Next Article in Journal
Experimental Study on Blocky Cuttings Transport in Shale Gas Horizontal Wells
Previous Article in Journal
Effects of Cyanobacteria on Competitive Interactions Between Different-Sized Cladoceran Species
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Water
Volume 17
Issue 7
10.3390/w17071015
water-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Distinct Flood Diversion Mechanisms and Comparable Effects on Discharge Fraction and Peak Water Levels over X-Shaped and H-Shaped Composite River Nodes

by
Yongjun Fang
1,2,
Xianwei Wang
2,3,4,*,
Jie Ren
4,5,*,
Huan Liu
4,5,
Peiqing Yuan
2,3,6 and
Yazhou Ning
2,3
1
School of Geomatics, Zhejiang University of Water Resources and Electric Power, Hangzhou 310018, China
2
School of Geography and Planning, Sun Yat-sen University, No. 135 Xingang West Road, Haizhu District, Guangzhou 510275, China
3
Guangdong Provincial Engineering Research Center for Public Security and Disasters, Guangzhou 510275, China
4
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519082, China
5
School of Marine Sciences, Sun Yat-sen University, No. 2 Daxue Road, Gaoxin District, Zhuhai 519082, China
6
Yellow River Institute of Hydraulic Research, Zhengzhou 450003, China
*
Authors to whom correspondence should be addressed.
Water 2025, 17(7), 1015; https://doi.org/10.3390/w17071015
Submission received: 26 February 2025 / Revised: 26 March 2025 / Accepted: 28 March 2025 / Published: 30 March 2025
(This article belongs to the Section Hydrology)
Browse Figures
Versions Notes

Abstract

:
River nodes play a crucial role in regulating water and sediment transport within river networks. The SiXianJiao (SXJ) node serves as a key exchange point between the West River (WR) and North River (NR) in the Pearl River Delta (PRD), South China. Understanding the differences in flood diversion dynamics between X-shaped and H-shaped configurations under varying geomorphic conditions is essential for flood management. This study employs the Delft3D-Flow model to investigate the flood diversion mechanisms of these composite river nodes. Results revealed distinct hydrodynamic behaviors: the X-shaped node facilitates greater water exchange due to a shared channel segment, whereas the H-shaped node experiences restricted exchange due to flow resistance in the connecting branch. Both configurations exhibit self-regulating flood diversion processes that significantly reduce flood risks. A critical flow fraction of approximately 75.9% [WR/(WR + NR)] is identified, at which water levels (WLs) at both ends of the SXJ node almost equalize. When the WR flow fraction exceeds this threshold, floodwaters are diverted toward the NR. Below it, the diversion direction reverses. Additionally, flood diversion synchronizes asynchronous flood waves, stabilizing the discharge fraction at Makou (Sanshui), which fluctuates around 75.8% (24.2%) for the X-shaped node and 76.6% (23.4%) for the H-shaped node. These findings enhance our understanding of flood diversion dynamics and provide valuable insights for optimizing flood mitigation strategies and hydraulic infrastructure planning in the PRD and comparable river systems worldwide.

1. Introduction

Water and sediment distribution at river nodes is governed by regional factors, such as upstream and downstream forcing, and local morphological features, including channel geometry, topography, meanders, and river bars, which shape internal hydrodynamics [1] Among these, the morphology of river nodes is one of the most important local factors in determining the distribution of water and sediment. Common configurations of river node morphologies include Y-shaped confluence nodes, inverse Y-shaped bifurcation nodes, and X-shaped nodes [2,3,4]. The Y-shaped confluences have been extensively studied through field monitoring, hydraulic experiments, and numerical simulations [5,6,7]. The inverse Y-shaped bifurcation node, typically formed by a single upstream river, is primarily influenced by channel dimensions and hydraulic roughness, which dictate the water and sediment distribution across downstream branches [1,8,9]. Studies using 1D and 2D theoretical models have explored their stability, sediment transport, and channel evolution [8,10,11]. The X-shaped node integrates a Y-shaped confluence with an inverse Y-shaped bifurcation, enabling upstream convergence and redistribution of water and sediment across downstream branches. A study introduced an H-shaped composite river node, exemplified by the SXJ node in the PRD, where a short natural channel connects the WR and NR [12]. This system, which combines an inverse Y-shaped bifurcation followed by a Y-shaped confluence, enables water and sediment to initially divert from one of the two upstream rivers into the lateral channel before converging at the Y-shaped confluence.
The PRD is characterized by a complex river network [13], with the SXJ node serving as a critical exchange node that modulates the allocations of water and sediment between the WR and NR, which is significant in the morphological development and progression of the delta. Around 2000 years ago, the SXJ river node was an extensive water domain featuring an X-shaped composite river node, where the WR and NR first converged and then split into two or more branches. This configuration allowed for the self-adjustment of upstream fresh flow and estuarine tidal dynamics [14,15,16]. To claim land from riverine and tidal flats and mitigate flood, large-scale levees and floodgates have been constructed and separated the WR and NR. Nowadays, these two rivers are connected by a 3-km-long, 400-m-wide SXJ waterway, which forms an H-shaped composite river node. At this site, water is initially diverted at an inverse Y-shaped bifurcation before converging at a Y-shaped confluence, enabling bidirectional flow exchange between the WR and NR [12].
The SXJ waterway plays a pivotal role in regulating water and sediment exchange between the WR, NR, and the downstream network [17,18,19]. This waterway generates an unstable hydrological state, characterized by substantial water level (WL) differences between the two ends of the SXJ node, which gradually approaches equilibrium as a result of mutual flow diversion. The flow in the SXJ waterway reverses direction twice daily, driven by the asynchronous semi-diurnal tide from the downstream branches and the irregular flood waves from the upstream WR and NR [12]. From 1959 to 1992, water primarily flowed from the NR to the WR, accounting for 19.0% of the annual freshwater flow from the upstream NR. At the Sanshui section, this flow accounted for only 14.5% of the total flow at both the Makou and Sanshui sections, which are crucial to the downstream branch. From 1993 to 2016, the annual average flow via the SXJ waterway had reversed direction, with water diverted from the WR to the NR, accounting for 31.9% of the freshwater from the NR. This shift led to a 62.0% increase in annual average flow at Sanshui, and the diversion ratio, the proportion of flow distribution between the WR and NR, increased to 24% [20,21].
The reversion of flow in the SXJ waterway and the increasing water and sediment diversion at Sanshui had significant implications for flood control, freshwater management, and aquatic ecosystem reservations in the PRD. Before the early 1990s, Makou received approximately 82% of the total flow from the WR and NR, but this proportion declined to 75% after the 1990s [17,20,22]. Consequently, the discharge at Sanshui nearly doubled during this period [23]. The abrupt changes around 1992 can primarily be due to the uneven riverbed undercutting in the downstream branches, along with a decrease in upstream sediment input induced by anthropogenic activities, such as reservoir construction and land use changes [17,24]. While extensive studies have quantified shifts in water and sediment distribution at Makou and Sanshui, the underlying mechanisms remain unclear. Specifically, it is still uncertain whether these changes are linked to alterations in the river node morphology.
The SXJ node has undergone significant morphological evolution, transitioning from an X-shaped to an H-shaped configuration. Historically, the X-shaped node facilitated simultaneous flow confluence and division, enabling unrestricted water and sediment exchange between the WR and NR [15,16]. In contrast, the present H-shaped SXJ node regulates exchange primarily through the hydraulic head difference between the two rivers while being constrained by the geometric dimensions of the SXJ waterway [12]. Proposals have been made to construct flow-regulating structures, such as sluice gates or submerged dikes, within the SXJ waterway, which could potentially reduce its modulating capacity. Over the past 2000 years, levee and sluice gate construction has progressively altered the exchange dynamics, reshaping the SXJ node from an expansive water domain into its current H-shaped form. However, there remains a research gap on the exchanging capacity limits and demands in different shapes of the SXJ waterway. Addressing such questions may help river dredging and management in the SXJ primary node and give some thoughtful suggestions for planning reasonable flood control facilities and risk management policies at a composite river node.
The primary objective of this research is to study the flood diversion capacity of X-shaped and H-shaped SXJ river nodes using numerical simulations, comparing their flood diversion mechanism and effects on flood modulations, including discharge fractions and peak flood stages in the downstream channels. However, limitations exist in capturing complex real-world interactions, particularly those influenced by historical evolution and anthropogenic modifications. The structure of the paper is as follows: Section 2 describes the material and methods. Section 3 presents the results, which are followed by the discussion in Section 4 and the conclusion in Section 5. This research advances the understanding of flood diversion in composite river nodes and offers insights for optimizing riverine facilities, especially at an H-shaped composite river node.

