhoon Nesat transited on 3 October, maintaining a total discharge in the basin exceeding 8500 m³/s for three consecutive days. This caused the daily average chloride concentration at GC to decrease to 228 mg/L and 56 mg/L on October 5 and 6, respectively, completely repelling the saltwater intrusion.

Figure 10 summarizes and compares the relationships between the changes in river discharge at the WZ station and total basin discharge before and after typhoon transit and the decrease of chloride concentration at the GC station. The rainfall brought by the typhoon generally increases the runoff in the upstream watershed by less than 10% of that generated in the lower section of the West River basin from WZ to the Modaomen Outlet and in the North River basin. If only the change in river discharge at the WZ station is considered, an unreasonable conclusion would be drawn: there is no correlation between the rate of decline in daily average chloride concentration at GC and the increase in river discharge at WZ (Figure 11a). A more reasonable approach is to consider the total change in basin discharge due to rainfall–runoff, i.e., the rate of decline in daily average chloride concentration at GC will increase with the increase in river discharge at WZ, with a lag effect. Considering the total basin discharge, including runoff from the downstream WZ and North River basins, provides a better explanation for the decreased rate and duration of storm surge-induced saltwater intrusion (Figure 10 and Figure 11). Due to the flow division to the other seven outlets, the actual discharge via the Modaomen Outlet is about one-third of the basin’s total discharge during the dry season.

4.3. Impact of Wind Direction on Saltwater Intrusion

Wind plays a crucial role in saltwater intrusion, and its impact can be categorized into two types: local winds within the estuary and remote winds outside the estuary. The mechanisms of wind-driven saltwater intrusion vary depending on the season and the dominant wind directions [35].

In the Modaomen Outlet, the prevalence of down-estuary winds, such as north and northeast winds, enhances the transport of surface water toward the sea. This promotes stratification within the estuary and strengthens the exchange flow. As a result, the landward movement of the bottom saltwater wedge is facilitated, leading to an increase in the range and duration of saltwater intrusion. Conversely, during periods of prevailing up-estuary winds, such as southern and southeastern winds, vertical mixing in the Modaomen Waterway intensifies. This weakens the estuarine gravitational circulation and mitigates saltwater intrusion [19,36]. Outside the estuary, remote winds can drive offshore water into the estuary, causing elevated water levels on the continental shelf and in the estuary itself. This, in turn, promotes saltwater intrusion [37,38]. The impact of remote winds is particularly pronounced before the landfall of a typhoon, especially when combined with storm surges, which raise the tidal level in the estuary. The rapid increase in tidal level leads to a swift rise in salinity in the Modaomen Waterway [4,25,35]. However, due to the inhibiting effect of high river discharge at the end of the flood season and the influence of typhoon rainfall, the range and duration of saltwater intrusion were generally limited (Figure 4, Figure 5, Figure 6 and Figure 7).

During the transit of a typhoon, the characteristic rotation of winds around the typhoon’s eye creates a cyclonic circulation. Consequently, the dominant wind direction in the Modaomen Outlet changes as the typhoon center moves closer (from east to west) [39,40]. Initially, when the typhoon center is approaching the PRE, the dominant wind direction is from the northeast. At this stage, the estuarine water level and positive pressure gradient are relatively small, and salt transport in the Modaomen Outlet is mainly governed by estuarine exchange flow [18]. As the typhoon center passes through the PRE, the dominant wind direction shifts to the southeast, resulting in up-estuary winds. The estuarine water level, influenced by both remote winds and storm surges, rises rapidly. Under the interaction of waves and currents, landward advection and tidal oscillatory salt transport are significantly enhanced, leading to severe storm surges and salinity augmentation [19]. As the typhoon center moves away from the PRE, the recession of the storm surge causes a decrease in the water level in the estuary. The dominant wind direction changes to up-estuary winds, intensifying mixing in the estuary, which facilitates the recession of saltwater intrusion in the Modaomen Outlet (Figure 12a–c) [39,40].

