Channel erosion and its impact on environmental flow of riparian habitat in the Middle Yangtze River

Abstract

Evaluating environmental flow (EF) is pivotal for conserving and restoring riverine ecosystems. Yet, prevalent EF evaluations presume that a river reach's hydraulic conditions are exclusively governed by inflow discharge, presupposing a state of equilibrium in the river channel. This presumption narrows the scope of EF evaluations in expansive alluvial rivers like the Middle Yangtze River (MYR), characterized by marked channel alterations. Here we show the profound channel erosion process and its impact on EF requirements for riparian habitats within the MYR. Our research unveils that: (i) pronounced erosion has led to a mean reduction of 1.0–2.7 m in the riverbed across four sub-reaches of the MYR; (ii) notwithstanding a 37–107% increase in minimal discharges post the Three Gorges Project, the lowest river stages at some hydrometric stations diminished owing to bed erosion, signifying a notable transformation in MYR's hydraulic dynamics; (iii) a discernible rightward shift in the correlation curve between the weighted useable area and discharge from 2002 to 2020 in a specific sub-reach of the MYR, instigated by alterations in hydraulic conditions, necessitated an increase of 1500–2600 m³ s⁻¹ in the required EF for the sub-reach; (iv) it is deduced that macroinvertebrate biomass rapidly decreases as the flow entrains the riverbed substrate, with the maximum survivable velocity for macroinvertebrates being contingent on their entrainment threshold. These findings highlight the importance of incorporating channel morphological changes in devising conservation strategies for the MYR ecosystem.

Graphical abstract

Highlights

• •Channel erosion lowers the Middle Yangtze River bed by 1.0–2.7 m.
• •The lowest river stage decreases with a 37–107% increase in minimal discharge.
• •Bed lowering necessitates a 1500–2600 m³ s⁻¹ increase in required environment flow.

Abbreviations

EF

Environment flow refers to the discharge required by aquatic creatures (m³ s⁻¹)

E

Erosion intensity (m³ km⁻¹ a⁻¹)

d₅₀

Medium diameter of bed substrate/material (mm)

d

Sediment diameter (mm)

D_m

Density of macroinvertebrates (ind m⁻²)

h

Water depth (m)

HSI

Habitat suitability index

OEF

Optimal environment flow with a discharge range [Qo₀, Qo₁] (m³ s⁻¹)

ΔOEF

Uncertainty of OEF, represented by the variation of Qo₁ (m³ s⁻¹)

Q

Water discharge (m³ s⁻¹)

Q_min

Lowest water discharge (m³ s⁻¹)

Q_h

Water discharge at which WUA_P is 100% (m³ s⁻¹)

TGP

Three Gorges Project

TEF

Threshold environment flow with a discharge range of [Q_T0, Q_T1] (m³ s⁻¹)

ΔTEF

Uncertainty of TEF, represented by the variation of Q_T1 (m³ s⁻¹)

u_cs

Sediment entrainment velocity (m s⁻¹)

u

Flow velocity (m s⁻¹)

WUA

Weighted usage area (m²)

WUA_max

Maximum value of WUA (m²)

WUA_P

The proportion of WUA to WUA_max (%) and the subscript P denotes the proportion

WUA-Q_LFB, WUA-Q_RTB

Obtained WUA_P-Q curves using the left and right boundaries of the Q-Z data, respectively

Z

River stage (m)

ΔZ

Difference of river stage (m)

Z_min

Lowest river stage (m)

u_cm

Critical velocity for disappear of macroinvertebrates (m s⁻¹)

α, β, k, m, k₁

Empirical coefficients

1. Introduction

The aquatic ecosystem is pivotal to sustaining our livelihoods, yet the intricate interplay between the demands for water and energy resources and their conservation poses significant challenges to decision-making and practical measures in water allocation and regulation [[1], [2], [3], [4], [5]]. In the 21st century, the deliberate and strategic design of resilient ecosystems, including freshwaters, is recognized as a major social scientific challenge [6]. Intensive efforts and remarkable progress have been made, including developing ecological goals and management standards that can be applied globally to streams and rivers across a spectrum of ecological, social, political, and governance contexts. For example, a new and flexible framework for developing regional environmental flow standards was proposed by Poff et al. [7], which synthesized several existing hydrological techniques and environment flow methods. However, translating general principles and knowledge into specific management rules for specified river systems remains challenging [8].