2. Materials and Methods

2.1. Study Area

The Pearl River Delta features a hydrological system where the West River (WR), North River (NR), and East River (ER) converge before dispersing into the sea through eight outlets. (Figure 1). This low-lying delta is interwoven with a complex network of tributaries and distributaries. The SXJ node, located in Foshan City, has a total catchment area of 3.99 × 105 km2, with the upstream watersheds of Gaoyao and Shijiao contributing 3.52 × 105 km2 and 3.84 × 104 km2, respectively. Gaoyao and Shijiao serve as the key hydrological stations, situated 45 km and 52 km upstream of the SXJ node, respectively [12]. As the principal exchange node between WR and NR, the SXJ node functions to regulate flow and sediment redistribution across the downstream river network.
About two thousand years ago, during the Qin and Han dynasties, the SXJ was a wide water domain rather than the narrow waterway it is today, consisting of numerous islands and sandbars [14,15]. During the Tang Dynasty and then after, the SXJ area remained a wide confluence of rivers; meanwhile, sedimentation began to silt up.
About one thousand years ago, during the Song Dynasty (960~1279), although a few sandbars were formed and low dikes were piled up, the SXJ was still a wide water domain with free exchanges of flows between two rivers (Figure 2a). With the advancement of the Suijiang Delta to the southeast, the tributaries of the SXJ district migrated to the southeast, and sandbars became larger and larger. During the Ming (1368~1644) and Qing dynasties (1644~1911), more and more artificial earth dikes were constructed, and the SXJ waterway was gradually formed as a flood regulation and navigation channel connecting the two rivers (Figure 2b,c). The SXJ area became a natural floodplain, and low-frequency flood events could easily be overtopped and destroy the dikes, such as the 1:100 years flood in 1915. From the point of view of flood flow, the SXJ was still an X-shaped composite river node.
About 60 years ago, large-scale joint levees and sluice gates were constructed to control floods at frequencies lower than 1:50 years. These hydraulic structures disrupted the natural water and sediment exchange between the WR and NR, transforming the X-shaped node into an H-shaped configuration (Figure 2d). The wide water domain was turned into a narrow SXJ waterway, which is about 3 km long from east to west, 200~400 m wide, and 10~25 m deep [12].

2.2. Methods

A flowchart for the study is illustrated in Figure 3. The data and model setup of this research are stepped from the study [12]. The Delft3D-Flow (Version 3.15) model was configured using Cartesian grids based on bathymetry data, which included information on levee heights and riverbed elevations [25]. The simulation domains span from the upstream locations of Gaoyao in the WR and Shijiao in the NR to the downstream points of Makou in the WR and Shijiao in the NR (Figure 1 and Figure 3). The model’s upper boundary is specified by the flow conditions at Gaoyao and Shijiao, while the downstream boundary is governed by stage-flow curves derived from historical flood flows and WLs observed at Makou and Sanshui.
The model was implemented using a depth-averaged 2D framework based on shallow water equations, which solve the unsteady flow dynamics, including horizontal motion equations, the continuity equation, and the transport equations for conservative elements [25]. The choice of depth-averaged 2D simulations was motivated by their comparability to 3D simulations in various coastal scenarios [25,26]. Model parameters followed the recommended values in the user manual [25]. The bed roughness was uniformly set to a Manning coefficient of 0.035 m−1/3s across all simulations, consistent with the setup in the study [12]. A time step of 3 s was applied to all simulations to maintain a Courant number below 4 2 , as required for the ADI scheme [25].
The model was calibrated using WL and flow data with the 2005 flood, optimizing parameters based on NSE (Nash-Sutcliffe efficiency coefficient), MAD (mean absolute difference), and RMAD (relative MAD). At Makou and Sanshui, NSE values of water and discharge are 0.99, with WL MADs of 0.19 m and 0.14 m, and discharge MADs of 1262 m3/s (RMAD = 5.2%) and 425 m3/s (RMAD = 5.6%). Validation with 2022 flood data showed NSE values of 0.94 and 0.97 for WLs and 0.93 and 0.96 for discharge, with WL MADs of 0.26 m, 0.17 m, and discharge MADs of 983 m3/s (RMAD = 2.8%) and 627 m3/s (RMAD = 5.8%) [12].
After calibration and validation, the model was used to simulate both flood events (unsteady flow) and various flood scenarios (steady flow). The two flood events occurred in 2006 and 2022. The flood scenario simulations include 121 combinations of the upstream inflow (Table 1), which were based on the design flood flow provided by the DWR (Department of Water Resources of Guangdong Province) in 2002 (Table 2). The inflow rates at Gaoyao ranged from 10,000 m3/s to 60,000 m3/s in 5000 m3/s increments, and at Shijiao, increments were from 2000 m3/s to 22,000 m3/s in 2000 m3/s. The model setup for the H-shaped river node followed a configuration used by a study [12], incorporating external forcing, Manning’s coefficient, viscosity, and other relevant parameters.
In addition, the topography of the SXJ node was modified to an X-shaped node by removing the levees and sluice gates at the northern peninsular of the SXJ waterway (Figure 4). The land elevation of the peninsular is 2–5 m, except for three small hills about 20 m high and 200 m wide, while the flood stage is over 10 m for a 1:50-year flood event. Using similar model parameters to those for the calibrated H-shaped node, the same flood events and scenarios were simulated again. Various comparisons and analyses are carried out.

