Impact of Partially Covered Vegetation on the Lateral Velocity Distribution of Open Channel Flow ()
1. Introduction
Vegetation widely exists in natural rivers. Vegetation plays an important role in the river environment since the vegetation provides a suitable environment for habitat creation and biodiversity along with the improvement of water quality and reduction of bank erosion, and so on (Nepf & Ghisalberti, 2008; Greet et al., 2011; Zhang et al., 2020). Various types of vegetation grow along natural rivers, and sometimes they are planted for the purpose of engineering or ecological requirement (Chembolu et al., 2019). The riparian vegetation retards flow and increases the flow resistance caused by the additional drag from the vegetation. Consequently, the vegetation affects the velocity, Reynolds stress, and turbulence intensity (Nepf & Vivoni, 2000; Lopez & Garcia, 2001; Tang & Knight, 2001, 2009; Tang et al., 2010, 2011; Zhao & Huai, 2016; Tang, 2018, 2019a, 2019b).
Previous studies on vegetated flow were mainly on the velocity and resistance of flow (Carollo et al., 2002; Stone & Shen, 2002; Tang & Knight, 2009; Tang & Ali, 2013), where the single layer vegetation was simulated by artificial cylindrical dowels of rigid or flexible type in laboratory flumes and in either emergent or submerged flow conditions (Tang et al., 2010, 2011; Yang et al., 2020; Yan et al., 2020). With the growing role of vegetation in the river ecological environment, much attention has been paid to understand the flow interactions with vegetation and the physical processes at various scales (Curran & Hession, 2013; Nezu & Sanjou, 2008; Nepf, 2012).
The characteristics of flow through vegetation may be interpreted using different mechanisms over the flow depth (Nikora et al., 2013; Huai et al., 2014; Rahami et al., 2020).Thus, the velocity profile can be modelled separately in a layer with each phenomenon described (Tang & Ali, 2013; Tang, 2018, 2019a, 2019b; Singh et al., 2019). The flow structure in vegetated channels has also been investigated through numerical modelling (e.g., Lopez & Garcia, 2001; Neary, 2003; Zeng & Li, 2014) and CFD simulation using FLUENT (Souliotis & Prinos, 2011; Anjum et al., 2018; Anjum & Tasnka, 2019; Rahimi et al., 2019).
In riparian environments, there exists multiple-layered vegetation such as grasses, shrubs, trees. Shorter vegetation is often submerged while the tall vegetation is emergent in high flow conditions. Thus, the flow structure becomes very complicated owing to the interaction between the flow and different-layered vegetation. To understand the influence of multiple-layered vegetation on the flow structure, some laboratory studies were carried out in an open-channel with the bed fully covered by a mixing array of short and tall vegetation (e.g., Liu et al., 2008; Anjum et al., 2018; Tang et al.. 2019; Rahimi et al.. 2019, 2020). Most recently, Tang et al. (2018, 2019, 2021) have conducted experiments on the flow with double-layer vegetation that covered one side of a channel bed. However, there is little study about the impact of non-evenly distributed multi-layered vegetation along two sides of a channel, which commonly exists in rivers. This knowledge gap becomes the aim of this paper.
This paper presents the experimental results of lateral velocity distribution in an open channel: one side of the bed is covered with single-layered vegetation whilst the other side is with two-layered vegetation in a linear pattern. The velocities at different locations were measured by ADV (Acoustic Doppler Velocimetry), aiming at investigating how the vegetation affects the lateral velocity of flow when the short vegetation is under submerged and emergent conditions.
2. Experimental Setting
The experiment was conducted in the 20 m-long titling flume at XJTLU (Xi’an Jiaotong-Liverpool University). The flume has a rectangular cross-section of 0.4 m wide by 0.5 m high and is set at a bed slope (So) of 0.003. The vegetation was simulated by 6.35 mm circular plastic dowels in two heights of 10 and 20 cm, denoting the short and tall vegetation, respectively. Both the short and tall dowels are arranged in a linear pattern with a spacing of 31.75 mm between the centres of dowels. The flume is sketched in Figure 1, where two different vegetation arrangements are along two sides of the bed (see Figure 2). Only tall dowels are in vegetation region 1 (left), with the distance of the nearest dowel to wall A being 25.38 mm. In vegetation region 2 (right), there are two rows of short dowels near the free region and two rows of tall dowels near wall B. Thus, each vegetation zone has the same width as the free region, i.e., one-third of the channel width.
![]()
Figure 2. The arrangement of vegetation array and measurement locations.
The measured locations are coded as follows (see Figure 2): BT and BS denote the measurement locations behind the tall dowel and the short dowel, respectively, while FR denotes the free region (i.e., the central zone without vegetation). The other notations are that BST = behind and side away from the tall dowel, NT = the location next to the tall dowel, NS = the location next to the short dowel, and NST = the location next to both short and tall dowels.
Two types of Nortek micro-ADV (downward- and side-probes) were used to measure velocity at various measurement locations in a cross-section. For most measurements, the downward-probe ADV was used to obtain 3D velocities in a vertical except for the 5 cm zone near the water surface, where the side-probe ADV was used instead. The sampling time of each measurement was set as 60 seconds. The WinADV software was used to process the velocity data of ADV. In the experiment, two flow depths of 9 and 14 cm were undertaken. The corresponding discharge was 6.1 and 11.1 L/s, representing the following two flow conditions: all dowels are emergent, the short dowels are submerged while tall dowels are emergent.