For the Yangtze River basin, Wang et al. [9] pointed out the major threats to the Yangtze River floodplain ecosystem and proposed two key scientific questions for biodiversity conservation based on their twenty years of efforts. Their second question is: what is the environment flow requirement of the Yangtze River floodplain? However, as noted by Poff et al. [7], the physical setting of a river segment will strongly influence how the flow regime is translated into the hydraulic habitats available to the riverine biota. Therefore, there should be another question that needs to be asked following the previous question of Wang et al. [9]: how would the channel evolution influence the environment flow? The impacts of channel evolution on the ecosystem have not been systematically investigated, even though channel erosion in the Middle Yangtze River (MYR) has been reported to directly impact the habitat condition on the main stem [10]. Arthington et al. [5] also pointed out that one of the major scientific challenges and areas for further exploration is quantifying/predicting the interaction between hydrology, sedimentary process, geomorphology, hydraulic, temperature, and ecological variables.

In terms of assessing the environment flow, numerous methods have been proposed since the 1940s, which could be categorized into four classes: hydrological rules, hydraulic rating methods, habitat simulation methods, and holistic methodologies [3,11]. Among those, the method of habitat simulation has been widely used [3,[12], [13], [14]]. Still, it has also been criticized for lacking biological meaning for the weighted useable area (WUA) and a thorough understanding of the feeding and predator avoidance concepts for the related species, etc. [15,16]. However, the habitat simulation serves as a good and simple tool to relate the physical process of a river system with the ecological process. This is very important to large river systems like the MYR, given that changes in the physical process of the river system will cause multifaceted concerns involving flood control management, navigation safety, riparian industry development, etc. The habitat simulation provides an opportunity to untangle the complicated relationships between the river flow, channel, and biota from a large-scale view, which will benefit the planning of the ecosystem conservation or restoration measures, such as designing an optimal ecologically-friendly dam operation strategy.

The major objectives of the current study are to (i) clarify the variation of the hydraulic conditions (particularly river stage and discharge) in the MYR caused by river bed erosion; (ii) identify the impacts of bed erosion on the requirement of environmental flow in the MYR, and (iii) discuss the relationship between river sediment entrainment with the survival of macroinvertebrates and the implications for the ecosystem reservation in the MYR.

2. Study area and data sources

2.1. Study reach

The Yangtze River originates from the Qingzang Plateau and empties into the East China Sea. It is geologically segmented into the upper, middle, and lower reaches. The MYR from Yichang to Hukou has a length of 955 km and includes four sub-reaches of Yizhi, Jingjiang, Chenghan, and Hanhu (Fig. 1a and b).

Fig. 1.

Sketches of Yangtze river basin (a) and the Jingjiang reach (b).

The upstream dam constructions have greatly altered the river hydrology, channel evolution, biodiversity, and riparian habitats in the MYR, resulting in intensive channel erosion, riparian habitat loss, and biodiversity loss [9,17]. Due to the operation of the Three Gorges Project (TGP) in June 2003, the maximum water discharge at the inlet of the MYR (Yichang) has decreased from 31,500 to 61,700 m³ s⁻¹ in 1990–2002 to 27,400–58,400 m³ s⁻¹ in 2003–2020, accompanied by a 90% decrease in sediment amount. This remarkable sediment reduction led to long-term bed erosion in the MYR, and the cumulative channel erosion in the MYR reached 2.6 billion m³ in 2002–2020, according to the statistics of the Changjiang Water Resources Commission (CWRC) [18]. Bed erosion is anticipated to persist by the end of this century [19], given that the sediment entering the TGP has also reduced greatly in recent years because of the operation of upstream cascade reservoirs and the implementation of soil conservation measures.