3. Results

The results present comparisons for unsteady flow simulations of two flood events in 2006 and 2022 and steady flow simulations of 121 flow combinations over the X-shaped and H-shaped SXJ nodes (Table 1).

3.1. Comparison of Unsteady Flow Simulations

3.1.1. The July 2006 Flood Event

The Delft3D hydrodynamic model was constructed using grids for both X-shaped and H-shaped composite nodes to simulate the July 2006 flood event (Figure 4). The flood persisted from 11 July to 31 July, with three peaks in the WR and two in the NR. The peak discharge at Gaoyao reached 34,564 m3/s, approximating a 2-year return period flood, while at Shijiao, it reached 18,245 m3/s, corresponding to a 60-year return period flood (Figure 5a), while the incoming flow percentage of the WR at Gaoyao varied from 58% to 95% (Figure 5b). The mean discharge fraction at Makou (Sanshui) slightly fluctuated around 76.4% (23.6%) for the X-shaped node and 76.8% (23.2%) for the H-shaped node. Based on the diversion process at the SXJ node, the flood event was categorized into four stages (Figure 5e,f).
During Phase I (11–16 July), the WR’s inflow ratio at Gaoyao remained around 94%, while the discharge fractions at Makou declined to 76.4% for the X-shaped node and 78.0% for the H-shaped node due to reduced flood diversion (Figure 5b). Although the peak diversion at the H-shaped (−4000 m3/s, Figure 5c) was slightly lower than that at the X-shaped node (−4480 m3/s, Figure 5d), the peak WL difference between the two ends of the SXJ waterway was significantly greater at the H-shaped node (28.0 cm) compared to the X-shaped node (1.0 cm). The flood diversion effectively mitigated peak WLs and flood risk in both rivers (Figure 5e,f). After the flood diversion, peak WLs at Makou dropped by 0.96 m and 0.83 m for the X-shaped and H-shaped nodes, respectively, while those at Sanshui rose by 3.72 m and 3.29 m. Consequently, the peak WL differences between the WR and NR decreased from 4.68 m to 0.01 m at the X-shaped node and from 4.12 m to 0.28 m at the H-shaped node, enhancing levee stability across the SXJ node.
During Phase II (16–20 July), the NR’s maximum incoming flow ratio at Shijiao increased from 10% to 42%, becoming the dominant factor influencing WLs at the SXJ node. Consequently, the flood was diverted from the NR to the WR (Figure 5). The maximum flood diversion reached 6800 m3/s over the X-shaped node (Figure 5c) and 6300 m3/s over the H-shaped node (Figure 5d). After flood diversion, the peak WLs at Sanshui decreased by 3.63 m and 3.35 m for the X-shaped and H-shaped nodes, respectively, while peak WLs at Makou increased by 1.51 m and 1.34 m. This redistribution reduced the peak WL differences between the WR and NR at the SXJ node from 5.14 m to 0.02 m and from 4.69 m to 0.29 m, respectively (Figure 5e,f).
During Phase III (20–27 July), the NR flood receded significantly faster than the WR (Figure 5a). Consequently, the inflow ratio at Shijiao declined from 42% to 9%, while at Gaoyao, it increased from 58% to 91% (Figure 5b). This shift led to a reversal in flood diversion, with water flowing from the WR to the NR, mitigating flood risk in the WR. The peak diversion reached −5100 m3/s at the X-shaped node (Figure 5c) and −4600 m3/s at the H-shaped node (Figure 5d). After the flood diversion, peak WLs at Makou dropped by 0.86 m and 0.75 m at the X-shaped and H-shaped nodes, respectively, while at Sanshui, they rose by 3.72 m and 3.40 m. This diversion reduced the peak WL differences between the WR and NR from 4.58 m to 0.01 m for the X-shaped node and from 4.15 m to 0.24 m for the H-shaped node (Figure 5e,f).
During Phase IV (27–31 July), the maximum inflow ratio at Shijiao increased from 9% to 40%, which once again dominated the WLs at the SXJ node and drove floodwater diversion from the NR to the WR (Figure 5). The peak diversion reached 3580 m3/s at the X-shaped node (Figure 5c) and 3300 m3/s at the H-shaped node (Figure 5d). After the flood diversion, peak WLs at Sanshui decreased by 2.35 m for the X-shaped node and 2.20 m for the H-shaped node. Meanwhile, the peak WLs at Makou increased by 0.98 m and 0.90 m, which reduced the peak WL differences between the WR and NR from 3.33 m to 0.01 m and from 3.10 m to 0.19 m over the X-shaped and H-shaped nodes, respectively (Figure 5e,f).
Figure 6 illustrates the instantaneous flow fields (WL replacement) simulated over the X-shaped and H-shaped composite river nodes at 12:00 on 18 and 23 July 2006, and their flow division showed distinct mechanisms. At the X-shaped composite river node, the WR and NR flow initially merged at the Y-shaped confluence node before splitting at the inverse Y-shaped bifurcation node (Figure 6a,b). On 18 July, the NR flood waves dominated the flood stages at the SXJ node, and the flood water mainly diverted from the NR to the WR (Figure 6a). On 23 July, the flood waves from the WR dominated the flood stages over the SXJ river node, with floodwater diversion from the WR to the NR (Figure 6b). At the H-shaped composite river node, on 18 July, floodwater first diverted from the NR to the SXJ waterway, which then converged into the WR (Figure 6c). Similarly, on 23 July, the flood water from the WR was diverted into the SXJ waterway, where it subsequently converged into the NR (Figure 6d).

