Potential Impact of a Large-Scale Cascade Reservoir on the Spawning Conditions of Critical Species in the Yangtze River, China

: Dam building and reservoir operations alter the downstream hydrological regime, and as a result, a ﬀ ect the health of the river aquatic ecosystem, particularly for large-scale cascade reservoirs. This study investigated the impact of the Gezhouba Reservoir (GR) and the Three Gorges Reservoir (TGR) on the spawning conditions of two critical taxa, i


Introduction
Global river systems have undergone major changes due to intensive anthropogenic activities, such as land use/cover change, irrigation, water diversion, and dam building and operations. According to incomplete statistics, more than half of the 300 biggest river systems around the world are controlled by or under the impact of dams [1]. Up to now, ≈2.8 million dams (with reservoir areas >10 3 m 2 ) have been built and more than 3700 hydropower dams (>1 MW) are currently planned or under construction worldwide [2]. Undoubtedly, dams or reservoirs play an important role in contributing water for domestic use and agriculture, sustaining transportation corridors, and enabling power generation and industrial production. However, they also affect many fundamental processes and functional

Study Region
The TGR is located on the main stem 44 km above the end of the upper Yangtze River. Its operations were done in four stages: the initial stage began in June 2003, with the reservoir water level increased to 135 m above sea level; the transitional stage started in October 2006 with the water level increased to 156 m; the reservoir water level rose to 173 m in November 2008 during the quasinormal stage; and then to the designed reservoir water level of 175 m in October 2010 during the normal stage with the total storage capacity of 39.3 billion m 3 [31]. According to the operational scheme of the Ministry of Water Resources [32], the TGR stores 22.15 billion m 3 of water to reach normal water level of 175 m from 145 m at the end of flood season (generally from 15 September to 31 October, lasting 47 days; it begins refilling from 10 September if the year is relatively dry). The GR, a run-of-river reservoir, located on the main stem of the Yangtze River, 38 km below the TGR. Its capacity is about of 1.58 billion m 3 , partly operated in 1981. The spawning of CS takes place just downstream of the GR, and many spawning grounds of FMC also are also located in the downstream of the two reservoirs. To investigate the impacts of reservoir construction and operation on the spawning conditions of these two critical taxa, the two key hydrological stations of Cuntan and Yichang were selected as study sites (Figure 1). The Cuntan station, the farthest end of the backwater of the TGR, is located 608 km above the TGR. The Yichang station, an important monitoring station for outflow from the TGR, is immediately downstream of the TGR, approximately 44 km below the TGR.

Data and Methods
Due to the ≈86% runoff at Yichang that comes from Cuntan and the alteration of the flow regime by human activities being limited in the region between Cuntan and Yichang, the Cuntan station just above the TGR was selected as a reference station to investigate the impacts of the TGR on its downstream hydrological regime in the Yangtze River basin. Since the spawning of the FMC and the

Study Region
The TGR is located on the main stem 44 km above the end of the upper Yangtze River. Its operations were done in four stages: the initial stage began in June 2003, with the reservoir water level increased to 135 m above sea level; the transitional stage started in October 2006 with the water level increased to 156 m; the reservoir water level rose to 173 m in November 2008 during the quasi-normal stage; and then to the designed reservoir water level of 175 m in October 2010 during the normal stage with the total storage capacity of 39.3 billion m 3 [31]. According to the operational scheme of the Ministry of Water Resources [32], the TGR stores 22.15 billion m 3 of water to reach normal water level of 175 m from 145 m at the end of flood season (generally from 15 September to 31 October, lasting 47 days; it begins refilling from 10 September if the year is relatively dry). The GR, a run-of-river reservoir, located on the main stem of the Yangtze River, 38 km below the TGR. Its capacity is about of 1.58 billion m 3 , partly operated in 1981. The spawning of CS takes place just downstream of the GR, and many spawning grounds of FMC also are also located in the downstream of the two reservoirs. To investigate the impacts of reservoir construction and operation on the spawning conditions of these two critical taxa, the two key hydrological stations of Cuntan and Yichang were selected as study sites (Figure 1). The Cuntan station, the farthest end of the backwater of the TGR, is located 608 km above the TGR. The Yichang station, an important monitoring station for outflow from the TGR, is immediately downstream of the TGR, approximately 44 km below the TGR.