The MYR is also a cradle of aquatic animals with abundant vegetation-covered riparian habitats and floodplains [20]. With the obvious changes in the hydrological processes, channel evolution, and over-exploitation, the ecosystems of the MYR have degraded [9]. For example, the biomass of macroinvertebrates in 2016 reduced by more than 90% in the reach from Yichang to Anqing (including the MYR, Fig. 1a), compared with the value in 1987 [21]. Ecosystem conservation and rehabilitation have become an urgent task in the MYR, with many encouraging measures and policies being issued recently, such as the ecological operation of the TGP during May–June to promote the spawning of four major Chinese carps [22].

A typical sub-reach of Jianli in the Jingjiang Reach of the MYR (Fig. 1b) was selected as the study objective to investigate the impact of channel erosion on the requirement of the environmental flow of riparian habitat in the current study, according to the following reasons. (i) This area serves as a critical spawning site and breeding area for the fish. (ii) It is currently in the spot area of river bed erosion in the MYR, and the hydraulic condition in this sub-reach is also obviously influenced by bed erosion. Therefore, this sub-reach provides a good chance to reveal the relationship between the environmental flow (EF) requirement and the channel evolution. (iii) Hydrological and biological data are all available in this sub-reach for achieving these objectives.

2.2. Data sources

The daily averaged flow data (including river stage and water discharge) and topography data in 2002–2020 in the MYR were collected from the CWRC. Other related data were obtained from the study of Ban et al. [23] and the data source of Ban [24], covering the biological preference data and the weight usage area (WUA)-discharge (Q) curves for different guilds of larval fish, macroinvertebrates, hygrophytes, and juvenile fish in the Jianli sub-reach.

3. Methods

To achieve the second objective, a modified habitat simulation method was adopted to calculate the environment flow in the study reach, with a relationship between the weight usage area and river stage being incorporated as a medium step of the calculation. This change incorporated the effect of river bed level change into the environment flow assessment.

3.1. Rating curve of river stage and water discharge

The rating curve between river stage and discharge is the simplest way to represent the hydraulic condition at a specific site, and its change is generally controlled by the combined effects of channel deformation, downstream base variation, bed roughness change, etc. Generally, the rating curve of river stage (Z) and discharge (Q) can be written as [25]:

$$Z=\alpha Q^{\beta }$$ (1)

where α is the coefficient, and β is the exponent. However, for the convenience of the following habitat simulation, equation (1) is transformed into:

$$Q=k{Z}^m$$ (2)

where k is the coefficient, and m is the exponent.

3.2. Habitat simulation

Fig. 2 shows the flow chart of the habitat simulation. Based on the field sampling data, the biological preferences for the target guilds can be determined, and the habitat suitability index (HSI) can be expressed as a function of hydraulic condition and river bed substrate condition. At the same time, the hydraulic conditions under different discharges (Q) can be calculated through a two-dimensional hydrodynamic model, and the habitat suitability index (HSI) distributions at different discharges can be summarized by calculating the weighted useable area (WUA), which has been traditionally computed as the sum of stream surface area weighted by multiplying HSI. Afterward, a relationship between the WUA and Q can be established.

Fig. 2.

Flow chart for the improved habitat simulation method. HSI: habitat suitability index; WUA: weighted useable area (m²); Q: discharge (m³ s⁻¹); Z: river stage (m). h: water depth (m). u: flow velocity (m s⁻¹). d₅₀: medium diameter of bed substrate/material (mm). A1 means assumption 1, and A2 means assumption 2.

There is an important underlying assumption (referred to as A1) in the calculation as mentioned above, that is, the river should be under an equilibrium or quasi-equilibrium state. Under this state, the hydraulic condition will not have a significant variation under the same discharge, and thus, discharge can be used as the controlling factor for WUA. However, this assumption is not solid for the rivers under intensive channel deformation, where the hydraulic condition will change in different periods, even though the inflow discharge remains unchanged.