3.1.2. The June 2022 Flood Event

In June 2022, the flood event comprised three distinct flood waves impacting both rivers (Figure 7a), with flood diversion alternatively governed by the inflows of the WR or NR. The peak discharge of the WR at Gaoyao reached 42,300 m3/s, corresponding to a 10-year return period, while the NR at Shijiao reached 19,400 m3/s, approaching a 100-year return period flood (Figure 7a). The inflow ratio from the WR varied from 65% to 81.6% (Figure 7b). The mean discharge fractions at Makou (Sanshui) fluctuated slightly, around 75.8% (24.2%) and 76.0% (24.0%) at the X-shaped and H-shaped river nodes, respectively (Figure 7b). Based on the flood diversion dynamics, the event was categorized into three distinct phases (Figure 7e,f).
During Phase I (5–14 June), the incoming flow ratio of the WR at Gaoyao varied around 81%, while the downstream flow fractions at Makou decreased to 75.8% for the X-shaped node and 76.2% for the H-shaped node (Figure 7b). The peak diversion flow was slightly higher at the X-shaped (−2400 m3/s, Figure 7c) than at the H-shaped node (−2100 m3/s, Figure 7d), with peak WL difference of 1.0 cm and 5.0 cm, respectively, between the two ends. The flood diversion greatly reduced the peak WLs and flood risk for both rivers (Figure 7e,f). Following the flood diversion, peak WLs at Makou decreased by 0.34 m and 0.30 m at the X-shaped and H-shaped nodes, respectively, while those at Sanshui increased by 1.60 m and 1.46 m. Consequently, the peak WL differences between the two ends at the SXJ waterway were reduced from 1.94 m to 0.01 m and from 1.76 m to 0.05 m, respectively.
During Phase II (14–19 June), the maximum inflow ratios of the NR at Shijiao increased from 18.3% to 26.5%, resulting in a near-equilibrium condition with similar WL heights at both ends of the SXJ waterway (east mouth–west mouth) at the X-shaped and H-shaped river nodes (Figure 7c,d). The flood diversion direction shifted three times, with average flow rates of 61 m3/s and 176 m3/s at the X-shaped and H-shaped nodes, respectively (Figure 7c,d).
During Phase III (19–26 June), a significant flood wave occurred in the NR, causing its maximum incoming flow ratio at Shijiao to increase from 21.6% to 35.1%, which dominated the WLs at the SXJ node. As a result, floodwater was diverted from the NR to the WR. The maximum flood diversion reached 5970 m3/s at the X-shaped node (Figure 7c) and 5700 m3/s at the H-shaped node (Figure 7d). After flood diversion, the peak WLs at Sanshui decreased by 2.98 m and 2.80 m at the X-shaped and H-shaped river nodes. Simultaneously, the peak WL at Makou increased by 1.13 m and 1.10 m, which reduced the peak WL differences between the WR and NR at the SXJ node from 4.11 m to 0.03 m and from 3.90 m to 0.16 m, respectively (Figure 7e,f).
Figure 8 illustrates the instantaneous flow fields (WL replacement) simulated over the X-shaped and H-shaped composite river nodes at 12:00 on 10 and 23 June 2022, and their flow division also showed distinct mechanisms. Over the X-shaped composite river node, on 10 June, the flood waves from the WR dominated the flood stages over the SXJ node, and the floodwater was primarily diverted from the WR to the NR (Figure 8a). On 23 June, the flood waves from the NR dominated, and the floodwater mainly diverted from the NR to the WR (Figure 8b). Over the H-shaped composite river node, on 10 June, floodwater first diverted from the WR to the SXJ waterway, which then converged to the NR (Figure 8c). Similarly, on 23 June, floodwater first diverted from the NR to the SXJ waterway, which then converged to the WR (Figure 8d).

3.2. Comparison of Steady Flow Simulations

In total, 121 steady flow scenarios, combining upstream WR and NR conditions, were simulated for both the X-shaped and H-shaped SXJ nodes, respectively.
The WL differences between the NR and WR approximated to zero over the X-shaped node even under different incoming flows from both rivers (Figure 9a). This indicates that the X-shaped node has enough diversion capability and the flood flow from either upstream river can freely exchange and redistribute. The discharge fractions are mainly determined by the geometric dimensions of the downstream branches, thus maintaining a nearly constant value once the flow at Gaoyao is larger than 20,000 m3/s (Figure 9e). In contrast, the WL differences between the western and eastern mouths of the SXJ waterway range from −0.69 m to 1.50 m (Figure 9b), primarily due to the limited diversion capacity of the SXJ waterway at the H-shaped node. Moreover, the discharge fractions have a much larger variation range than those over the X-shaped node (Figure 9f). When the flow at Gaoyao is larger than 30,000 m3/s, the discharge fraction at Makou (Sanshui) fluctuates between 74.3% and 77.7% (22.4% and 25.7%). This suggests that the discharge fractions at the H-shaped node are determined not only by the geometry of the downstream branches but also by the inflow from the WR or NR and the flood diversion rate through the SXJ waterway.
The flood diversion rates over the X-shaped node range from −13,000 m3/s (WR to NR) to 14,300 m3/s (NR to WR) (Figure 9c), which directly reduces the peak WL at Makou (Sanshui) by up to 1.65 m (7.50 m), while increasing the peak WL at Sanshui (Makou) by up to 8.62 m (3.60 m) (Figure 10a). Likewise, the flood diversion rates over the H-shaped node vary from −11,930 m3/s (WR to NR) to 11,990 m3/s (NR to WR) (Figure 9d), which directly reduces the peak WL at Makou (Sanshui) up to 1.48 m (6.05 m); meanwhile, the peak WL at Sanshui (Makou) also increases up to 8.02 m (3.23 m) (Figure 10b). As a result of the flood diversion, the peak WL reduction in one river and rise in another river make both rivers reach similar peak WLs and greatly reduce the flood risk over the SXJ node. The flood diversion effect is larger for the NR than the WR, and the X-shaped node shows a larger flood diversion effect than the H-shaped node.

4. Discussion

This section mainly discusses the diversion mechanism, the flood diversion effects, and the implications over the X-shaped and H-shaped composite river nodes.

4.1. Flood Diversion Mechanisms

For a single upstream and inverse Y-shaped node, the downstream flow division is mainly influenced by local factors such as bends and branch channel geometry [28,29,30], while flow division at an X-shaped or an H-shaped river composite node is affected not only by local factors such as branch channel geometry and cross-channel circulation [4,16] but also by regional factors such as the incoming flow ratios from both rivers [12]. As illustrated in the flow fields (WL replacement), the flow division showed distinct mechanisms over an X-shaped and an H-shaped composite river node (Figure 5 and Figure 7). Over the X-shaped composite river node, the flows from the WR and NR initially converge at the Y-shaped confluence node and then diverge at the inverse Y-shaped node (Figure 5a,b and Figure 7a,b). The regular flow and flood water could freely redistribute between the two rivers or channels [4,16]. In contrast, at the H-shaped composite river node, such as on 10 June 2022, the floodwater first diverted from the WR to the SXJ waterway, which then converged to the NR (Figure 9c). The SXJ waterway is essential for facilitating floodwater diversion between the WR and NR. The relative magnitude of flood waves from each river governs the flood stages and the WL difference, which drives the flood diversion through the SXJ waterway (Figure 5a,d and Figure 6a,d). The diversion capacity is further constrained by the geometrical features of the waterway, including its width, depth, and length [12].
Meanwhile, flood diversion at the SXJ node is also influenced by downstream tidal forcing, including spring tides, neap tides, and storm surges, especially when the flood stage at the SXJ node is less than 3 m above sea level [24]. The higher tidal levels were reported to be more beneficial to the flow division in the downstream branch of the NR [31]. Since this study focuses on flood diversion for flood stages higher than 6 m or for flood events over a 2-year return period, the influence of tidal force on flood diversion is nearly negligible and is not investigated in this study [12].