For example, during the landfall and passage of Typhoon Hagupit in 2008 and Typhoon Mangkhut in 2018, both typhoons brought up-estuary southeast winds (Figure 12a–c,g–i). As a result, storm surges of 1.77 m and 1.64 m occurred, with total water levels reaching 3.20 m and 2.95 m at the Modaomen Outlet (GC station). The storm surges propagated inland, reaching the top of the delta at the Makou (in Foshan) and Gaoyao (in Zhaoqing) locations [41]. The storm surge caused the water level in the PRE to be significantly higher than in the upstream reaches of the West River in the PRD (Figure 8), inducing the rapid landward transport of strongly mixed high-salinity water along the river channels. This led to a sharp increase in chloride concentration in the Modaomen Waterway, reaching 9600 mg/L and 6200 mg/L, respectively, during the two typhoons. The saltwater intrusion range exceeded 20 km, reaching the PG station. However, due to the end of the flood season conditions, the WZ river discharge was 5000 m³/s and 4910 m³/s, respectively, for the two typhoons. The typhoon-induced rainfall further increased the discharge to over 6200 m³/s and 8500 m³/s, significantly shortening the range and duration of saltwater intrusion (Figure 4, Figure 5, Figure 6 and Figure 7, Table 2). Notably, Typhoon Mangkhut, with a closer landfall location and trajectory to the PRE, was predominantly influenced by southern and southeastern winds for an extended period, resulting in a longer duration of saltwater intrusion compared to Typhoon Hagupit.

The distance from the typhoon path to the PRE determines the minimum pressure and maximum wind speed and influences the dominant wind direction, thereby affecting storm surge and salinity augmentation in the PRE. For instance, Typhoon Saudel in 2020 had a minimum distance of 460 km to the PRE. Before the typhoon’s landfall, north winds prevailed, shifting to northeast winds after the landfall (Figure 12j–l). Consequently, the storm surge at the Modaomen Outlet (GC station) was only 0.11 m, with a maximum water level of 1.75 m. The SISA was 3400 mg/L, and the range and duration of saltwater intrusion were smaller compared to other typhoons under equivalent runoff and tidal conditions (Figure 4, Figure 5, Figure 6 and Figure 7, Table 2).

Chen et al. found that the dominant wind field of east and southeast winds promotes the landward transport of strongly mixed high-salinity water, intensifying saltwater intrusion in the PRE [42]. Typhoons generated in the northwest Pacific primarily move in westward tracks (38%), northward tracks (21%), or west–eastward turn tracks (31%) [43]. Typhoons with a westward track have the highest landfall ratio (46%) and cause the most significant storm surge and salinity augmentation in the PRE (Figure 4, Figure 5, Figure 6 and Figure 7, Table 2). Specifically, for Typhoon Nesat in 2022, the typhoon initially moved westward and then shifted to a southwestern direction parallel to the shore of Guangdong Province. This resulted in prolonged northeast winds in the outer sea of the PRE (Figure 12p–r), causing continuous storm surge and salinity augmentation in the Modaomen Outlet for three consecutive days (Figure 4h). Saltwater intrusion reached the RY station, located 45 km from the GC station (Figure 6b), and chloride concentrations at the DCK and XH (PG) stations exceeded 2500 mg/L for 118 h and 58 h, respectively, during the five-day period (Figure 7h). These conditions significantly affected the water supply for water plants along the Modaomen Waterway.

5. Conclusions

Based on in situ hydrological and chloride data, this study conducted a comparative analysis of the impact of storm surges on saltwater intrusion in the Modaomen Waterway of the PRE during the passage of ten autumn typhoons from 2006 to 2022. The concept of Storm surge-Induced Salinity Augmentation was introduced to elucidate the characteristics of intensity, range, and duration of saltwater intrusion. The study also discussed the main influencing factors of SISA, leading to the following conclusions:

(1)

The SISA predominantly occurs in September and October, particularly in years and months, with typhoon landfalls affecting the PRE.

(2)

The distance between the typhoon path and the PRE affects SISA’s intensity, range, and duration. Typhoons with a closer landfall location and trajectory to the PRE tend to result in stronger salinity augmentation and longer durations of saltwater intrusion.

(3)

The dominant wind direction during a typhoon, particularly east and southeast winds, promotes the landward transport of strongly mixed high-salinity water, intensifying saltwater intrusion in the PRE.

(4)

Typhoons with a westward track cause the PRE’s most significant storm surges and salinity augmentation.

The conclusions of this study are primarily based on in situ chloride data along the Modaomen Waterway. Our findings elucidate SISA’s characteristics and principal influencing factors, offering potential improvements in forecasting SISA occurrences, understanding its ramifications on domestic and industrial water supply, and guiding administrators in allocating reservoir resources to mitigate saltwater intrusion. In future research, our objective is to implement a 3D numerical model to simulate the storm surge and salinity transport in the PRE, aiming for a more precise characterization and quantification of the spatiotemporal distribution and dynamic driving mechanisms of SISA.