Therefore, river stage Z is instead adopted as the controlling factor of hydraulic condition in the current paper. The reason is that the flow velocity u is proportional to the depth h according to the Chézy formula (u ∝ h^2/3), and thus, the dependence of HSI on velocity can be transferred into a dependence on the depth as well. If the substrate condition of the riparian zone (represented by the medium diameter in d₅₀, in Fig. 2) slowly changes, the water depth will control HSI. Therefore, the WUA will be primarily related to the submerged area of the riparian zone, which is generally controlled by the river stage. The underlying assumption (referred to as A2) is that the riparian hydraulic condition will not significantly change under the same river stage. This assumption can be satisfied if the riparian topography changes slowly, given that around 92% of the channel erosion volume in the MYR is mainly concentrated in the low-flow channel. Therefore, the topography change in the riparian zone is much less intensive.

However, the WUA-Q curve is still necessary, considering that the river stage is not a good regulation index for dam operation from the perspective of practical engineering. Therefore, using the rating curve of Q-Z (equation (2)), the WUA-Z relationship can be further converted into a WUA-Q relationship. Including the Q-Z curve in the flow chart of habitat simulation enables us to include the impact of hydraulic condition variation on the assessed environmental flow.

4. Results

4.1. River bed change

Fig. 3a illustrates the temporal changes in erosion intensity (E) within the four sub-reaches of the MYR, where E represents the annual erosion volume per unit length in a sub-reach, measured in m³ km⁻¹ a⁻¹. The erosion processes in the MYR can be divided into three periods: the 145 m impoundment period of the Three Gorges Project (PI, 2003–2008), the 175 m impoundment period of the Three Gorges Project (PII, 2009–2013), and the cascade reservoirs operation period (PIII, after 2013). From PI to PIII, the average erosion intensity in the MYR noticeably increased from 9.4 × 10⁵ to 2.1 × 10⁵ m³ km⁻¹ a⁻¹ due to the enhanced impacts of upstream sediment retention. The hotspot area of channel erosion gradually propagated downstream from the Yizhi Reach to the Chenghan Reach. For example, the maximum erosion intensities in PI and PIII were 6.2 × 10⁵ and 8.7 × 10⁵ m³ km⁻¹ a⁻¹, respectively, occurring in the Yizhi and Chenghan Reach. The intensive channel erosion resulted in a significant lowering of the average river bed in the MYR. The average erosion depths in the four sub-reaches of the MYR during 2002–2020 were 2.7, 2.6, 1.0, and 1.7 m, respectively, as shown in Fig. 3b. Furthermore, the rate of river bed erosion in the MYR was much faster than previously predicted, as shown in Fig. 3c. This is because the sediment amount entering the TGP was only 1/3 of the anticipated value before the operation of the TGP, which thus further enhanced the erosion downstream of TGP, as compared with previous anticipation by Huang and Huang [26]. Besides, the incoming sediment in the MYR is anticipated to remain low by the end of this century [19]. This means that in the coming decades, the river bed will continue to undergo intensive erosion, significantly altering the physical characteristics of the MYR.

Fig. 3.

River bed erosion in the MYR. a, Temporal changes in the erosion rate in different sub-reaches. b, Temporal changes in the average river bed level in different sub-reaches. Bed levels in the Yizhi and Jingjiang sub-reaches are depicted using the left-hand coordinate axis, whereas those in the Chenghan and Hanhu sub-reaches are presented on the right-hand coordinate axis. c, Ratio of cumulative bed erosion mass in N years to that in ten years. Reprinted with permission from Ref. [33]. Copyright 2022, Elsevier B.V. TGP: Three Gorges Project. PI: 145 m impoundment period of TGP. PII: 175 m impoundment period of TGP. PIII: cascade reservoirs operation period.