4.2. Flood Diversion Effects

The asynchronous and unbalanced flood waves from the upstream WR (Gaoyao) or NR (Shijiao) dominated the flood stages and determined the diversion direction of flood water over the SXJ node. This led to synchronized flood discharge (Figure 4a and Figure 6a) and nearly constant discharge fractions (Figure 4b and Figure 6b) at Makou and Sanshui in the downstream channels during the two flood events. The discharge fractions were almost identical over the X-shaped or H-shaped composite river node. Furthermore, during the two flood events of 2006 and 2022, the difference in peak flood diversion over the X-shaped and the H-shaped river node was relatively small compared to the peak flood flow (e.g., over 40,000 (W) vs. 15,000 (N) m3/s) in either river, i.e., 6800 (X) vs. 6300 (H) m3/s and 6070 (X) vs. 5700 (H) m3/s, respectively, resulting in similar flood discharge in the downstream branches over the two river nodes (Figure 4b and Figure 6b,c).
For steady flow simulations of 121 flow combinations between the WR and NR, the discharge fractions were constant over the X-shaped river node while having a larger range over the H-shaped river node (Figure 9e,f). Over the X-shaped river node, the flow can be distributed more evenly and effectively during floods [4,16], resulting in larger diversion rates than those over the H-shaped river node (Figure 8c,d). This indicates that the geometric dimensions and roughness of the downstream branches primarily influenced the discharge fractions over the X-shaped river node. In contrast, at the H-shaped node, diversion flow in the SXJ waterway is constrained by both the imbalanced incoming flow (i.e., WL difference) and the geometric characteristics of the SXJ waterway and downstream branches.
Without flood diversion, the peak flood stages would increase by 0.96 m and 0.83 m at Makou and by 3.63 m and 3.35 m at Sanshui during the 2006 flood (Figure 5e,f). Similarly, during the 2022 flood, the peak stages would increase by 0.34 m and 0.30 m at Makou and by 2.98 m and 2.80 m at Sanshui at the X-shaped and H-shaped river nodes, respectively (Figure 7e,f). For extreme unbalanced incoming flow combinations and without flood diversion, the peak flood stages would increase by 1.65 m and 1.48 m at Makou and by 7.50 m and 6.05 m at Sanshui over the X-shaped and H-shaped river nodes, respectively (Figure 10).
Flood diversion not only mitigated the flood stages in the downstream branches (Figure 10) but also reduced the WL differences over the SXJ node and the upstream rivers (Figure 11). Without flood diversion, the WL differences between the two rivers would range from −4.50 m to 4.90 m and from −4.46 m to 5.14 m, respectively. However, with flood diversion, these differences were reduced to 0.02 m and 0.29 m over the X-shaped and H-shaped river nodes during the 2006 flood (Figure 11a,b). The reduction of WL differences over the SXJ node also reduced the WL difference up to 1.20 m at the 8 km upper away from the SXJ waterway. In addition, the flood diversion from the WR to the NR resulted in an increase in the peak WLs at Shijiao, indicating a significant backwater effect in the NR [31]. Meanwhile, the effect of flood diversion from the WR on the downstream NR involves additional flood diversions in the lower-order river network nodes, sluice gate operations, and tidal forces, which is beyond this study and needs further investigation.
In contrast to artificial, one-way flood diversion channels within river networks [32,33], the mutual flood diversion at the SXJ node facilitates synchronization of downstream discharge, significantly reducing the peak flood stages and the WL difference (i.e., flood hazards) between the two rivers. This interaction promotes hydraulic equilibrium at the SXJ node and ensures the maintenance of a nearly constant discharge fraction in the downstream branches [12,31,34]. At present, the flood diversion capacity of the H-shaped SXJ node could well redistribute the 1:100 years flood water for the NR, and any reduction of it could increase the flood risk over the SXJ node and the downstream branches, especially for the NR, thus being proscribed.

4.3. Implications of Results

The findings presented above underscore the significant influence of local factors [1], namely the X-shaped and H-shaped nodes, on flood diversion dynamics. Both nodes can divert asynchronous and unbalanced flood waves from upstream channels to synchronous flood discharges and maintain nearly constant discharge fractions in downstream branches. However, the mechanisms and effects of flood diversion differ between the X-shaped and H-shaped nodes.
In the X-shaped node, the river with a larger flood first partially converges and then diverges to another river without any constraint. The flow is freely redistributed between the interconnected rivers or channels, resulting in minimal or no water level difference between them. Conversely, in the H-shaped node, the river with higher flood waves first diverges into the lateral channel, which then converges to another river with lower flood. The water exchange is constrained by the flow resistance in the connecting branch, resulting in significant WL differences between the two rivers [12], particularly when one river experiences flooding. This change in WL difference has a significant effect on the distribution of floods in the downstream channels.
Furthermore, the effect of flood diversion differs between X-shaped and H-shaped nodes. Higher diversion rates can be accommodated more effectively in X-shaped nodes due to the greater space available in the connecting branch, resulting in lower flow resistance. Consequently, understanding the differences between X-shaped and H-shaped nodes can inform water resource allocation strategies during dry seasons and guide flood diversion practices during flood periods. These insights provide valuable guidance for river management in regions with similar river morphologies, providing a basis for optimizing water resource utilization and enhancing flood mitigation strategies.

5. Conclusions

This study investigated the mechanisms and impacts of flood diversion over the X-shaped and H-shaped river nodes using the Delft3D-Flow model. Simulations of two flood events (2006 and 2022) and 121 flood scenarios were conducted to analyze the flood diversion mechanism and effects at the SXJ node. The key conclusions are as follows:
(1)
The X-shaped and H-shaped river nodes show distinct mechanisms in flood diversion. Over the X-shaped node, the river with a larger flood first partially converges and then diverges to another river without a few constraints. In contrast, over the H-shaped node, the river with a higher flood first diverges into the lateral channel, which then converges to another river with a lower flood. The diversion capability is constrained not only by the relative magnitude of flood waves from the upstream rivers but also by the geometrical dimensions and roughness of the lateral waterway and downstream branches.
(2)
Both composite river nodes create a self-adaptive mode for mutual diversion and reduce the WL difference between both rivers over the SXJ node and the peak flood stages in both rivers, thus greatly reducing the flood hazards not only in the downstream branches but also over the SXJ node and in the upstream rivers.
(3)
Flood diversion over the SXJ node synchronizes the asynchronous upstream flood waves and leads to nearly constant discharge fractions in the downstream branches. The discharge fractions at Makou (Sanshui) fluctuated around 75.8% (24.2%) and 76.6% (23.4%) for the X-shaped and H-shaped nodes, respectively, which is crucial for optimizing flood mitigation strategies and hydraulic infrastructure planning in the PRD.

