How does the daily regulation hydropower station reduce the hydrological regime impact? A case study in upper Yellow River

The natural hydrological regime plays a crucial role in maintaining the ecological health and biodiversity of rivers (Radinger et al. 2018; Cui et al. 2020). However, human activities such as the operation of hydropower stations, irrigation, and land use changes have substantially impacted these regimes, contributing to significant alterations in approximately 24% of the world's largest rivers (Pfeiffer & Ionita 2017; Sarauskiene et al. 2021; Yang et al. 2022). Among these activities, the operation of hydropower stations stands as one of the most influential factors, leading to the disruption of ecological continuity in more than half of the world's large rivers, resulting in extensive river system fragmentation (Wang et al. 2018; Gierszewski et al. 2020; Knott et al. 2024). Given this scenario, it is imperative to thoroughly assess the impact of hydropower stations to support the ecological sustainability of river systems.

The analysis of hydrological regimes is primarily based on runoff data, with numerous studies highlighting the substantial influence of such changes on biodiversity and ecosystem integrity (Ma et al. 2014). Over 170 hydrological indicators have been proposed to characterize hydrological regimes, such as average flow conditions, variability in mean daily flow, skewness in flow and peak discharges, short-term estimates of flood frequency, seasonal distributions of monthly flows, flow and flood frequency duration curves, and time series of annual discharge (Olden & Poff 2003; Lu et al. 2018; Wang et al. 2018). However, due to correlations among some of these indicators, leading to redundant information, the indicators of hydrologic alteration (IHA) method, proposed by Richter et al. (1996) at the end of the 20th century, has facilitated the measurement of complex changes and the assessment of effects on hydrological ecosystems. The IHA method, which evaluates the differences between pre-impact and post-impact periods using 33 hydrological indicators, has significantly enhanced our comprehension of the interplay between flow regimes and river ecosystems (Zuo & Liang 2015; Wang et al. 2023). For instance, Monk et al. (2006) applied the IHA method to explore relationships with macroinvertebrate community metrics in 83 rivers in England and Wales, identifying the magnitude and duration of annual extreme water conditions as the ‘best’ predictors. Similarly, Mwedzi et al. (2016) utilized indicators of hydrological alterations (IHAs) to assess hydrological alterations caused by six dams in the Manyame catchment, Zimbabwe, revealing significant adverse effects on low flows, extreme low flows, and an increased number of zero-flow days. Across various regions, the application of IHA has unveiled diverse changes in hydrological regimes.

To quantify the overall extent of hydrological regime change using various indicators, several approaches have been proposed, such as the range of variability approach (RVA) (Richter et al. 1997), histogram matching approach (Shiau & Wu 2008), histogram comparison approach (Huang et al. 2017), among others. Notably, the RVA method stands out as one of the most widely adopted techniques, likely being the earliest method devised for assessing alterations in hydrological regimes (Huang et al. 2017). This method has proven instrumental in water resource and ecosystem management by effectively delineating optimal environmental flows that balance both ecosystem and human needs objectives (Shiau & Wu 2008). This method, based on the frequency of a hydrologic parameter falling within the target range and assigning equal weight for the parameter value within the target range, offers a simple calculation and low data requirements (Ge et al. 2018). For instance, Hu et al. (2008) utilized the RVA method to assess hydrologic alterations, demonstrating the profound influence of dams on eco-hydrological conditions in the Huaihe River basin (China), particularly highlighting substantial impacts on the timing, magnitude, and frequency of flow, especially during dry seasons. Similarly, by applying the RVA method based on IHAs, Li et al. (2014) analyzed four large-scale controlled hydropower stations in the mainstream of the Yellow River in China, revealing marked differences in hydrological regimes, resulting in an overall hydrological change of 87%. Furthermore, Pfeiffer & Ionita (2017) employed the RVA method to investigate the effects of climate change on the hydrological regimes of the Elbe and Rhine rivers in Germany, revealing significant changes in several hydrological indicators, particularly in the number of runoff reversals. In the RVA method, calendar years are utilized instead of the hydrological year, a modification that has been recently compared and refined (Ramesh & Thampi 2023; Zhou et al. 2023). Ge et al. (2018) enhanced the traditional RVA by incorporating the inherent characteristics of hydrological years, revealing that the conventional RVA method underestimated the alterations in hydrological regimes. Furthermore, the RVA method, when considering the hydrologic year, can more accurately depict changes in IHAs for monthly runoff and extreme runoff. Widely adopted and validated in numerous studies, the RVA method has significantly advanced the assessment of hydrological regime alterations (Pirnia et al. 2019). However, whether the RVA method effectively captures the impact of daily regulation hydropower stations remains uncertain, as previous studies have predominantly focused on the effects of controlled regulation hydropower stations operating on a yearly or multiyear scale.

In addition to the RVA method, various statistical hypothesis tests are commonly applied in hydrological regime alteration studies, including the Mann–Kendall (M-K) test, Spearman rank correlation test, and Pettitt test. The M-K test, a nonparametric statistical test recommended by the WMO (World Meteorological Organization), has found widespread use in hydrology research due to its low data requirement and high accuracy (Serrano et al. 1999; Tian et al. 2019; Guo et al. 2022a). For instance, Pirnia et al. (2019) utilized the M-K test to analyze the Tajan River in Iran, discovering significant decreases in runoff in the rainy season and increases in the dry season following the construction of dams. Fang et al. (2023) applied the M-K test to 77 indicator parameters, including IHAs, finding significant changes in 61 hydrological indicators after the construction of the Liujiaping hydropower station in China.

Given the widespread distribution of daily regulation hydropower stations in China and their cumulative impact on downstream hydrological regimes (Fang et al. 2023), it is crucial to prioritize an in-depth analysis of the effects of daily regulation hydropower stations. Therefore, this study seeks to employ the RVA and M-K methods to quantify the effects of daily regulation hydropower stations on changes in hydrological regimes based on IHAs. The upper Yellow River is chosen as the study area due to the presence of two daily regulation hydropower stations (Nina and Laxiwa), alongside a multiyear regulation hydropower station (Longyangxia). The downstream hydrological station at Guide station provides the necessary flow data to elucidate the impact of daily regulation hydropower station operations on runoff, thus enabling river ecologists to better comprehend and predict the effects of daily regulation hydropower stations on river hydrology.

The study utilizes the observed daily runoff data collected from the Guide hydrological station, spanning from 1954 to 2020, encompassing information obtained from the ‘Hydrologic Data Yearbook’ published by the Yellow River Conservancy Commission (YRCC). Even though the daily operations of hydropower stations do indeed alter the intraday runoff distribution and regimes, this study utilizes daily data instead of hourly data for analysis. This choice is informed by the understanding that ecological impacts are the result of long-term, cumulative changes in runoff (Shu et al. 2010; Binh et al. 2020). In addition, conducting our analysis on a daily basis offers greater practicality for both ecological and water resource management purposes. The controlled basin area spans approximately 1.33 × 10⁵ km², characterized by an average elevation exceeding 3,000 m a.m.s.l. (Figure 1(a)). This region exhibits a plateau continental climate, with an average annual precipitation of 288 mm and an annual average temperature of 7.2 °C. Notably, over 70% of the precipitation occurs between May and September, with temperature ranges spanning from −23.8 to 34 °C (Yang et al. 2010).

The regulation of hydropower stations has substantially altered the hydrological and topographic characteristics of the upper Yellow River region (Wang et al. 2007; Jin et al. 2020). Figure 1(b) presents the annual runoff over the past 70 years, indicating an average runoff of 205.54 m³/s. Notably, prior to 1986 when no hydropower stations were present, the annual runoff demonstrated significant fluctuations, with the maximum difference between annual runoff reaching up to 175.97 m³/s. After that, the difference between annual runoff decreases visibly, the 5-year sliding average flow shows a sudden drop around 1985 and a continued decline for some time thereafter. Nina hydropower station operated around 2004, and the flow showed an upward trend.

In this study, the impact of hydropower stations on the hydrological regime within the upper reaches of the Yellow River is assessed using IHA, the M-K trend test, and the M-K mutation test to characterize the significance of the change in indicators. In addition, the overall influence of hydropower stations on the hydrological regime is evaluated using the RVA method.

Assessment of natural flow regime change is essential for understanding river ecosystems (Zeilhofer & de Moura 2009; Zhao et al. 2012). This study utilizes IHAs to evaluate the impact of hydropower stations on hydrological conditions in the upper Yellow River. The IHA method, as proposed by Richter et al. (1997), is employed to characterize inter-annual runoff variability. The IHA method comprises 33 hydrological indicators that encompass five types of hydrological characteristics within the basin. These indicators are used to analyze changes in hydrological conditions by comparing historical flow regimes of river systems across pre- and post-impact periods (Park et al. 2020). The 33 hydrological indicators are categorized into five groups: (1) magnitude of monthly water conditions, (2) magnitude and duration of annual extreme water conditions, (3) timing of extreme flow occurrences, (4) frequency and duration of high and low pulses, and (5) flow change rate and frequency (Table 2).

As all observed daily flows within our study area during the research periods were greater than zero, the IHAs ‘number of zero-flow days’ is not considered in this study. Instead of using the traditional IHA based on the calendar year, we applied an improved IHA method (Ge et al. 2018). This method is based on the hydrological year, which more accurately captures the fluctuations of hydrological indicators through a full rainy and dry cycle (Ramesh & Thampi 2023). In addition, the minimum 1-day flow in this region typically occurs during the transition between calendar years, which can lead to misleading results if the traditional calendar-based approach is used. For instance, a shift in the timing of such an event from December 31 to January 1 could falsely indicate a significant change. In line with the hydrological patterns of this area, the hydrological year begins on April 1 and concludes on March 31 of the following year.

The IHA method analyzes changes by comparing flow regimes in pre- and post-impact periods, which is normally determined as the influencing time of hydropower stations. In the current study, the period 1954–1986 is defined as the ‘natural period’ and referred to as the P1 period, since there are no hydropower stations or reservoirs. The period 1990–2010 is referred as the P2 period to evaluate the influence from Longyangxia reservoir or hydropower station. Since the construction of Nina and Laxiwa hydropower stations was completed in relatively close in time, and both stations regulate on a daily scale, their influences are considered together here and the period 2011–2020 is referred as the P3 period.

To ascertain the significance of changes in hydrological indicators, IHAs from different periods are compared using the M-K trend test and the M-K mutation test to evaluate the impact of additional regulation from daily hydropower stations on these indicators. Detailed information regarding the M-K test can be found in various studies (Lin et al. 2017; de Oliveira et al. 2021).

This study aims to investigate the impact of daily regulation by hydropower stations on downstream hydrological regimes, which have been affected by multiyear reservoirs, in conjunction with observed runoff data in Guide station, and three subperiods (P1, P2, and P3) are applied in following results and analysis.

Beyond the overall change, distinct differences are detected among IHAs across different groups. Analysis of IHAs within Group 1 shows that the averaged IHAs 6–10 during the P1 period are larger compared to the other two periods, indicating a disruption of the natural seasonal hydrological regime, where the river flow increases in the dry season and decreases in the flood season (Yang et al. 2020; Mezger et al. 2021). Moving to the IHAs in Group 2, the colors between minimum and maximum IHAs present visible differences during the P2 and P3 periods besides the period 2018–2020, because the extreme magnitudes are polished by the hydropower station (Vicente-Serrano et al. 2017). For the occurrence time of hydrological extremes (IHAs in Group 3), the inter-annual difference in the P1 period seems much smaller than the period with hydropower stations' operation. The averaged IHAs in Group 4 are close to 0 in the P2 and P3 periods, differing substantially from the P1 period, suggesting significant impacts on hydrological extremes due to hydropower station operation (Wang et al. 2022; Zhang et al. 2023). Moreover, IHAs within Group 5 show smaller differences between years during the P2 and P3 periods compared to the P1 period, indicating a transition toward more stable general flow conditions. Overall, the operation of hydropower stations has led to significant alterations in hydrological regimes, and the noticeable differences between the P2 and P3 periods indicate that we could not ignore the influences of the daily regulation hydropower stations.

Annual extreme flow plays a pivotal role in maintaining the equilibrium of competition among organisms and in shaping river ecosystems. Table 3 illustrates the mean values of extreme flow in three distinct periods, highlighting a decrease in maximum statistics and an increase in minimum statistics for the P2 period. This pattern reflects the primary impact of hydropower stations, which regulate based on reservoirs specifically designed to alter the hydrological regime in this manner (Brunner 2021; Lazin et al. 2023). Nevertheless, the extent of these changes varies significantly across different statistics and periods. Notably, when the statistics are analyzed within a weekly timeframe, the maximum statistics decrease by over 60% from the P1 to the P2 period. Similarly, Sarauskiene et al. (2021) examined hydrological alterations in Lithuanian rivers and observed a substantial impact of hydropower generation on the maximum discharge of 1-, 3-, and 7-day periods. Such substantial changes can heavily disrupt the distribution of plant and animal populations in floodplains (Junk & Bayley 2008). When incorporating the operation of daily regulation at hydropower stations, the maximum flow notably increases, while the variation is reduced by approximately 40% compared to the results observed between the P1 and P3 periods. The occurrence of high-flow events enhances the connectivity between the floodplain and the main channel, playing a crucial role in delivering nutrients to wetlands and providing fish and other mobile organisms with increased opportunities to move through rivers into floodplains for breeding or into shallower habitats (Xie 2003; Arthington et al. 2010; Guo et al. 2024). As for the minimum flow statistics, a positive relationship is evident between the extent of change and the timescale, consistent for both the P2 and P3 periods. For instance, from the P1 to the P2 period, the minimum 1-day flow increases by 32.16% and further rises to 66.47 and 126.71% for the minimum 7- and 90-day periods. This phenomenon is reasonable and expected, since raising the minimum daily flow is one of the functions of hydropower station-based reservoirs, which aims to meet the downstream domestic and ecological water needs. This difference and impacts are cumulated to be larger with the increased time, as well as more hydropower stations. Contrasting the extent of change between the P2 and P3 periods from the baseline P1 period, a larger average change is apparent in the P3 period. This change could not only enhance water purification but also increase the contact time between air and water, thereby effectively ensuring the safety and stability of the hydrological ecology in the basin (Cheng et al. 2018).

Table 4 illustrates the variation in the count and duration of high- and low-flow pulses, and the operation of Longyangxia hydropower station has greatly reduced the count and duration of both high and low flows. The count of pulses has decreased to less than 10 occurrences, with duration of less than 2 days and a change rate exceeding 90%. High flows are primarily a result of heavy rainfall and snowmelt, and the substantial reduction in pulse count and duration indicates a reduction in downstream flooding, ultimately leading to a more stable channel (Gao et al. 2012; Yang et al. 2020). However, Yang et al. (2020) and Guo et al. (2022a) pointed out that the reduction of high and low counts diminishes hydrological connectivity, hindering the exchange of nutrients and organic matter, and fish need high flows to stimulate reproduction. When comparing the P2 period, the average count of high-flow pulses increased from 7.05 in the P2 period to 47.70 in the P3 period, with the average duration extending to 2.42 days, indicating a more favorable environment for the ecosystem. He (2012) argues that an appropriate increase in the high-flow pulse can mitigate impacts such as the increased water temperature or low oxygen conditions caused by the low flow and provide nutritional supplementation. Moreover, it can also avoid excessive high-flow pulse resulting in various pollutants being washed into the river, causing water pollution. It merits attention that there is a notable reduction in the count and duration of low-flow pulses from the P2 to P3 period. This trend is plausible, given that the daily hydropower station should both meet the power generation demand and ensure ecological water use. In summary, with the inclusion of daily hydropower stations, contrast alteration trend between high and low flow is obtained, suggesting a reduction in low-flow occurrences, as noted by Kuriqi et al. (2021) and Wang et al. (2023). This change becomes more different from natural conditions, but might not be a bad phenomenon. As Rolls et al. (2012) have highlighted, the disappearance of low flow could prevent severe blooms and promote diversity of biological association.

The preceding results demonstrate the temporal variation of IHAs, revealing noticeable changes. To ascertain the significance of these variations, the M-K test method is applied for all IHAs. The P3 period exhibits compounded impacts from both Longyamngxia and the daily regulation hydropower stations, and when delineating the specific impacts attributable to the daily operation of hydropower stations, it is essential to exclude the impact from Longyangxia Station. For this purpose, refined datasets are employed following the approach from Fang et al. (2023).

Clear disparities are evident when comparing IHAs across different periods, signifying shifts in hydrological regimes. To elucidate the degree of change for each indicator and the overall impact of hydropower stations, the RVA method is employed and the findings are presented in the following subsections.

Figure 9(b) presents the distribution of IHAs across change categories. At first glance, the proportion of IHAs in the low and moderate groups is minimal for the P2 period, with a similar distribution appearing between the moderate and high alteration groups for the P3 period. The significant 28.13% decrease in the high alteration category from the P2 to P3 period indicates that the combined operation of daily regulation hydropower stations effectively reduces the substantial impact from controlled hydropower stations.

Table 5 outlines the degree of change for IHAs during the P2 and P3 periods in groups, calculated using Equation (3). Examining the results from the P2 period, IHAs in Group 1 exhibit the lowest degree of change, but still reaches a high alteration, with an overall hydrological change of 82.29% falling into the high alteration category. With the additional regulation from daily hydropower stations, the degree of change in Groups 2 and 3 significantly decreases, with Group 3 experiencing a 34.92% reduction from the P2 period. However, an opposite trend is observed in Group 5, with the change degree notably increasing to 95.8%, suggesting additional damage to hydrological regimes from the supplementary daily regulation of hydropower stations. Although the overall hydrological change degree decreases to 75.24%, indicating that the general hydrological regimes align more closely with natural conditions due to the inclusion of daily regulation hydropower stations, the additional impact observed in Group 5 warrants attention in future applications.

The burgeoning regulation of hydropower stations has sparked substantial debate surrounding the alteration of hydrological and ecological regimes in contemporary hydrological research. This study conducted an analysis in the upper Yellow River, where the hydrological regimes are impacted by three hydropower stations, employing the M-K test and RVA method with the IHAs to assess the influences of these hydropower stations. Through the comparison of impacts from multiyear and daily regulation hydropower stations, the following conclusions have emerged from this study:

• (1) The impact of daily regulation hydropower stations on runoff dynamics is of significant consequence. Upon the integration of facilities for daily regulation with the Longyangxia, which is characterized by multiyear regulation capabilities, several observations have been made regarding the runoff patterns. Namely, the runoff patterns exhibited seasonal variability, and the annual extreme flow values have undergone changes, with their timing now more closely aligning with the natural period. The implementation of daily regulation hydropower stations has, in the majority of instances, led to the imitation of natural hydrological attributes in IHAs.

• (2) The supplementary operation of daily regulation hydropower stations has the potential to mitigate the changes or impacts induced by multiyear hydropower stations on hydro-ecological environments, although not universally. The change degree decreased for 16 IHAs with the additional regulation of daily hydropower stations. According to the M-K test, the total number of IHAs exhibiting a significant trend decreased from 27 to 18 with the inclusion of daily regulation hydropower stations alongside the multiyear regulated hydropower station, and the operation of daily regulation hydropower stations caused mutations in 18 IHAs in the P2 period.

• (3) With the operation of daily regulation hydropower stations (Nina and Laxiwa), the overall degree of change in the hydrological regime in the upper Yellow River decreased to 75.24% from 82.29% when solely operating the multiyear hydropower station, Longyangxia. Nevertheless, the change in IHAs within Group 5 increased by approximately 16.10%.

The findings underscore the complex interplay of multiyear and daily regulation hydropower stations on hydrological and ecological dynamics, emphasizing the need for continued investigation and monitoring to comprehensively understand and mitigate their impacts.

Xue Yang and Fengnian Li: conceptualization, methodology, funding acquisition, writing – original draft. Shi Li and Xiaohua Fu: data curation and analysis, methodology, investigation, writing – review and editing. Jungang Luo and Ganggang Zuo: data curation, funding acquisition and analysis – review and editing. Chong-Yu Xu: review and editing.

The National Natural Science Foundation of China (Grant Nos. 52209035 and 52309034) and the Education Department of Shaanxi Provincial Government (Project No. 21JT032) supported this work.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.