Data and Methods
Due to the ≈86% runoff at Yichang that comes from Cuntan and the alteration of the flow regime by human activities being limited in the region between Cuntan and Yichang, the Cuntan station just above the TGR was selected as a reference station to investigate the impacts of the TGR on its downstream hydrological regime in the Yangtze River basin. Since the spawning of the FMC and the CS occurs primarily during May-June and October-November, respectively, the time series of daily discharge, sediment concentration, and water temperature during these two periods, both at Cuntan and Yichang, were obtained from the Changjiang Water Resources Commission (CWRC), China ( Table 1). The data homogeneity and reliability have been firmly checked and assured by the CWRC before its release. It should be noted that the daily water temperature during 2001-2006 and 2010-2012 were averaged to a ten-day time scale for this analysis. Due to the scarcity of fish data, only annual numbers of the CS during 1981-2014 were collected from published references [33][34][35][36][37][38][39][40]. All these data for the CS were collected from the same spawning ground downstream of the GR during its spawning season. Based on the long-term historical data of the FMC (1997)(1998)(1999)(2000)(2001)(2002) and the CS  from the Jianli river cross-section downstream of the GR in the Yangtze River, Guo [40] investigated the spawning characteristics of these two taxa in the Yangtze River, and concluded that the suitable discharge, velocity, and water temperature for the FMC's spawning fall within the range of 7500-12500 m 3 /s, 0.2-0.9 m/s, and 18.6-25.5 • C, respectively; for the CS, the suitable discharge, velocity, sediment concentration, and water temperature fall within the ranges of 10,000-15,000 m 3 /s, 1.0-1.2 m/s, 0.1-0.3 kg/m 3 , and 18-20 • C, respectively.
To investigate the impacts of the GR and the TGR on the propagation of the FMC and the CS, the study period was divided into five sub-periods, i.e., pre-GR period (before 1981), post-GR period , post-TGR period (2003)(2004)(2005) for the first stage, 2006-2008 for second stage, 2009-end for the third stage). Frequency distributions of the daily discharge, sediment concentration, and water temperature during May-June and October-November were found and compared among the different sub-periods. The satisfying degrees, i.e., the suitable daily discharge/sediment concentration/water temperature for the two taxa, and the flow rising characteristics during May-June, as well as the falling ones during October-November, were also examined. In addition, the flow rising (falling) difference between Yichang and Cuntan during different sub-periods was also tested using the two-sample t-test at a confidence level of 95% [41]. Definitions of frequency, satisfying degree, and flow rising and falling characteristics were provided as follows.
The frequency of the daily discharge/sediment concentration/water temperature was estimated according to the empirical frequency expectation formula [42,43]: F(x i ) = m n+1 × 100%, where x i is the data from a given streamflow/sediment concentration/water temperature series and F(x i ) is the frequency estimator of x i ; m is the rank of observations x i arranged in rising order, and n is the number of data points.
The satisfying degree (SD) was defined as: . For the FMC, the N suitable was for the days of daily flow or water temperature within the range of 7500-1000 m 3 /s and 18.6-25.5 • C, respectively, while the N total is for the total days from May to June; for the CS, the N suitable was for the days of daily discharge or sediment concentration or water temperature within the range of 10,000-15,000 m 3 /s or 0.1-0.3 kg/m 3 or 18-20 • C, respectively, and N total was for the total days from October to November. The flow rising (falling) characteristics included the rising (falling) count, rising (falling) magnitude, and rising (falling) duration. The rising (falling) count was defined according to the following algorithms: ∆Q i = Q t+i − Q t+i−1 . If ∆Q i remained positive (negative) for at least three days in a row, then one rise (fall) was recorded; and the related rising (falling) magnitude for this rise (fall) was defined as i=m i=1 ∆Q i , while the m days is defined as the related rising (falling) duration for the rise (fall).

Alterations in Flow, Sediment, and Water Temperature Frequency Distributions
Relative to the Cuntan station, it was clear that the effect of the TGR's regulation on the downstream discharge was very strong. The monthly discharge distribution remained relatively stable during the pre-GR and pre-TGR periods, whilst over the post-TGR period, a big drop during July-December occurred with the maximum reduction in September and October, along with an increase over January-June with the maximum increase occurring in May and June ( Figure 2). Figure 3   was defined as ∑ ∆ , while the days is defined as the related rising (falling) duration for the rise (fall).