Author Contributions

Conceptualization, Y.F. and X.W.; Methodology, Y.F. and Y.N.; Software, P.Y.; Validation, H.L.; Investigation, X.W.; Resources, J.R.; Data curation, J.R.; Writing—original draft, Y.F. and X.W.; Visualization, P.Y. and Y.N.; Supervision, H.L.; Funding acquisition, X.W. All authors have read and agreed to the published version of the manuscript.

Funding

This study is funded by the National Key R&D Program of China (No. 2021YFC3001000), the Innovation Group Project of Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) (No. 311021004), and the Zhejiang University of Water Resources and Electric Power Research Fund (No. 88106324008). We also greatly appreciate the WL, stream flow, and bathymetry data provided by the Pearl River Hydraulic Research Institute and Department of Water Resources (DWR) of Guangdong Province, as well as the Deft3D model developed by the Deltares in the Netherlands. In addition, all the authors thank the editors and two anonymous referees for their constructive comments and suggestions.

Data Availability Statement

The research data are confidential, and the authors are not authorized to share them.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Kleinhans, M.G.; Jagers, H.R.A.; Mosselman, E.; Sloff, C.J. Bifurcation dynamics and avulsion duration in meandering rivers by one-dimensional and three-dimensional models. Water Resour. Res. 2008, 44, W08454. [Google Scholar] [CrossRef]
  2. Davoren, A.; Mosley, M.P. Observations of bedload movement, bar development and sediment supply in the braided Ohau River. Earth Surf. Process. Landf. 1986, 11, 643–652. [Google Scholar] [CrossRef]
  3. Ferguson, R.I.; Ashmore, P.E.; Ashworth, P.J.; Paola, C.; Prestegaard, K.L. Measurements in a braided river chute and lobe: 1. Flow pattern, sediment transport, and channel change. Water Resour. Res. 1992, 28, 1877–1886. [Google Scholar] [CrossRef]
  4. Ferguson, R.I. Understanding braiding processes in gravel-bed rivers: Progress and unsolved problems. Geol. Soc. Lond. Spéc. Publ. 1993, 75, 73–87. [Google Scholar] [CrossRef]
  5. Best, J.L. Flow dynamics at river channel confluences: Implications for sediment transport and bed morphology. In Recent Developments in Fluvial Sedimentology; Ethridge, F.G., Flores, R.M., Harvey, M.D., Eds.; SEPM: Tulsa, OK, USA, 1987; Volume 39, pp. 27–35. [Google Scholar] [CrossRef]
  6. Best, J.L.; Roy, A.G. Mixing-layer distortion at the confluence of channels of different depth. Nature 1991, 350, 411–413. [Google Scholar] [CrossRef]
  7. Ashmore, P.E.; Ferguson, R.I.; Prestegaard, K.L.; Ashworth, P.J.; Paola, C. Secondary flow in anabranch confluences of a braided, gravel-bed stream. Earth Surf. Process. Landf. 1992, 17, 299–311. [Google Scholar] [CrossRef]
  8. Wang, Z.B.; De Vries, M.; Fokkink, R.J.; Langerak, A. Stability of river bifurcations in 1D morphodynamic models. J. Hydraul. Res. 1995, 33, 739–750. [Google Scholar] [CrossRef]
  9. Kleinhans, M.G.; Ferguson, R.I.; Lane, S.N.; Hardy, R.J. Splitting rivers at their seams: Bifurcations and avulsion. Earth Surf. Process. Landforms 2013, 38, 47–61. [Google Scholar]
  10. Bolla Pittaluga, M.; Repetto, R.; Tubino, M. Channel bifurcation in braided rivers: Equilibrium configurations and stability. Water Resour. Res. 2003, 39, 1046. [Google Scholar] [CrossRef]
  11. Bertoldi, W.; Tubino, M. River bifurcations: Experimental observations on equilibrium configurations. Water Resour. Res. 2007, 43, W10437. [Google Scholar] [CrossRef]
  12. Fang, Y.; Wang, X.; Ren, J.; Liu, H.; Wang, Y. Driving Forces and Influences of Flood Diversion on Discharge Fraction and Peak Water Levels at an H-Shaped Compound River Node in the Pearl River Delta, South China. Water 2023, 15, 1970. [Google Scholar] [CrossRef]
  13. Yang, T.; Xu, C.; Shao, Q.; Chen, X. Regional flood frequency and spatial patterns analysis in the Pearl River Delta region using L-moments approach. Stoch. Environ. Res. Risk Assess. 2010, 24, 165–182. [Google Scholar] [CrossRef]
  14. Zeng, Z. Evolution of Sixianjiao waterway. J. South China Norm. Univ. (Nat. Sci. Ed.) 1979, 1, 52–74. (In Chinese) [Google Scholar]
  15. Zeng, Z.; Huang, S. Study on the historical geomorphology of the convergence area of branching channels in the Pearl River Delta, taking the development of a sandbar in Sixianjiao as an example. Pearl River 1982, 4, 25–29. (In Chinese) [Google Scholar]
  16. Ying, Z.; Chen, Z.; So, C.L. Formation and Evolution of the X-Shape waterways by Sixianjiao Channel. Sun Yat-Sen Univ. Forum 1988, 2, 8–14. (In Chinese) [Google Scholar]
  17. Zhang, W.; Du, J.; Zheng, J.; Wei, X.; Zhu, Y. Redistribution of the suspended sediment at the apex bifurcation in the Pearl River Network, South China. J. Coast. Res. 2014, 293, 170–182. [Google Scholar] [CrossRef]
  18. Wu, Y.; He, Y.; Wu, M.; Lu, C.; Gao, S.; Xu, Y. Multifractality and cross-correlation analysis of streamflow and sediment fluctuation at the apex of the Pearl River Delta. Sci. Rep. 2018, 8, 16553. [Google Scholar] [CrossRef]
  19. Wu, Y.; Zhang, W.; Zhu, Y.; Zheng, J.; Ji, X.; He, Y.; Xu, Y. Intra-tidal division of flow and suspended sediment at the first order junction of the Pearl River Network. Estuar. Coast. Shelf Sci. 2018, 209, 169–182. [Google Scholar] [CrossRef]
  20. Liu, F.; Xie, R.; Luo, X.; Yang, L.; Cai, H.; Yang, Q. Stepwise adjustment of deltaic channels in response to human interventions and its hydrological implications for sustainable water managements in the Pearl River Delta, China. J. Hydrol. 2019, 573, 194–206. [Google Scholar] [CrossRef]
  21. Wang, B.; Yang, H.; Ou, S.; Fu, L.; Cai, H.; Yang, Q. Water Surface Profile Dynamics and Underlying Mechanism of Transverse Channel in Pearl River Channel Networks: A case Study of the Dongping Channel. Trop. Geogr. 2021, 41, 410–422. (In Chinese) [Google Scholar]
  22. Liu, Q.; Wu, C. Hydrodynamic Characteristics of Waterway Network of the Pearl River Delta in the 1950’s. Port Waterw. Eng. 2005, 3, 66–69. (In Chinese) [Google Scholar]
  23. Xie, P.; Tang, Y.; Chen, G.; Li, C. Variation analysis of hydrological and sediment series in North River and West River Delta: Case study of Makou Station and Sanshui Station. J. Sediment Res. 2010, 5, 26–31. (In Chinese) [Google Scholar]