3. Results and Discussion
In the subsequent figures, the velocity is normalized by the cross-sectional mean velocity U. The vertical distance (z) above the channel bed is normalized by the height of short vegetation (h).
3.1. Lateral Change of Velocity Profiles
To understand the change of profiles in the different regions (the vegetation and non-vegetation region), the comparisons of velocity profiles at the representative locations (i.e., at locations BST, FR, BSS and BSTS) are shown in Figure 3 and Figure 4 for the cases of 9 cm and 14 cm, respectively.
Figure 3 and Figure 4 clearly reveal that the velocities in the free region (FR) are much higher than those in the vegetation region, indicating that the vegetation has a considerably retarding effect on the velocity. Furthermore, in the emergent case (Figure 3), the velocity profiles are similar for various locations
![]()
Figure 3. Lateral variation of velocity profiles for the flow depth of 9 cm.
![]()
Figure 4. Lateral change of velocity profiles for flow depth of 14 cm.
(BST, BSS and BSTS) within the vegetation region. However, the velocities tend to show some differences near the top of short vegetation depending on the locations for the partially submerged case (Figure 4), where the velocity between short dowels (BSS) has the highest value while the velocity between tall dowels (BST) is the lowest. At the locations of BSS and BSTS (i.e. short vegetation in submerged conditions), their velocities start to increase from certain distances (z/h at about 0.6) below the top of short vegetation, indicating a penetration depth due to the strong shear between the upper free flow and lower vegetated flow (Nepf & Vivoni, 2000; Tang et al., 2021).
3.2. Lateral Distribution of Depth-Averaged Velocity
To show the impact of different vegetation configurations on the lateral distribution of velocity, the depth-averaged velocity (ud) is calculated and shown in Figure 5, where y is the lateral distance from wall A. Figure 5 shows that the depth-averaged velocity increases rapidly around the interface between the vegetation region (1 or 2) and the free region. This implies that a strong momentum exchange occurs near the interface between the vegetation and no-vegetation regions. The large lateral velocity gradient at the interface is caused by the velocity difference between the two regions, where the flow in the free region (i.e. center) is faster than that in the vegetation region where the slow-moving flow is due to the additional resistance of vegetation. These findings are similar to those observed by Tang et al. (2019, 2021).
Besides, as the water depth increases, when the short vegetation is fully submerged but the tall vegetation is emergent (case H = 14 cm), the vegetation will reduce the relative velocity of the free region more than in the vegetation region, resulting in a smaller gradient of lateral velocity in the transition layer between the vegetation and non-vegetation regions. This result indicates that the influence of depth on the resistance in the vegetated region is relatively smaller compared with that in the free region (Tang et al., 2019, 2021).
Further examination on the lateral velocity distribution of 14 cm case shows that the velocity gradient in the transition zone between the free region and region 2 (i.e., two-layered vegetation) becomes larger than that between the free region and region 2 (tall vegetation). This effect may be due to the relatively increasing resistance of flow in the submerged short vegetation, which has additional strong vertical mixing.
3.3. Discharge in Each Region
The discharge in each region can be calculated based on the lateral distribution of depth-averaged velocity. Figure 6 shows the percentage of discharge in each region for the two cases. The discharge in the free region is about 61% - 63.5% of the total discharge, where it is nearly the same for each vegetation region, although the free region takes only one third of the channel width. With increasing water depth, the discharge percentage in the free region slightly decreases whereas it increases in both vegetation regions. This result may be caused by the relative submergence of vegetation, because of reduced resistance from the short vegetation in region 2 changes from emergent to submerged condition when flow depth changes from 9 cm to 14 cm.
In general, the discharge percentage through the entire vegetation region slightly increases from 39% to 37% as the depth of flow increases from 9 cm to 14 cm, i.e. from the emergent to partially submerged conditions.
![]()
Figure 5. Lateral distribution of depth-averaged velocity ud (B = 40 cm).
![]()
Figure 6. Discharge percentage in each region for three flow depths: 9 cm and 14 cm.
4. Conclusion
Based on the novel experimental study on the open-channel with two bedsides covered with different vegetation patterns, the results show that the vegetation pattern will affect the lateral velocity distribution. The following points may be drawn:
• The velocity profiles are different laterally. The velocity profiles in the free region are much higher than those in vegetated regions. Although the vertical variation of velocity in all regions is small in the emergent conditions, the velocity starts to increase from a certain distance below the top of short vegetation when the flow depth increases to make the short vegetation under submergence.
• The depth-averaged velocities in the free region are much larger than in the vegetation regions, indicating that a strong momentum exchange exists between the vegetation and non-vegetation regions and that the presence of the vegetation has a noticeably retaining effect on the flow.
• As the flow depth increases, the vegetation will reduce the relative velocity in the free region more than in the vegetation region, resulting in a smaller gradient of lateral velocity in the transition layer between the vegetation and non-vegetation regions.
• In the free region, the velocity near its center is much larger than that near the vegetation regions.
Acknowledgements
The project was partly supported by the National Natural Science Foundation of China (11772270) and by the funds (KSF-E-17; RDF-16-02-02) of XJTLU.