Alterations in Flow, Sediment, and Water Temperature Frequency Distributions
Relative to the Cuntan station, it was clear that the effect of the TGR's regulation on the downstream discharge was very strong. The monthly discharge distribution remained relatively stable during the pre-GR and pre-TGR periods, whilst over the post-TGR period, a big drop during July-December occurred with the maximum reduction in September and October, along with an increase over January-June with the maximum increase occurring in May and June ( Figure 2).  Relative to Cuntan, significant changes in the monthly sediment concentration occurred at Yichang due to the GR and the TGR. Sediment concentration during January-May and November- was defined as ∑ ∆ , while the days is defined as the related rising (falling) duration for the rise (fall).

Alterations in Flow, Sediment, and Water Temperature Frequency Distributions
Relative to the Cuntan station, it was clear that the effect of the TGR's regulation on the downstream discharge was very strong. The monthly discharge distribution remained relatively stable during the pre-GR and pre-TGR periods, whilst over the post-TGR period, a big drop during July-December occurred with the maximum reduction in September and October, along with an increase over January-June with the maximum increase occurring in May and June ( Figure 2). Figure   Relative to Cuntan, significant changes in the monthly sediment concentration occurred at Yichang due to the GR and the TGR. Sediment concentration during January-May and November- Relative to Cuntan, significant changes in the monthly sediment concentration occurred at Yichang due to the GR and the TGR. Sediment concentration during January-May and November-December was reduced by 50-80% over 1981-2002; a reduction was also found during June-October (3-15%) ( Figure 4). These changes were statistically significant at the 95% confidence level. After the TGR's implementation, sediment concentration in January-December substantially decreased (76-95%), with the decrease being more than 94% during April-June and October-November. The decrease during March-November was also found to be statistically significant at the 95% confidence level. Frequency distributions of daily sediment concentrations in October-November during pre-and post-dam periods are shown in Figure 5. The frequency curve at Yichang generally overlapped with that at Cuntan during 1981-2002, while the curve at Yichang was clearly above that at Cuntan during 1953-1980. During 2003, the curve at Yichang was completely below that at Cuntan, particularly for the higher frequency part. The decrease in the higher frequency part for Yichang was exacerbated over 2006-2008. Since the sediment concentration at Cuntan was rather low after 2009, the decrease at Yichang was not as clear as that during the past two periods. December was reduced by 50-80% over 1981-2002; a reduction was also found during June-October (3-15%) ( Figure 4). These changes were statistically significant at the 95% confidence level. After the TGR's implementation, sediment concentration in January-December substantially decreased (76-95%), with the decrease being more than 94% during April-June and October-November. The decrease during March-November was also found to be statistically significant at the 95% confidence level. Frequency distributions of daily sediment concentrations in October-November during preand post-dam periods are shown in Figure 5.    Water temperature conditions at Cuntan did not change much, as seen by the distribution curves being very close during the three periods. However, water temperature at Yichang experienced a December was reduced by 50-80% over 1981-2002; a reduction was also found during June-October (3-15%) ( Figure 4). These changes were statistically significant at the 95% confidence level. After the TGR's implementation, sediment concentration in January-December substantially decreased (76-95%), with the decrease being more than 94% during April-June and October-November. The decrease during March-November was also found to be statistically significant at the 95% confidence level. Frequency distributions of daily sediment concentrations in October-November during preand post-dam periods are shown in Figure 5.    Water temperature conditions at Cuntan did not change much, as seen by the distribution curves being very close during the three periods. However, water temperature at Yichang experienced a Water temperature conditions at Cuntan did not change much, as seen by the distribution curves being very close during the three periods. However, water temperature at Yichang experienced a remarkable decrease from March to July and an increase from September to February after the TGR was implemented, while a very limited change occurred between pre-and post-GR periods ( Figure 6). The changes during January, March to May, and October to December were all significant at the 95% confidence level. Frequency distributions of ten-day water temperature in May-June and October-November during pre-and post-dam periods are presented in Figure 7. remarkable decrease from March to July and an increase from September to February after the TGR was implemented, while a very limited change occurred between pre-and post-GR periods ( Figure  6). The changes during January, March to May, and October to December were all significant at the 95% confidence level. Frequency distributions of ten-day water temperature in May-June and October-November during pre-and post-dam periods are presented in Figure 7.     remarkable decrease from March to July and an increase from September to February after the TGR was implemented, while a very limited change occurred between pre-and post-GR periods ( Figure  6). The changes during January, March to May, and October to December were all significant at the 95% confidence level. Frequency distributions of ten-day water temperature in May-June and October-November during pre-and post-dam periods are presented in Figure 7.        The satisfying degree kept increasing over the period of 2009-2017 with more water being released from the TGR with the increase of water storage during the period from the middle of September to early November. The lower flows continued to increase and finally fell within the range of 7500-10,000 m 3 /s. In contrast, the relative satisfying degree of discharge for the CS showed an opposite behavior. It diminished slightly during 1981-2002 from the period of 1950-1980 whilst it improved significantly during 2003-2005. It is clear that the reservoir water storage decreased some daily flows to 10,000-15,000 m 3 /s. Nevertheless, the satisfying degree declined again during 2006-2008. Although the daily flow higher than 15,000 m 3 /s was reduced by the reservoir operation, some discharge falling within the suitable range was also reduced by the TGR. Unfortunately, the decreased days of suitable discharge was more than the increased days of high discharge. Thus, the relative satisfying degree of suitable flow significantly dropped during 2009-2017 ( Table 2). Since no publicly admitted information of the suitable sediment concentration for the FMC is available, only the suitable sediment concentration for the CS was examined. Analysis of sediment data showed ( Table 3)

Alterations in Discharge Rising/Falling Characteristics
During the post-GR period, there was an obvious increase in annual falling counts and a mild increase in annual falling/rising duration and rising/counts, along with a remarkable reduction in annual rising/falling magnitude, observed in comparison with those during the pre-GR period (Figure 8). A t-test indicated that only the annual falling magnitude and counts were statistically significant at the 95% confidence level. Since the falling magnitude and counts showed opposite changing trends, it can be considered that the contributions from the changes of these two factors had counteracting effects on the downstream CS's spawning. However, the falling duration showed a moderate increase during the post-GR period; thus, it seems that the GR had a negative impact on the downstream CS's propagation. The GR also had negative influences on the downstream FMC's propagation due to the obvious decrease in rising magnitude since the increase of rising duration and counts was small. It should be noted that the conclusion obtained from the above was based on the assumption that the flow falling/rising magnitude, counts, and duration contributed equally to the spawning of the CS and FMC. Notes: SWT-suitable water temperature, FMC-four major carps, CS-Chinese sturgeon; CT, YCdegree of suitable water temperature at Cuntan and Yichang respectively, YC/CT-ratio of the satisfying degree of suitable water temperature at Yichang to that at Cuntan.

Alterations in Discharge Rising/Falling Characteristics
During the post-GR period, there was an obvious increase in annual falling counts and a mild increase in annual falling/rising duration and rising/counts, along with a remarkable reduction in annual rising/falling magnitude, observed in comparison with those during the pre-GR period ( Figure 8). A t-test indicated that only the annual falling magnitude and counts were statistically significant at the 95% confidence level. Since the falling magnitude and counts showed opposite changing trends, it can be considered that the contributions from the changes of these two factors had counteracting effects on the downstream CS's spawning. However, the falling duration showed a moderate increase during the post-GR period; thus, it seems that the GR had a negative impact on the downstream CS's propagation. The GR also had negative influences on the downstream FMC's propagation due to the obvious decrease in rising magnitude since the increase of rising duration and counts was small. It should be noted that the conclusion obtained from the above was based on the assumption that the flow falling/rising magnitude, counts, and duration contributed equally to the spawning of the CS and FMC. Comparison of flow characteristics between the post-TGR and pre-TGR period shows a clear increase in annual falling magnitude, but a pronounced decrease in the annual rising magnitude, rising/falling counts, and duration (Figure 8). A two-sample t-test indicated that only the falling magnitude and duration were statistically significant at the 95% confidence level. These results also suggest some trade-off between the changes in falling magnitude and duration in contributing to Comparison of flow characteristics between the post-TGR and pre-TGR period shows a clear increase in annual falling magnitude, but a pronounced decrease in the annual rising magnitude, rising/falling counts, and duration ( Figure 8). A two-sample t-test indicated that only the falling magnitude and duration were statistically significant at the 95% confidence level. These results also suggest some trade-off between the changes in falling magnitude and duration in contributing to CS's reproduction because of their counteracting tendencies. Therefore, the TGR had a negative effect on the downstream CS's propagation due to remarkable decline in falling counts. Additionally, the TGR also had a negative impact on the downstream FMC's spawning since the annual rising counts and duration decreased much more than the rising magnitude.

Discussion
Generally, the GR exerted no obvious effects on the frequency distribution of the downstream discharge in both May-June and October-November, whereas the TGR decreased its downstream flow significantly during May-June and October-November. What is more important is that the influence from the TGR was more pronounced in October-November, relative to May-June, since the TGR began storing water from the middle of September to the end of October or early November. Generally, the higher water storage in the TGR reservoir led to a larger decline, with the maximum impact during the TGR's normal stage in May-June and October-November. The GR and the TGR also affected the downstream flow rising and falling characteristics, with a larger impact on the falling characteristics than the rising ones. Correspondingly, the TGR and the GR had negative impacts on the spawning of the CS and the FMC.
Significant trapping of sediment was induced both by the GR and the TGR, particularly by the TGR during the refilling operations due to the higher water level. For instance, the average daily sediment concentration at Yichang reduced from 0.406 kg/m 3 over 1981-2002 to 0.032 kg/m 3 , 0.010 kg/m 3 , 0.004 kg/m 3 during the first stage, second stage, and third stage, respectively, of the post-TGR period. The low sediment in the clear water with high energy can trigger serious riverbed scouring and riverbank collapse in the downstream channels [8,44,45]. In addition, low sediment may also harm the spawning of Chinese sturgeons as they are rather sensitive to the sediment condition during their propagation. The present study suggested that the sediment condition in the downstream of the TGR was not suitable anymore for the spawning of the CS after 2006. A similar conclusion for the FMC by Ban et al. [27] disclosed that the sediment concentration played the leading role in the spawning while its reduction was beyond the limited range. It is worthy to note that the dramatic decline of sediment could also be partly attributed to the sediment trapping above the Cuntan station since several other large reservoirs (i.e., Xiluodu reservoir, Xiangjiabang reservoir, Pubugou Reservoir) were built in recent years. This is one of the most important reasons that the Cuntan was chosen as a reference station to investigate the impact of the TGR and the GR on their downstream hydrological regimes.
The GR had a very mild thermal hysteresis effect on its downstream water temperature for both the May-June and October-November periods. Comparatively, the hysteresis impacts of the TGR were much more significant, i.e., hindering the warming trend in water temperature in the warm season of May-June and impeding the cooling trend in the cooling season of October-November. The thermal hysteresis effects were amplified significantly with increasing reservoir storage volumes. The larger the reservoir capacity, the longer the residence time of the water inside the reservoir, as a result of more severe stratification of the water. Since the spawning of the FMC and the CS are closely associated with the water temperature, the lower temperatures in the water released from the TGR, as a result of the cooling effect, delayed the FMC spawning timing to middle May, and the spawning period was shortened from April-July to May-June [46]. Du et al. [47] also reported that the timing of the water temperature reaching 20 • C (the upper limit of the optimal water temperature for the CS) was postponed by 27 days.  Figure 9). Moreover, the Ministry of Agriculture (MOA) also reported that no CS spawning activity was detected in 2013 and 2014 [48]. This evidence implies that integrated changes/variations in flow, sediment, and water temperature regime due to dam construction and operation have severely affected the CS spawning ground and habitat, leading to a significant reduction in the Chinese sturgeon population. It is worth noting that due to difficulties in fish data collection, only the annual population of adult CS was obtained from different publications. Although these data series may not be very homogenous, they are from the same CS spawning ground during the spawning season, and their variation could serve as a reliable reference for the change in CS spawning habitat between the pre and post-dam construction.
Although these data series may not be very homogenous, they are from the same CS spawning ground during the spawning season, and their variation could serve as a reliable reference for the change in CS spawning habitat between the pre and post-dam construction.
To alleviate the adverse effects of dam regulation, the TGR has advocated releasing an experimental flow increase process to simulate the natural flow in duration, magnitude, and daily increasing rate during the spawning season to enhance the natural population of the FMC since 2011 [23]. These ecological experimental operations have been launched continuously over the past seven years and the 11th ecological operation test was carried out by the Changjiang Flood Control and Drought Relief Headquarter on 19 May 2018 [49]. It was reported that the water temperature and rising characteristics were considered reasonable during this experimental operation. For the spawning of the CS, there is not as much information of ecological experimental operations as that for the FMC. However, the spawning condition of the CS is much more complicated and sensitive to the environment than that of the FMC. Since the CS is in a near-extinct condition, the ecological operation for the CS's spawning should be the highest priority and should at least take flow, sediment concentration, water temperature, and flow falling characteristics into account simultaneously. Therefore, an urgent ecological operation, including at least three sub-operations, i.e., ecological flow, sediment concentration, and thermal condition, is necessary for the GR and the TGR in the near future. Furthermore, we also emphasize the need for long-term monitoring of the FMC and the CS after the TGR commenced operation in order to understand the ecological health responses to hydrological alterations for effective resource management in regulated rivers [50]. Note: Blue square represents the data collected from Ke et al. [33], red dot from Wei et al. [34], blue triangle from Qiao et al. [35], green inverted triangle from Tao et al. [36], purple diamond from Huang et al. [37], left brown triangle from Wu et al. [38], right green triangle from Huang et al. [39].