  24. Wang, X.; Guo, Y.; Ren, J. The Coupling Effect of Flood Discharge and Storm Surge on Extreme Flood Stages: A Case Study in the Pearl River Delta, South China. Int. J. Disaster Risk Sci. 2021, 12, 1–15. [Google Scholar] [CrossRef]
  25. Deltares. Delft3D-FLOW. User Manual 3.15.57696; Deltares: Delft, The Netherlands, 2018. [Google Scholar]
  26. Lesser, G.R.; Roelvink, J.V.; van Kester, J.T.M.; Stelling, G.S. Development and validation of a three-dimensional morphological model. Coast. Eng. 2004, 51, 883–915. [Google Scholar] [CrossRef]
  27. Department of Water Resources (DWR) in Guangdong Province. Manuals for the Design Flood Stages in the Downstream of Xijiang and Beijiang and in the River Networks in the Pearl River Delta; Internal Document Issued; Department of Water Resources in Guangdong Province: Guangzhou, China, 2002. (In Chinese)
  28. Buschman, F.A.; Hoitink, A.J.F.; van der Vegt, M.; Hoekstra, P. Subtidal flow division at a shallow tidal junction. Water Resour. Res. 2010, 46, W12515. [Google Scholar] [CrossRef]
  29. Edmonds, D.A. Stability of backwater-influenced river bifurcations: A study of the Mississippi-Atchafalaya system. Geophys. Res. Lett. 2012, 39, L08402. [Google Scholar] [CrossRef]
  30. Kästner, K.; Hoitink, A.J.F. Flow and suspended sediment division at two highly asymmetric bifurcations in a river delta: Impli-cations for channel stability. J. Geophys. Res. Earth Surf. 2019, 124, 2358–2380. [Google Scholar]
  31. Liu, J. The Pearl River Delta Sixianjiao Channel, Tianhe Node Split Ratio on Law. Pearl River 2016, 37, 15–20. (In Chinese) [Google Scholar]
  32. Mel, R.A.; Viero, D.P.; Carniello, L.; D’Alpaos, L. Multipurpose Use of Artificial Channel Networks for Flood Risk Re-duction: The Case of the Waterway Padova–Venice (Italy). Water 2020, 12, 1609. [Google Scholar] [CrossRef]
  33. Mel, R.A.; Viero, D.P.; Carniello, L.; D’Alpaos, L. Optimal floodgate operation for river flood management: The case study of Padova (Italy). J. Hydrol. Reg. Study 2020, 30, 100702. [Google Scholar] [CrossRef]
  34. Li, T. The role of Sixianjiao in the July 1996 flood. Guangdong Water Resour. Hydropower 1997, 1, 22–24. (In Chinese) [Google Scholar]
Water 17 01015 g001
Figure 1. The SXJ node (red rectangle) and the river network in the PRD, South China. Black triangles indicate water level (WL) stations. The eight outlets are labeled as HUM (Humen), JM (Jiaomen), HQM (Hongqimen), HM (Hengmen), MDM (Modaomen), JTM (Jitimen), HTM (Hutiaomen), and YM (Yamen).
Figure 1. The SXJ node (red rectangle) and the river network in the PRD, South China. Black triangles indicate water level (WL) stations. The eight outlets are labeled as HUM (Humen), JM (Jiaomen), HQM (Hongqimen), HM (Hengmen), MDM (Modaomen), JTM (Jitimen), HTM (Hutiaomen), and YM (Yamen).
Water 17 01015 g001
Water 17 01015 g002
Figure 2. The schematic evolution of the SXJ node in the PRD across different periods: (a) Song Dynasty, (b) Ming Dynasty, (c) Qing Dynasty, and (d) The year 2023. The black triangles represent water level (WL) stations.
Figure 2. The schematic evolution of the SXJ node in the PRD across different periods: (a) Song Dynasty, (b) Ming Dynasty, (c) Qing Dynasty, and (d) The year 2023. The black triangles represent water level (WL) stations.
Water 17 01015 g002
Water 17 01015 g003
Figure 3. A flowchart for the study.
Figure 3. A flowchart for the study.
Water 17 01015 g003
Water 17 01015 g004
Figure 4. Hydrodynamic grids and topography of (a) the X-shaped and (b) H-shaped SXJ nodes. Red-filled circles indicate junctions—west mouth, east mouth, upper north mouth, and upper west mouth—where WLs are analyzed to assess flood diversion impacts. The upper west and upper north mouths are located approximately 8 km upstream of the west and east mouths, respectively. The black triangles represent water level (WL) stations.
Figure 4. Hydrodynamic grids and topography of (a) the X-shaped and (b) H-shaped SXJ nodes. Red-filled circles indicate junctions—west mouth, east mouth, upper north mouth, and upper west mouth—where WLs are analyzed to assess flood diversion impacts. The upper west and upper north mouths are located approximately 8 km upstream of the west and east mouths, respectively. The black triangles represent water level (WL) stations.
Water 17 01015 g004
Water 17 01015 g005
Figure 5. The simulated flow and WLs during the July 2006 flood event at the X-shaped and H-shaped SXJ river nodes. Panels (a,b) present the inflow and outflow rates, along with the flow fractions of the WR and NR, where the flow fraction represents the proportion of each river’s contribution to the total flow. Panels (c,d) illustrate the diversion flow rates and the WL differences between the east and west mouths of the SXJ waterway. Panels (e,f) display WL changes due to flood diversion, calculated as the difference between actual WLs with and without diversion at the X-shaped and H-shaped nodes, respectively, while the vertical dashed line divides the flood event into different phases.
Figure 5. The simulated flow and WLs during the July 2006 flood event at the X-shaped and H-shaped SXJ river nodes. Panels (a,b) present the inflow and outflow rates, along with the flow fractions of the WR and NR, where the flow fraction represents the proportion of each river’s contribution to the total flow. Panels (c,d) illustrate the diversion flow rates and the WL differences between the east and west mouths of the SXJ waterway. Panels (e,f) display WL changes due to flood diversion, calculated as the difference between actual WLs with and without diversion at the X-shaped and H-shaped nodes, respectively, while the vertical dashed line divides the flood event into different phases.
Water 17 01015 g005
Water 17 01015 g006
Figure 6. The water level (WL) simulated over the X-shaped (a,b) and H-shaped (c,d) river nodes at 12:00 on 18 July (a,c) and 23 (b,d), 2006.
Figure 6. The water level (WL) simulated over the X-shaped (a,b) and H-shaped (c,d) river nodes at 12:00 on 18 July (a,c) and 23 (b,d), 2006.
Water 17 01015 g006
Water 17 01015 g007