4.2. Hydraulic condition change

4.2.1. Hydraulic condition changes along the reach

Fig. 4 gives the temporal changes in the lowest water discharge Q_min and lowest river stage Z_min at different hydrometric stations in the MYR from 1998 to 2020. As indicated, after the operation of TGP, the values of Q_min in 2020 increased by 37–107%, compared with the values in 1998. The water discharge increased from 3690 m³ s⁻¹ in 1998 to 6340 m³ s⁻¹ in 2020 at Shashi station and from 3740 to 6510 m³ s⁻¹ at Jianli station. However, the river stage did not monotonically increase with the discharge but remained almost unchanged or even decreased at some stations after 2009. For example, the river stage at Shashi station increased from 30.54 m in 1998 to 31.16 m in 2009 but then decreased to 29.74 m in 2020 (even lower than the value before the TGP). The river stage at Jianli station increased from 24.09 m in 1998 to 24.89 m in 2012 but then decreased to 24.49 m in 2020. The inconsistent changes between Q_min and Z_min were caused by the intensive erosion of the river bed, which changed the channel boundary conditions and thus influenced the hydraulic conditions in the MYR. Besides, the changes in hydraulic conditions will become increasingly obvious with the continuous development of channel erosion.

Fig. 4.

Temporal changes in the lowest water discharge and river stage at different hydrometric stations in the Middle Yangtze River from 1998 to 2020. a, Yichang; b, Shashi; c, Jianli; d, Hankou. The arrow indicates a changing trend.

4.2.2. Hydraulic condition variation at a typical station

Fig. 5a shows the data pairs of the measured river stage Z and water discharge Q at the Jianli station from 2002 to 2020, and Fig. 5b depicts the fitted Q-Z curves in 2002 and 2020. It can be seen that the Q-Z curve was changed. For the low and medium flows (Q < 25,000 m³ s⁻¹), the river stage was decreased at the same discharge because of the river bed lowering. For example, at the discharge of 15,000 m³ s⁻¹, the corresponding river stage was lowered by around 1.2 m from 2002 to 2020. Besides, the Q-Z data pairs in Fig. 5a exhibit high scattering and are enclosed by four boundaries. The upper boundary is generally controlled by the requirement of flood control management, and the lower boundary is controlled by the navigation requirement. The left and right boundaries are because of natural fluctuation of river flow, channel deformation, and downstream tributary confluence. At the discharge of 20,000–30,000 m³ s⁻¹, the difference between the river stages on the left and right boundaries is 6–8 m. This implies that the different choices of the Q-Z curve in the hydrodynamic simulation may result in obviously different results when assessing the environment flow.

Fig. 5.

Relationship between the river stage and water discharge at Jianli. a, Data points during the period 2002–2020. b, Rating curves in different years. Q: discharge (m³ s⁻¹); Z: river stage (m); ΔZ: the difference of river stage (m).

4.3. Impacts of channel erosion on the environmental flow on the main stem

The WUA-Q curves (Fig. 6a) for different guilds in Ban et al. [23], specific to the hydraulic condition in 2019, were first transformed into WUA-Z curves, with the Q-Z curve at Jianli station in 2019 being used. From Fig. 6b, it can be found that at the river stage of around 30 m (close to the bankfull level of around 33 m), the WUA values of larval fish, macroinvertebrates, and juvenile fish reaches the highest values, whereas the WUA value of hygrophytes reaches the highest value at a river stage below 20 m, which is beyond the variation range of the measured river stages during the period 2002–2020 in Fig. 5a. The obtained WUA-Z curves for different guilds were assumed to remain unchanged in different years and were then used to calculate the WUA-Q curves in different years of 2002–2020, using the Q-Z curves in the corresponding years.

Fig. 6.

Changes of weighted useable area WUA with the discharge Q and river stage Z at Jianli in 2019. a, WUA-Q curves (original data are obtained from Ref. [24]). b, WUA-Z curves converted from the WUA-Q curves.