Conclusions
This study investigated the impact of the GR and the TGR on their downstream spawning of the FMC and the CS in the Yangtze River using frequency distribution, satisfying degree, and rising/falling characteristics of hydrological conditions. Results revealed that the GR had no obvious Figure 9. Temporal variations of adult CS population at the Gezhouba spawning site in the Yangtze River. Note: Blue square represents the data collected from Ke et al. [33], red dot from Wei et al. [34], blue triangle from Qiao et al. [35], green inverted triangle from Tao et al. [36], purple diamond from Huang et al. [37], left brown triangle from Wu et al. [38], right green triangle from Huang et al. [39].
To alleviate the adverse effects of dam regulation, the TGR has advocated releasing an experimental flow increase process to simulate the natural flow in duration, magnitude, and daily increasing rate during the spawning season to enhance the natural population of the FMC since 2011 [23]. These ecological experimental operations have been launched continuously over the past seven years and the 11th ecological operation test was carried out by the Changjiang Flood Control and Drought Relief Headquarter on 19 May 2018 [49]. It was reported that the water temperature and rising characteristics were considered reasonable during this experimental operation. For the spawning of the CS, there is not as much information of ecological experimental operations as that for the FMC. However, the spawning condition of the CS is much more complicated and sensitive to the environment than that of the FMC. Since the CS is in a near-extinct condition, the ecological operation for the CS's spawning should be the highest priority and should at least take flow, sediment concentration, water temperature, and flow falling characteristics into account simultaneously. Therefore, an urgent ecological operation, including at least three sub-operations, i.e., ecological flow, sediment concentration, and thermal condition, is necessary for the GR and the TGR in the near future. Furthermore, we also emphasize the need for long-term monitoring of the FMC and the CS after the TGR commenced operation in order to understand the ecological health responses to hydrological alterations for effective resource management in regulated rivers [50].

Conclusions
This study investigated the impact of the GR and the TGR on their downstream spawning of the FMC and the CS in the Yangtze River using frequency distribution, satisfying degree, and rising/falling characteristics of hydrological conditions. Results revealed that the GR had no obvious impacts on the frequency distribution of its downstream discharge, a weak impact on water temperature in the spawning season of the FMC and the CS, but a significant impact on the sediment concentration in the CS spawning season. Due to GR operations, the satisfying degree of the suitable discharge and water temperature for the FMC propagation increased, while that of the suitable flow, sediment concentration, and water temperature for the CS decreased. The TGR significantly reduced the downstream flow and sediment concentration in the spawning season of these two species, with a very dramatic decrease in the CS's season. The TGR had a significant hysteresis effect on the downstream water temperature by hindering the warming trend of the water temperature during the FMC spawning season while impeding the cooling trend in the CS's season. During the post-TGR period, the satisfying degree change showed a pronounced ascending tendency in the suitable flow while displaying a remarkable descending trend in the water temperature for the FMC; the satisfying degree change presented a significant decreasing trend for the suitable discharge, sediment concentration, and water temperature for the CS, and the importance of each effect on the decrease was ranked as sediment concentration > flow discharge > water temperature. The TGR and the GR had negative impacts on the spawning of the CS and the FMC in terms of the rising/falling characteristics. The impact of the TGR was generally much more significant than that of the GR, and the impact of the TGR increased with the reservoir storage.