Figure 7. Simulated flow and WLs during the June 2022 flood event at the X-shaped and H-shaped SXJ nodes. Panels (a,b) show the inflow and outflow rates along with WR and NR flow fractions. Panels (c,d) depict the diversion flow rates and the WL differences between the east and west mouths of the SXJ waterway. Panels (e,f) illustrate the WL changes due to flood diversion, derived by subtracting the WLs without diversion from those actual WLs with diversion at the X-shaped and H-shaped nodes, respectively, while the vertical dashed line divides the flood event into different phases.
Figure 7. Simulated flow and WLs during the June 2022 flood event at the X-shaped and H-shaped SXJ nodes. Panels (a,b) show the inflow and outflow rates along with WR and NR flow fractions. Panels (c,d) depict the diversion flow rates and the WL differences between the east and west mouths of the SXJ waterway. Panels (e,f) illustrate the WL changes due to flood diversion, derived by subtracting the WLs without diversion from those actual WLs with diversion at the X-shaped and H-shaped nodes, respectively, while the vertical dashed line divides the flood event into different phases.
Water 17 01015 g007
Water 17 01015 g008
Figure 8. The water level (WL) simulated over the X-shaped (a,b) and H-shaped (c,d) nodes on June 10 (a,c) and 23 (b,d), 2022.
Figure 8. The water level (WL) simulated over the X-shaped (a,b) and H-shaped (c,d) nodes on June 10 (a,c) and 23 (b,d), 2022.
Water 17 01015 g008
Water 17 01015 g009
Figure 9. Simulated flood diversion and discharge fractions for 121 combinations based on incoming flows at the X-shaped (a,c,e) and H-shaped (b,d,f) river nodes. The X-axis is the inflow at Gaoyao from the WR, while various colors are the inflow at Shijiao from the NR. Panels (a,b) show the WL differences between the WR and NR at the SXJ node. Panels (c,d) illustrate the flood diversion at Ganggen, located at the midpoint of the SXJ waterway. Panels (e,f) display discharge fractions at Makou and Sanshui, with the dashed line indicating threshold fractions when the flow within SXJ waterway approaches negligible levels.
Figure 9. Simulated flood diversion and discharge fractions for 121 combinations based on incoming flows at the X-shaped (a,c,e) and H-shaped (b,d,f) river nodes. The X-axis is the inflow at Gaoyao from the WR, while various colors are the inflow at Shijiao from the NR. Panels (a,b) show the WL differences between the WR and NR at the SXJ node. Panels (c,d) illustrate the flood diversion at Ganggen, located at the midpoint of the SXJ waterway. Panels (e,f) display discharge fractions at Makou and Sanshui, with the dashed line indicating threshold fractions when the flow within SXJ waterway approaches negligible levels.
Water 17 01015 g009
Water 17 01015 g010
Figure 10. The WL changes at Makou and Sanshui due to flood diversion at the X-shaped node (a) and H-shaped node (b). Red and blue indicate WL increases at Makou and Sanshui, respectively, while magenta and green represent WL decreases. The arrow indicates the direction of flood diversion, where W → N signifies flow from WR to NR, and W ← N denotes flow from NR to WR.
Figure 10. The WL changes at Makou and Sanshui due to flood diversion at the X-shaped node (a) and H-shaped node (b). Red and blue indicate WL increases at Makou and Sanshui, respectively, while magenta and green represent WL decreases. The arrow indicates the direction of flood diversion, where W → N signifies flow from WR to NR, and W ← N denotes flow from NR to WR.
Water 17 01015 g010
Water 17 01015 g011
Figure 11. WL differences (NR–WR) between the NR and the WR at 8 km upstream (blue) and the SXJ waterway (orange solid) during the 2006 (a,b) and 2022 (c,d) floods, with flood diversion at the X-shaped (a,c) and H-shaped (b,d) river nodes. The green dotted line represents the WL differences at the SXJ waterway without flood diversion.
Figure 11. WL differences (NR–WR) between the NR and the WR at 8 km upstream (blue) and the SXJ waterway (orange solid) during the 2006 (a,b) and 2022 (c,d) floods, with flood diversion at the X-shaped (a,c) and H-shaped (b,d) river nodes. The green dotted line represents the WL differences at the SXJ waterway without flood diversion.
Water 17 01015 g011
Table 1. The incoming flow at Gaoyao and Shijiao in the simulations.
Table 1. The incoming flow at Gaoyao and Shijiao in the simulations.
Rivers/StationsFlow (×103 m3/s)
West River (WR)/
Gaoyao		1015202530354045505560
North River (NR)/
Shijiao		246810121416182022
Table 2. Flood flow and return periods for Gaoyao and Feilaixia [27].
Table 2. Flood flow and return periods for Gaoyao and Feilaixia [27].
River/StationsReturn Periods (1:n Year) and Flow Rates (×103 m3/s)
5 10 20 30 50 100 200 300
West River (WR)/Gaoyao37.945.049.750.852.254.055.957.5
North River (NR)/Feilaixia11.913.815.516.717.719.220.721.6
Note: Shijiao is situated 40 km downstream of Feilaixia.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Fang, Y.; Wang, X.; Ren, J.; Liu, H.; Yuan, P.; Ning, Y. Distinct Flood Diversion Mechanisms and Comparable Effects on Discharge Fraction and Peak Water Levels over X-Shaped and H-Shaped Composite River Nodes. Water 2025, 17, 1015. https://doi.org/10.3390/w17071015

AMA Style

Fang Y, Wang X, Ren J, Liu H, Yuan P, Ning Y. Distinct Flood Diversion Mechanisms and Comparable Effects on Discharge Fraction and Peak Water Levels over X-Shaped and H-Shaped Composite River Nodes. Water. 2025; 17(7):1015. https://doi.org/10.3390/w17071015

Chicago/Turabian Style

Fang, Yongjun, Xianwei Wang, Jie Ren, Huan Liu, Peiqing Yuan, and Yazhou Ning. 2025. "Distinct Flood Diversion Mechanisms and Comparable Effects on Discharge Fraction and Peak Water Levels over X-Shaped and H-Shaped Composite River Nodes" Water 17, no. 7: 1015. https://doi.org/10.3390/w17071015

APA Style

Fang, Y., Wang, X., Ren, J., Liu, H., Yuan, P., & Ning, Y. (2025). Distinct Flood Diversion Mechanisms and Comparable Effects on Discharge Fraction and Peak Water Levels over X-Shaped and H-Shaped Composite River Nodes. Water, 17(7), 1015. https://doi.org/10.3390/w17071015

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop