Quantifying Environmental Flow in the Form of Pulse Flow for Fish Protection

Abstract

Pulse flow, which includes base flow, peak flow, flow duration, occurrence time, and frequency, is a manifestation of environmental flow in rivers. This paper describes a methodological framework for determining pulse flow based on an analysis of fish spawning behavior and presents its application to the second Songhua River in northeastern China. Peak flow was determined based on the hydrographic-habitat relationship for fish spawning in conjunction with physical habitat simulation. The flow duration was determined based on the incubation period of fish eggs. The occurrence time and frequency were determined based on the suitable temperature for fish spawning. This application differs from conventional methods for dividing fish spawning periods and selecting target species in the corresponding period. Fish were divided into cold-water fish, hypothermal-water fish, and warm-water fish according to spawning temperature, and the target species in each month of the spawning period (April to July) were identified. For the same spawning period, the hydrographic-habitat relationships of target species with adhesive and drifting eggs were weighted to determine the peak flow. The most suitable peak flow for fish spawning from April to May in the research region is 900 m³/s and 1200 m³/s for June to July. Using the proposed framework, an ecological hydrograph from April to July was obtained by combining the method of pulse flow with habitat simulation. Fifteen days, eight days, and five days were chosen as the flow durations for April, May, and June to July, respectively. We recommend scheduling two high pulse flows each month from May to July while limiting the occurrence to only one in April. The results show that this framework offers a promising approach for developing environmental flows in rivers with a variety of fish species.

1. Introduction

Freshwater ecosystems are one of the most impacted and threatened ecosystems, where the decline in biodiversity is 2–6 times faster when compared to marine and terrestrial ecosystems [1,2]. One of the reasons for this deterioration is habitat loss due to the alteration of the flow regime caused by the water conservancy project [3,4]. In order to reduce the negative impact, stream regulation to release river flows, including water quantity, quality, and timing, was developed as an important approach [5,6,7]. Environmental flows have been proven to have tremendous environmental benefits for ensuring the sustainability of ecosystem services and being the theoretical foundation for stream regulation.

Various methods can be used to estimate environmental flow, which could be classified as hydrologic approaches, hydraulic approaches, habitat-based approaches, or a synthesis of the three [8,9]. In recent years, fish habitat protection has been regarded as a critical element for environmental flow research [10,11]. Some researchers have confirmed that fish show strong preferences for certain values of hydrodynamic parameters such as water depth, current velocity, and streambed substrate [12]. Physical habitat simulation can be used to predict changes in fish habitat by coupling hydrodynamic modeling with habitat suitability curves for target species [8,13]. Habitat suitability models employing numerical algorithms such as artificial neural networks, fuzzy logic methods, and genetic algorithms have been widely used for habitat suitability curves [14,15,16,17]. However, this approach normally delivers information about the hydro-environmental preferences of an individual species, while the demands of other species with disparate spawning habits are ignored [18]. The environmental flow should be a time series of flows within a range, considering different life stages of target species and hydrological seasonality. Especially for rivers with diversified spawning habits, a comprehensive analysis of all spawning patterns is essential for fish preference curves.

Successful environmental flow prescriptions require an accurate understanding of the linkages among flow elements and biotic responses [7]. The prescriptions include not only flow amount but also flow duration and occurrence time. To address the issue above, pulse flow, which comes from the flood pulse concept (FPC), was selected as an expression of environmental flow. The FPC was to describe the periodic change in the river hydrological regime [19]. The flood pulse is the pivotal driving force that maintains equilibrium in river–floodplain systems, leading to periodically variable wetting-drying boundaries. The horizontal connection between the main channel and floodplains enriches the habitat area and lateral exchange, increasing the recirculation of nutrients and organic matter transformation [20]. On the basis of FPC, the pulse flow consists of base flow, peak flow, flow duration, and occurrence time.

In this study, we described a framework to determine pulse flow during the whole spawning period based on a requirement analysis of fish habits. The primary goal was to quantify environmental flows in rivers with various spawning types of fish species. We defined flow elements as follows: (1) the peak flow was determined with preference velocity and water depth of fish spawning; (2) the occurrence time was determined with suitable water temperature for spawning of various fish species; and (3) the flow duration was determined with the incubation period of fish eggs. The physical habitat model was applied to calculate a suitable area of fish habitat. A case study of the Second Songhua River is presented to examine the feasibility of the proposed approach.

2. Materials and Methods

2.1. Study Site

The Songhua River is located in the northeastern region of China, and is one of China’s seven major rivers. The Upper Songhua River, known as the Second Songhua River (SSR, for short), originates from Changbai Mountain and travels 958 km, draining a land area of 73,400 km². The studied reach takes about 370 km and extends from the Fengman Reservoir downstream to the Fuyu gauging station (Figure 1). The lower part of the reach is 2000 m wide on average, with numerous river islands, and flows slowly along a small hydraulic gradient. With sufficient solar radiation and a broad water area, the SSR creates a suitable environment for fish reproduction [21]. However, the existence of dams and reservoirs severely disrupted natural flow regimes, resulting in a gradual decrease in fish resources [22,23]. Investigations have shown that the construction of large reservoirs has severely damaged the spawning grounds of the drift channel in the lower reaches and has also caused the migratory fish to become rare species in the SSR. Meanwhile, the SSR is the major freshwater source for industry, agriculture, and domestic water in Northeast China [22,24]. Consequently, the fish habitat in this area has been subject to intense disturbance due to water resource exploitation and utilization.

2.2. Framework for Quantifying Pulse Flow

The pulse flow consists of the base flow, the peak flow, the flow duration, and the occurrence time. The base flow is intended to maintain a minimum continuous flow to provide essential fish habitat and protect migration channels. The base flow is usually determined by the Tennant method [25]. A high flow pulse, where the flow rate exceeds the base flow but falls below the full bank flow rate, is the major stimulus to fish reproduction and migration. At the same time, this process helps to shape the river configuration and alleviate the pressure on the water environment caused by low flow conditions. These alternating appearances can induce biological activity that affects the distribution, abundance, and stability of fish communities. The duration of the pulse flow is required to cover the length of time that the fish are stimulated to spawn by the current and the time that the eggs hatch. Otherwise, the drifting eggs will sink to the bottom of the river due to insufficient hydrodynamic conditions, and the adhesive eggs may be exposed to the bank due to falling water levels. The occurrence time of pulse flow is determined based on the natural flow rate of the hydrological stations and should be as consistent as possible with natural flow processes and with fish spawning periods under natural conditions.

In this paper, pulse flow is quantified by combining fish assessment, physical habitat simulation, and natural flow analysis, as shown in Figure 2. The first component of the framework involves assessing the spawning habits of fish species by analyzing the spawning temperature, spawning period, and spawning type to classify the main fish species. Habitat preferences and egg hatching times were then evaluated for fish species with different spawning periods and spawning types. The second component involves configuring a hydrodynamic and hydrological model using land use maps, topography data, meteorological data, and hydrologic data. The third component involves evaluating habitat suitability by integrating the target species’ preference curves with simulated water depth and velocity. We establish the relationship between environmental flow parameters and biological responses to determine the peak flow. The last component is to determine the suitable environmental flow with natural flow analysis. In fact, the method of natural flow analysis can also calculate the peak flow, but the results are slightly less accurate than the habitat simulation. Consequently, the recommended hydrograph, consisting of base flow, peak flow, frequency, time of occurrence, and duration, was provided.

2.3. Fish Assessment

Forty-five fish species have been recorded before strong human disturbance in the SSR, as shown in

Table 1. Classification and spawning habits of major fish species in the Second Songhua River (Classification based on the spawning water temperature and eggs properties).

Classification	Fish Species	Spawning Season	Spawning Temperature	Egg Properties	Incubation Duration (Day)
Cold-water fish	Brachymystax lenok	April	5 °C–10 °C	adhesive	15–20
	Hucho taimen	April	5 °C–10 °C	adhesive	20
	Esox reicherti	April	4 °C–7 °C	adhesive	9–10
	Thymallus arcticus grubei	April	6 °C–10 °C	adhesive	22–24
	Cottus poecilopus	April	7 °C–11 °C	adhesive	7–9
	Leuciscus waleckii	April–May	6 °C–14 °C	adhesive	8–10
	Hypomesus olidus	April–May	4 °C–12 °C	adhesive	8–9
	Hypomesus nipponensis	April–May	6 °C–16 °C	adhesive	6–8
Hypothermal-water fish	Barbatula nuda	May	10 °C–15 °C	adhesive	7
	Lampetra reissneri	May	10 °C–16 °C	adhesive	7–8
	Lampetra japonica	May	12 °C–16 °C	adhesive	10–11
	Rhodeus sericeus	May	12 °C–18 °C	adhesive	4–5
	Phoxinus phoxinus	May	12 °C–15 °C	adhesive	12–14
	Rhynchocypris percnurus	May	14 °C–16 °C	adhesive	5
	Phoxinus lagowskii	May	14 °C–17 °C	adhesive	5–7
	Saurogobio dabryi	May–June	12 °C–20 °C	adhesive	3–4
	Aphyocypris chinensis	May–June	15 °C–20 °C	adhesive	3–4
	Perccottus glenii	May–June	15 °C–20 °C	adhesive	5–6
warm-water fish with adhesive eggs	Carassius auratus gibelio	May–June	17 °C–22 °C	adhesive	3–5
	Hemibarbus maculatus	May–June	16 °C–23 °C	adhesive	4–5
	Abbottina rivularis	June–July	18 °C–25 °C	adhesive	5–6
	Cultrichthys erythropterus	June–July	18 °C–23 °C	adhesive	3–4
	Cyprinus carpio	June–July	18 °C–25 °C	adhesive	4–5
	Hemibarbus labeo	June–July	17 °C–22 °C	adhesive	3
	Megalobrama skolkovii	June–July	20 °C–26° C	adhesive	2
	Pelteobagrus nitidus	June–July	22 °C–30 °C	adhesive	3
	Pelteobagrus fulvidraco	June–July	22 °C–30 °C	adhesive	3
	Pseudobagrus ussuriensis	June–July	20 °C–25 °C	adhesive	2
	Squalidus chankaensis	June–July	18 °C–25 °C	adhesive	5
Warm-water fish with drifting eggs	Gobiobotia pappenheimi	June–July	11 °C–26 °C	drifting	4
	Aristichthys nobilis	June–July	20 °C–28 °C	drifting	2
	Culter alburnus	June–July	22 °C–25 °C	drifting	2
	Elopichthys bambusa	June–July	18 °C–26 °C	drifting	2
	Chanodichthys mongolicus	June–July	20 °C–24 °C	drifting	2
	Ctenopharyngodon idellus	June–July	18 °C–28 °C	drifting	2
	Hemiculter leucisculus	June–July	20 °C–26 °C	drifting	2
	Hypophthalmichthys molitrix	June–July	18 °C–28 °C	drifting	2
	Parabramis pekinensis	June–July	16 °C–25 °C	drifting	2
	Pseudaspius leptocephalus	June–July	17 °C–23 °C	drifting	3–4
	Pseudobrama simoni	June–July	24 °C–26 °C	drifting	1–2
	Rostrogobio amurensis	June–July	20 °C–26 °C	drifting	2
	Sarcocheilichthys lacustris	June–July	24 °C–26 °C	drifting	4
	Siniperca chawtsi	June–July	22 °C–24 °C	drifting	2–3
	Xenocypris argentea	June–July	20 °C–26 °C	drifting	2
	Xenocypris microlepis	June–July	18 °C–26 °C	drifting	3

[4,26]. By assessing the fish fauna and spawning patterns, we classified these species into four types. The fish species spawning at water temperatures ~4–10 °C were defined as cold-water fish, which generally spawn right after the freezing period. The fish species spawning at water temperatures of 10–18 °C were defined as hypothermal-water fish, which mostly spawn in May. The fish species spawning at water temperatures above 18 °C were defined as warm-water fish, which normally spawn during summer seasons. The preferred water temperature determines the spawning season for each fish type and, thus, the target species for each month. In the SSR, the target species are cold-water fish in April, hypothermal fish in May, and warm-water fish in June and July. Therefore, we divided the entire spawning period into three phases, including April, May, and June to July, and then quantified the pulse flow in each phase. Depending on the type of egg, warm-water fish can be further divided into adhesive egg and drifting egg fish. Notably, all cold-water fish and hypothermal-water fish spawn adhesive eggs.

According to relevant literature such as “Heilongjiang Fishery” [27] and “Freshwater Fishes in Northeast China” [28], fish species that exhibit different egg types vary widely in hydrological preferences. The adhesive eggs are laid among aquatic plants and pebbles, preferring shallow water for their spawning grounds. For example, the water depth of the spawning ground for Phynchocypris lagowskii is 0.3–0.6 m, and for Leuciscus waleckii it is 0.15–0.5 m. In general, the optimum water depth for most of the adhesive eggs is between 0.2 m and 1 m. Considering the water depth suitable for vegetation growth, we set the maximum water depth for spawning grounds at 1.5 m. Based on this, the optimum depth range is then suggested to be 0.4 m to 1 m, while the lower limit value is 0.2 m. In general, fish with adhesive eggs are suitable for moderate flow, while those with drifting eggs prefer higher current speeds. The adhesive eggs require a certain amount of flow to maintain water quality and nutrient transport, but the velocity should not be overly strong. For example, Leuciscus waleckii prefers a velocity not exceeding 0.4 m/s, with 0.1 m/s as the lower limit. Both Lampetra reissneri and Lampetra japonica are suitable for spawning in slow-flowing areas. The suitable velocity for adhesive-egg fish was set at 0.2–0.5 m/s, with a maximum flow rate not exceeding 0.8 m/s. Based on the above analysis, the habitat suitability curves for adhesive-egg fish are illustrated in Figure 3.

In contrast to the adhesive-egg fish, drifting-egg fish require greater flow velocity stimulation and deeper water depths during their spawning period. According to our tests, Ctenopharyngodon idellus and Hypophthalmichthys molitrix prefer 0.4 to 1 m/s during the spawning season [4]. The production of Chanodichthys mongolicus and Xenocypris argentea requires strong flow stimulation with an appropriate current of approximately 0.6–1 m/s. And Xenocypris microlepis is most likely to spawn when the current speed is between 0.4 and 0.8 m/s. Based on a comprehensive analysis of the relevant species, the optimal velocity range is 0.4 m/s–1.2 m/s. The minimum velocity of the supporting egg drifting is 0.25 m/s, and the maximum velocity is set at 1.5 m/s, depending on swimming ability. Drifting-egg fish have relatively relaxed water depth requirements. For example, the proper water depth for Parabramis pekinensis is 0.6–2 m, and for Xenocypris argentea is 0.2–2.3 m. We set 1–3 m as the well-posedness depth and 0.5 m as the t lower limit. Based on the above analysis, the habitat suitability curves for drifting egg fish are illustrated in Figure 3.

2.4. Fish Habitat Modeling

An eco-hydraulic simulation tool was created by combining the 1D hydraulic model based on DHI MIKE ZERO 2014 (MIKE 11 for short) with a fish habitat module from Physical Habitat Simulation (PHABSIM for short) [8]. Modeling was carried out from Fengman Reservoir downstream to the Fuyu gauging station. A series of data sources, including topographic data, cross-sectional data, and stream flow, were used to configure the hydrodynamic model. We selected 249 transects with cross-sectional data to divide the computational units. Historical hydrological data from hydrographic stations was also obtained from the watershed authorities. Measurements of daily flow rates and water levels at hydrological stations from 2000 to 2013 were obtained from the China Water Yearbook. The flow rate at the Jilin gauge station and the daily water level at the Fuyu gauge station were used as inflow boundary conditions for the hydrodynamic model. The measured water level at the Songhuajiang gauge station was used for model validation. In addition, slight changes in the river’s topography are common at extreme flows. However, if significant changes do occur, these data are treated as independent datasets in the hydraulic model calibration.

Habitat modeling transforms the results of hydraulic simulation into an index of habitat suitability at each flow increment by combining habitat suitability curves [8]. This index is referred to as a weighted usable area (WUA for short) and is quantitatively evaluated using Equation (1):

$$WUA={\displaystyle \sum _{i=1}^n}{A}_i\times \sqrt{HSI(v{)}_i\times HSI(d{)}_i}$$

(1)

where Ai (m²) is the area of computational unit i, HSI_i is the corresponding habitat suitability index of unit i, n is the total number of computational units, and HSI(v) and HSI(d) are the habitat suitability index for current velocity and water depth, respectively.

For a given flow event, the MIKE 11 model is used to calculate the velocity and water depth for each computational unit. The corresponding HSI at each computational unit is obtained by combining the habitat suitability curves. The sum of all computational units in the spawning ground represents the WUA for this flow event. Repeating the calculation for various flow rates of interest, the WUA-flow rate fitting curve is established. The study area was divided into several sections according to hydrological and eco-logical conditions. The main spawning ground for drifting-egg fish is 161 km–224 km from the estuary, while for adhesive-egg fish it is 80 km–136 km [29].

2.5. Nature Flow Analysis

We determined the time and frequency of pulse flow by analyzing natural flow. The nature flow paradigm (NFP) proposed that natural flow without human disturbance was critical to river ecosystems and species diversity [30,31]. The magnitude, frequency, timing, duration, and change rate of natural flow were described as hydrological processes. According to their magnitude and frequency, flow events can be divided into base flow, high flow pulses, small flood pulses, and major flood pulses.

The base flow, calculated using the Tennant method, corresponds to 60% of the annual average runoff. We defined the threshold for high-pulse flow based on the analysis of the river cross section. The flow rate corresponding to the level of the overflowed main channel has been suggested as a lower threshold for high-pulse flow. We obtained the overbanked water level by analyzing the topography of the river channel. The flow rate threshold is then determined using the relationship between water level and flow rate established from the measured data. In this paper, we took the cross-section of Songhuajiang gauging station as an example, whose terrain is illustrated in Figure 4a, and the water level-flowrate relation is illustrated in Figure 4b. The instream flow will brim the main channel when the water level rises to 151.1 m. Thus, 151.1 m was defined as the level of the overbanked water, and the 585 m³/s corresponding to 151.1 m was identified as the threshold for high-pulse flow. Similarly, threshold values for flood pulses are calculated from historic deluge data. The experiential frequency is calculated with the P-Ⅲ frequency curve, and statistical parameters are calculated with the matrix formula method. The results are 2913 m³/s and 6583 m³/s for two-year and ten-year small floods, respectively.

Based on the thresholds of high-pulse flow, small flood, and large flood obtained from the above analysis, natural flow processes were classified by SPSS to obtain high-pulse flow events. High-flow pulse events must satisfy both magnitude and duration requirements. The flowrate is necessary to exceed the lower threshold of the high-flow pulse, and the pulse duration is also required to reach the hatching time of the fish egg. The SPSS results can illustrate the time and frequency of high pulse flow in each month without human disturbance. In this paper, natural flow during 1956–2000 was derived from a distributed hydrological model (WEP model) illustrated in Hu [32].

3. Results

3.1. Peak Flow

The WUA-discharge relationship for adhesive-egg fish (a) and drifting-egg fish (b) is illustrated in Figure 5, respectively. Figure 5a shows the trend of WUA within the scope from 320 m³/s (base flow) to 2000 m³/s for adhesive-egg fish. With 900 m³/s as the dividing point, the WUA in the range of 300 m³/s–900 m³/s is increasing. When the flow exceeds 900 m³/s, the WUA gradually decreases to a plateau after 1700 m³/s. The relation curve varying from 320 m³/s to 2500 m³/s for the drifting egg fish is shown in Figure 5b. Overall, the WUA increases with flow rates below 2500 m³/s. However, this curve reveals a painfully slow growth trend between 1200 m³/s and 2000 m³/s, with the first inflection at 1200 m³/s. This distribution is due to the abundance of central islands at low water levels. As the flow rate continues to increase, the suitability does not change significantly in tandem with the area of inundation as the water level rises to the level of the overburden. This may have occurred because when the flow rate exceeds 1200 m³/s, a portion of the cross section may be experiencing water at the level of the overbanked area.

Given that cold-water fish in the SSR were essentially laying adhesive eggs, as illustrated in Table 1, the flow rate corresponding to the maximum WUA for adhesive-egg fish (900 m³/s) should be chosen as the peak flow from April to May. However, for the drifting-egg fish, the flow rate corresponding to the maximum WUA is overly large, and the natural runoff process is difficult to meet. Current studies provide different methods to determine the suitable flow, such as selecting the first obvious turning point in the relationship between WUA and flow rate [33]. Considering operability, some scholars have suggested that the flow rate corresponding to the first significant turning point in the WUA-flow rate curve is the objective flow, that is, the minimum ecological flow. From this, we define 1200 m³/s as the appropriate flow rate for drifting egg fish. At the same time, adhesive-egg and drifting-egg fishes have disparate habitat preferences and respond differently to flow rates. Thus, we need to balance the two habitat requirements for setting the peak flow for June-July. In this reach, the WUA for the adhesive-egg fish at 1200 m³/s is 76 km², while the max WUA at inflection is 85 km². And the WUA for the drifting egg fish at 1200 m³/s is 60 km², while the max WUA is about 75 km². At a flow rate of 1200 m³/s, the WUA of both adhesive-egg and drifting-egg fishes can reach 80% of the maximum WUA. In conclusion, an optimal flow rate of 1200 m³/s was suggested for June-July, combing both spawning types.

3.2. Flow Duration

The flow duration is summarized by the incubation time requirements of fish species that spawn in different months. As listed in Table 1, we collected 27 warm-water fish species, 10 hypothermal-water fish species, and 8 cold-water fish species, which were recorded in the SSR. We analyzed the distribution of egg hatching times for different types of fish, as shown in Figure 6. As a result, 89% of warm-water fish have incubation durations of less than five days, and 75% have incubation durations of less than four days. Based on the 80% criterion, five days were chosen as the flow duration for June and July. For hypothermal-water fish, the average time for egg incubation is about seven days, with 80% of species requiring less than eight days. Similarly, the pulse duration in May was eight days, meeting the spawning requirements of 80% of the species during this period. For cold-water species, the average incubation duration is twelve days, and 80% require less than fifteen days. Therefore, fifteen days were confirmed as the pulse flow duration in April, which is sufficient for almost all cold-water fish. The incubation period is mainly affected by the water temperature. The water temperature is lower in April, which implies a long incubation period, while it rises to about 20 °C in June or July, significantly reducing the incubation time.

3.3. Occurrence Time

Based on threshold values, Figure 7 depicts the ecological flow in conjunction with natural runoff from 1956 to 2000. High-pulse flows, with peak flows of around 900–1200 m³/s, tend to occur about five times a year, mainly between April and August. Meanwhile, there were several major floods during this period, with max peak flow rates exceeding 8000 m³/s. According to statistical results, the high pulse from May to July accounts for more than 75 percent of the total occurrence, which is essentially the same as the spawning time. The peak flow of the high pulse process in April was relatively low. Conversely, extreme high flows occurred in August as a consequence of heavy rainfall. Given this, we recommend scheduling two high pulse flows each month from May to July while limiting the occurrence to only one in April and August. As for the occurrence time, water temperature is the dominant indicator for fish spawning and also for ecological operations. Basically, the ecological flow should be guaranteed separately in the first half and second half of May and June, and the same for late April after the melting season and July before the flood season.

3.4. Recommended Pulse Flow

Combining the method of pulse flow with habitat simulation, the ecological hydrograph from April to July is obtained.

Table 2. The recommended pulse flow in the Second Songhua River.

Occurring Time	Duration (Day)	Peak Flow (m3/s)	Base Flow (m3/s)
Mid–late April	15	700	320
Early May	10	900	320
Late May	8	900	320
Early June	8	1000	320
Late June	5	1200	320
Early July	5	1200	320
Late July	5	2000	320

presents the recommendations for each process during the main spawning season. In mid-to-late April, after the melting season, the peak flow needs to reach at least 700 m³/s, corresponding to 90% of the maximum suitable habitat area. Considering the incubation of cold-water species, the flow fluctuates between 320 m³/s and 700 m³/s for about 15 days. Similarly, in early May, a flow process is required with a peak flow of 900 m³/s and a duration of 8 days. Then, in late May and early June, the pulse flow lasts for 8 days with a peak flow rising to 1000 m³/s in order to create spawning grounds for adhesive-egg species and others with lower velocity preferences. When the water temperature rises to 20 °C in late June, the peak flow increases to 1200 m³/s, and the duration decreases to 5 days. This is to stimulate warm-water species migration and spawning at the required velocity and water depth while creating buoyancy forces for drifting eggs. However, most-warm fish spawn in July, resulting in the requirement for twice flow processes. In late July, a flow with a peak rising to 2000 m³/s is suggested for certain species that require higher velocities. The duration needs to last for 5 days to ensure that the eggs and larvae are carried into the main channel. If possible, higher peaks of 3000 m³/s should be seen in August to ensure small flood processes, which can be used to maintain biological communities.

4. Discussion

4.1. Contrast of Habitat-Based Environmental Flow and Recommended Pulse Flow

The habitat simulation approach is widely used in ecological studies to understand the interactions between species and their environment. Given its higher accuracy, we used habitat simulation to evaluate the peak flow. We discarded the historical practice of selecting a single indicator species and classified the various spawning patterns by fish habitat preferences. The recommended flow rate for adhesive-egg species and drifting-egg species, corresponding to the optimal habitat area, was chosen as a reference. Unlike traditional habitat-based methods, we focus more on flow fluctuations. The recommended ecological hydrograph for the entire spawning season was suggested, combined with the determined elements of the pulse flow.

The recommended ecological hydrograph was depicted in Figure 8, and the environmental flow based on the habitat simulation was also illustrated for comparison. Clearly, the proposed ecological hydrograph is more favorable for reservoir scheduling, with less wasteful discharge and simpler operation. On the one hand, the habitat-based method requires maintaining peak flow rates for a prolonged period, while the method proposed in our study allows for fluctuations within a certain range. This can reduce economic losses while ensuring fish habitat, thus balancing the conflict between ecological water demand and social water use. On the other hand, the pulse flow obtained in this paper reflects a continuous rising and falling flow process, which is also more conducive to reservoir operations. The flow rhythms, such as water level fluctuations, will enrich the floodplains and habitats for diverse fish species. The results indicate that this method is a promising approach for developing environmental flow strategies to provide fundamental support for ecological restoration.

4.2. The Practical Significance of Pulsed Discharge to River Management

With the increase in water scarcity, water withdrawal from aquatic systems is occurring globally. The SSR, where the existence of the Fengman reservoir has severely disrupted the natural flow regime, was selected as the study area. The fish in the SSR have been experiencing anthropogenic stress from the control of stream flow for hydropower development and other human uses. Unfortunately, this reach is threatened by declining biodiversity, particularly for cold-water fish. One important measure for reducing ecosystem degradation is to determine ecological flow and implement it through reservoir operation [34].

Successful environmental flow prescriptions require an accurate understanding of the linkages among flow events, geomorphic processes, and biotic responses [7]. Previous studies on environmental flow (or ecological flow) in the SSR have emphasized water quantity. Some scholars have conducted ecological base flows and eco-environmental flows based on hydrologic analysis [23,32]. However, the ecological flow focused on fish protection is still in its infancy. The proposed framework determines the peak flow, flow duration, occurrence time, and frequency during the spawning period to provide fundamental support for ecological restoration. By coupling habitat simulation with the method of pulse flow, our results revealed the adaptability and limitations of the two complementary approaches. The selection of target groups of fish species ensures a strong representation of aquatic ecosystems and sensitivity to hydrological changes. Previous studies in the SSR have focused on certain species while ignoring the requirements of other fish with different spawning patterns [23]. We believe that the ecological hydrograph proposed in this study will help with water management. The implementation of ecological operations helps provide suitable flow regimes for spawning species and is of great significance for fish conservation.

Further discussion is needed regarding the biological response to high pulse flow, which meets the spawning requirement. In April, relatively low flow provides costimulatory signals to migratory fish. During May and June, this process increases noticeably, producing sufficient nutrients and dissolved oxygen to create optimal conditions for fish spawning. Then, the peak flow in July increases to its maximum value, water levels rise, and larvae travel from the spawning ground to the river channel.

5. Conclusions

To offer comprehensive protection to fish species, this paper presents a methodological framework to determine the ecological hydrograph of rivers by coupling the Pulse Flow Concept with habitat simulation and applying it to the second Songhua River in northeastern China. This method enables environmental flow to be quantified in the form of pulse flow during the spawning season (from April to July), including base flow, peak flow, occurrence time, and flow duration. The comprehensive analysis suggests that the appropriate peak flow for fish spawning in the SSR is 900 m³/s in April–May and 1200 m³/s in June–July. The flow duration suggested in this paper is based on fish spawning and incubating behaviors. Fifteen days, eight days, and five days were chosen as the flow duration for April, May, and June–July, respectively. We recommend scheduling two high pulse flows each month from May to July, while limiting the occurrence to only one in April and August. This method is a promising approach for developing environmental flow strategies in rivers with diverse fish species. The results indicate that the proposed pulse flow requires less water discharge from reservoirs than the habitat-based environmental flow and can reduce the economic losses caused by reservoir operation. The fluctuation of pulse flow better stimulates fish spawning; however, the stimulating effects need to be further verified based on monitoring of fish at early life history stages during reservoir operation.