For the convenience of comparison between different guilds, the WUA-Q curve was then converted into the WUA_P-Q curve, where WUA_P = WUA/WUA_max and WUA_max represents the maximum value of WUA for each guild. The optimal (OEF) and threshold environmental flows (TEF) were defined as the discharge ranges [Q_O0, Q_O1] and [Q_T0, Q_T1] that corresponded to WUA_P ≥ 90% and ≥50%, respectively. Besides, the obtained WUA_P-Q curves using the left and right boundaries of the Q-Z data in Fig. 3a (WUA_P-Q_LFB and WUA_P-Q_RTB) were used to estimate the uncertainties of TEF and OEF caused by the hydraulic condition variation. The uncertainty of TEF (ΔTEF) was defined as the difference between the calculated values of Q_T1 from the curves of WUA-Q_LFB and WUA-Q_RTB, and the same definition was applied to the uncertainty of OEF (ΔOEF).

Fig. 7 presents the calculated WUA_P-Q curves for guilds in different years in the Jianli sub-reach. The temporal change of the WUA_P-Q curve can be easily identified, which gradually moves rightward to the curve of Ban et al. [23]. For the larval fish, the discharges (Q_h) corresponding to a WUA_P of 100% increase from 14,612 to 17,252 m³ s⁻¹ during 2002–2020, and the TEF changes from <19,875 to <22,303 m³ s⁻¹. For juvenile fish, Q_h has the same value as larval fish, whereas the TEF increased from <20,038 to <22,455 m³ s⁻¹. For macroinvertebrates, Q_h increases from 16,388 to 18,985 m³ s⁻¹, and the TEF changes from <29,836 to <31,308 m³ s⁻¹. For hygrophytes, Q_h is less than 3000 m³ s⁻¹, and the TEF changed from <9910 to <12,469 m³ s⁻¹. However, the discharges of less than 3000 m³ s⁻¹ nearly disappeared after the TGP because the lowest discharge was designed to be greater than 6000 m³ s⁻¹ to keep sufficient water depth for the downstream navigation.

Fig. 7.

WUA_P-Q curves for different guilds in different years in the Jianli sub-reach of the MYR: a, Larval fish; b, Macroinvertebrates; c, Hygrophytes; d, Juvenile fish. WUA_P: the proportion of WUA to WUA_max (%); WUA: weighted useable area (m²); WUA_max: the maximum value of WUA (m²); Q: water discharge (m³ s⁻¹); ΔTEF: the uncertainty of threshold environment flow (m³ s⁻¹).

The arrows in Fig. 7 represent the potential variation ranges of WUA_P for larval fish, macroinvertebrates, and juvenile fish at discharge levels of 20,000 and 30,000 m³ s⁻¹. It is observed that the variation range generally increases with the discharge for all three guilds. For instance, at a discharge of 20,000 m³ s⁻¹, the WUA_P for larval fish ranges from 20% to 88%; at a discharge of 30,000 m³ s⁻¹, it ranges from 5% to 86%. Similarly, for macroinvertebrates, the WUA_p ranges from approximately 61%–100% at 20,000 m³ s⁻¹ and around 29%–100% at 30,000 m³ s⁻¹.

Table 1 provides the uncertainties (ΔTEF and ΔOEF) associated with the estimated TEF and OEF. The uncertainties are high, with ΔTEF ranging from 10,332 to 35,056 m³ s⁻¹ and ΔOEF ranging from 2098 to 20,795 m³ s⁻¹. These uncertainties arise from the annual and inter-annual variations in the hydraulic conditions.

Table 1.

Uncertainties in the calculated threshold and optimal environment flow (ΔTEF and ΔOEF) caused by the annual and inter-annual variations of the hydraulic condition.

Guilds	ΔTEF (m3 s−1)	ΔOEF (m3 s−1)
Larval fish	22,343	17,618
Macroinvertebrates	35,056	20,795
Hygrophytes	10,332	2098
Juvenile fish	22,546	18,268

5. Discussion

5.1. Impact of channel erosion on the survival of macroinvertebrates

Wang [10] has pointed out that intensive channel erosion was the major reason for the decreasing density and biomass of macroinvertebrates on the main stem of the MYR. From the macro perspective, we know how the river bed lowering caused by channel erosion changed the required EF of macroinvertebrates. From the mechanical perspective, it can also be inferred that the entrainment of bed material/substrate will closely relate to the survival of macroinvertebrates.

According to the investigations of Wang [10] and Zhang et al. [27], the optimal ranges of water depth and flow velocity for the macroinvertebrates in the MYR are 0–15 m and 0–1.8 m s⁻¹, respectively. Based on the formula of sediment entrainment velocity u_cs = k₁h^1/6d₅₀^1/3, where k₁ is 5.91 in MYR according to the research of Yu and Lu [28], it can be calculated that the sediment entrainment velocity for the bed material in the MYR is around 0.5 m s⁻¹, at a water depth of 7.5 m (the medium value of the optimal ranges of water depth for macroinvertebrates) and a medium diameter d₅₀ of 0.20 mm. If the actual flow velocity is below this critical value, the bed will generally remain stable; otherwise, erosion will occur.

Fig. 8 shows the relationship between the density of the macroinvertebrates and the flow velocity in the MYR, proposed by Ma [21]. The maximum velocity for macroinvertebrates is around three times the average entrainment velocity of bed material. This indicates that macroinvertebrates are tolerant to bed erosion to some degree. However, the biomass of the macroinvertebrates quickly decreases when the flow velocity is larger than the entrainment velocity, indicating an underlying quantitative relationship may exist between the movement of the river bed and the survival of the macroinvertebrates. Furthermore, the curve in Fig. 8 can be divided into three regions representing stable, unstable, and very unstable beds regarding the macroinvertebrates' survival. In the stable region, the flow velocity is not a major factor influencing the biomass of macroinvertebrates; whereas in the unstable region, the flow velocity is large enough to entrain the bed sediment (substrate) and thus becomes the major factor.

Fig. 8.

Relationship between the density of macroinvertebrates and the near-bed velocity. Reprinted with permission from Ref. [21]. Copyright 2019, China Academic Journal Electronic Publishing House. D_m: density of macroinvertebrates (ind m⁻²); u_cs: sediment entrainment velocity (m s⁻¹); u_cm: the critical velocity for the disappearance of macroinvertebrates (m s⁻¹).

Analogous to the calculation of the entrainment velocity of sediment, it can be found that the entrainment velocity for macroinvertebrates will be around 1.8 m s⁻¹ at a water depth of 7.5 m when the average size of macroinvertebrates is around 10 mm. Therefore, it may be inferred that when the river bed substrate is entrained by the flow, the biomass of the macroinvertebrates will quickly decrease, whereas the maximum velocity for the survival of macroinvertebrates is controlled by the entrainment of the organisms themselves. Taking a further step, the sediment entrainment near the river bed has long been the hot research area of river dynamics and fluid mechanics [[29], [30], [31]], and thus, if these theories can be refined to understand the biokinetics of macroinvertebrates, maybe some new progress can be achieved.

5.2. Implications for ecological restoration works in the MYR

The above analysis has indicated that the intensive river bed erosion has led to an increase in the requirement on the environment flow in the MYR, and this kind of change will possibly continue in future decades. Therefore, for the ecological restoration in the MYR, it's necessary to consider and predict the long-term channel evolution process in the MYR and prepare for the impacts of physical setting change.

The inherent uncertainties in hydraulic conditions within natural rivers, which mirror natural flow variations, are generally considered beneficial for ecosystems. However, these high uncertainties also hinder the accurate assessment of environment flow. Therefore, when assessing the environment flow in the MYR, it is necessary to consider the annual and inter-annual changes in the hydraulic condition, and the environment flow needs to be re-assessed after a relatively short period (a few years). Furthermore, this raises the question of determining to what extent these uncertainties (or natural flow fluctuations) should be preserved to maintain a healthy ecosystem in the MYR.

Wang et al. [9] summarized the main threats to the Yangtze River floodplain ecosystem, including (i) the habitat loss caused by river channelization, sharp shrinkage of lake area, degradation of lakeshore zones, and intensive sand mining; (ii) the alteration of hydrological regimes caused by cascade reservoir operation, river-lake disconnection, and water pollution; and (iii) the over-exploitation of biological resources.

In the authors’ view, among the second threat, the change in the sediment transport process and its impact on the ecosystem in the MYR have become a very tough problem to address. This has gained much less attention than the water resources in the MYR, even though the important role played by the sediment has been recognized in previous studies [5,32]. There are two major reasons why the sediment transport process should be included in the overall designation of ecosystem conservation in the MYR: (i) the associated intensive and long-term channel evolution, as stated above, will cause an obvious change in the physical settings of the MYR, and thus leads to the biological adjustment; and (ii) sediment is an important carrier of many biochemical matters, such as phosphorus and heavy metals, and the changes in sediment sources are implications for the changes in sources of these bio-chemical matters. Besides, in extreme conditions like flood events, when sediment load changes, flow velocity and depth might not be the primary factors influencing aquatic creatures.

The MYR has transferred from a sediment transport-limited system into a sediment supply-limited system after the TGP operation, with most of the sediment being eroded from the river bed and the tributary input also obviously decreasing [33]. Besides, the intensive bank protection works along the MYR also limited the sediment resuspension in the riparian floodplain zone. The sediment trajectory in the MYR has been obviously changed, and sediment flux is thus more “unnatural” than the flow flux. Evaluating and preserving an optimal sediment flux to keep a healthy river system needs more scientific effort.

6. Conclusions

Environment flow assessment is key in ecosystem conservation and restoration in the middle Yangtze River (MYR). However, the impacts of channel evolution on the environment flow requirement have seldom been investigated. Based on field measurements, the current study investigated the intensive channel erosion process and analyzed its impacts on the requirement of environment flow for riparian habitats in the MYR. The study leads to the following conclusions:

• (i)The intensive erosion has led to the average river bed level being lowered by 1.0–2.7 m in four sub-reaches of the MYR from 2002 to 2019, which changed the hydraulic condition in the MYR. Even with an increase of 37–107% in the lowest discharges after the Three Gorges Project, the lowest river stages at some hydrometric stations decreased to a value lower than before the TGP because of the bed erosion.
• (ii)Calculated environment flow in a typical sub-reach of the MYR was found to change with the process of channel evolution. The threshold environment flow increased from <around 20,000 to < around 22,500 m³ s⁻¹ for juvenile and larval fish, from <around 29,800 to <31,300 m³ s⁻¹ for macroinvertebrates; and from <9900 to <12,500 m³ s⁻¹ for hygrophytes in 2002–2020. This indicated that the hydraulic condition variation due to river bed lowering resulted in an obvious increase of 1500–2600 m³ s⁻¹ in the required environment flow.
• (iii)It was found that the critical flow velocity for the decrease of biomass of macroinvertebrates agreed with the average sediment entrainment velocity of around 0.5 m s⁻¹ in a sand river bed in the MYR. However, the maximum flow velocity at which macroinvertebrates can survive is possibly related to the organisms' entrainment. For the ecological restoration in the MYR, it's necessary to predict the long-term channel evolution processes in the MYR and prepare for the impacts of physical setting change because of the channel evolution.

CRediT authorship contribution statement

Shanshan Deng: Writing - Original Draft, Methodology, Investigation. Junqiang Xia: Conceptualization, Supervision, Resources. Hengzhu: Visualization, Investigation, Data Curation. Jie Liang: Writing - Reviewing & Editing, Validation. Huiwen Sun: Visualization, Validation. Xin Liu: Data Curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.