The natural flow regime heavily influences river geomorphological and ecosystem processes, affecting diverse abiotic and biotic components (Poff et al., 1997; Poff and Zimmerman, 2010). Consequently, changes in the flow regime have a potential negative impact on aquatic ecosystems and functions (Bunn and Arthington, 2002), not limited to biodiversity and productivity. Altered hydrology can result in perturbed sediment dynamics, limited habitat formation, poor biogeochemical cycling, invasion of non-native species, and disruption of phenological cues (Poff et al., 1997; Bunn and Arthington, 2002; Grantham et al., 2010; Poff and Zimmerman, 2010; Yarnell et al., 2010). Recently, there has been increased attention on how reservoir operations and water diversions affect the natural flow threatening riverine ecosystem (Grantham et al., 2010; Rheinheimer and Viers, 2015; Yang et al., 2018). Many rivers have already lost their natural conditions, due to reservoir operations such as hydropower, irrigation, and other human activities (Poff and Zimmerman, 2010; Fong et al., 2016; Hayes et al., 2018; Grill et al., 2019; Kuriqi et al., 2019). Although several efforts have been made to manage environmental flows and establish the scientific foundation to understand the mechanism of flow alteration (Richter et al., 1996; Poff et al., 1997; Bunn and Arthington, 2002; Poff and Zimmerman, 2010), adequate studies that have investigated the effect of both reservoir operations and climate change specifically for the San Joaquin River Basins where hydropower generation is predominant do not exist to date.

Understanding the harmful impacts caused by regulation and climate change on flow alteration are equally important to maintain healthy ecosystems. The fourth Climate Change Assessment suggests global temperatures will increase between 3.1 and 4.9°C (5.6–8.8°F) while the water supply from snowpack will decline by two-thirds by 2,100 (Pachauri and Reisinger, 2007). Such changes will significantly influence the timing and magnitude of river flow (Hidalgo et al., 2009; Das et al., 2013; Knowles et al., 2018). Knowles et al. (2018) suggest that projections of streamflow implied by climate models are essential to understand climate change impacts on hydrology and operations, as climate change can further stress already impacted ecosystems. However, the nature of the climate change involves uncertaintities and is not fully understood yet, in a systematic way. Moreover, there is no single climate model that fully captures the intricate details of hydrological processes, however considering historical hydrology might lead to poor environmental water planning (John et al., 2020). Therefore, there is a crucial need to understand the impacts caused by reservoir operations and anticipated climate change, in order to set realistic managed environmental flows and their ecological goals.

Gathering reliable ecological data can be costly and time-consuming. Further, many times existing datasets are insufficient to analyze complex ecological characteristics (Richter et al., 1996; Yang et al., 2018). However, many hydrologic indices have been proposed for different studies to overcome that problem (Richter et al., 1996; Clausen and Biggs, 1997, 2000; Pettit et al., 2001; Olden and Poff, 2003). Richter et al. (1996) offer a statistical approach, the Indicators of Hydrological Alteration (IHA), to characterize the temporal variability in hydrologic regimes, from which biologically relevant statistical attributes can be derived (Gibson et al., 2005; Law, 2013; Huang et al., 2017). These IHA metrics are widely used in assessing the impacts of anthropogenic activities such as dam regulation and agricultural diversion as well as climate change on river flow regimes (Richter et al., 1996, 1998; Magilligan and Nislow, 2005; Yan et al., 2010; Xue et al., 2017; Kakouei et al., 2018, 2018). The use of IHA metrics can improve establishing suitable strategies for planning, conservation and restoration of impacted ecosystems with the anticipated knowledge of flow regimes at the various phases of project implementation, including climate change studies.

The primary goal of this research was to quantify the impact of both reservoir operations and climate change on the river flow of the four major river basin in the San Joaquin River (SJR) in California, United States. For this, we used modeled hydrology, simulated realistic reservoir operations (e.g., flood control rules, instream flow requirements and urban/agricultural deliveries), and evaluated IHA metrics as a proxy for streamflow changes create four scenarios: free-flowing and regulated hydrology, both under past and future climate conditions. This analysis aims to answer the following research questions: 1) How do the flow metrics vary between the past and near future projections (2031–2060)? 2) What changes do they show under natural and regulated flow regimes, in both, past and near future projections? 3) What is the degree of hydrological alteration caused by climate change and river regulation?

To answer these questions, we integrated a water allocation model for the SJR basins. In particular, this research examines the flow regime at the outlet of all these basins, concentrating on ecologically relevant metrics that best represent the main aspects of the flow regimes and their sensitiveness to climate change (Olden and Poff, 2003; Rheinheimer and Viers, 2015; Langhammer and Bernsteinová, 2020). To analyze the statistical significance of differences, we employed nonparametric trend tests for each metric. We further quantified the degree of alteration integrating all metrics as a complementary result, showing a deviated hydrological regime from historical flow regimes at the basin outlet. Finally, we discussed policy and management implications from this study, followed by key limitations and future research avenues.

The San Joaquin River (SJR) is one of California’s two major river systems draining California’s Central Valley into the ecologically sensitive Sacramento-San Joaquin Delta (CWB, 2019). With an area of 41,130 km², the SJR basin yields an average annual surface runoff of about 2 km³ (1.6 million-acre-feet). The SJR comprises nine tributaries, including Cosumnes, Mokelumne, Calaveras, Stanislaus, Tuolumne, Merced, Chowchilla, Fresno, and Upper San Joaquin Rivers. This study considers a highly regulated network of reservoirs and hydropower facilities within the four most agriculturally important basins with the main tributary rivers: Stanislaus, Tuolumne, Merced, and Upper San Joaquin located in the Central Sierra Nevada (Figure 1). These subbasins are highly regulated and are crucial to water management in the state. The main catchment characteristics of each basin are presented in Table 1, and the hydropower facilities within each basin are included in Supplementary Tables S1–S4.

Each basin has a major storage reservoir at its lowest elevation (basin outflow), formed by terminal “rim dams.” These multipurpose reservoirs manage water for flood control, recreation, urban and agricultural demand, environmental needs, and hydropower generation. The reservoirs and powerhouses in the region have a wide range of storage and generation capacities, ranging from less than 12 to 3,085 million m³ and from less than 5 MW to over 500 MW, respectively (Table 1 and Supplementary Tables S1–S4).

The main purpose for regulation in the Stanislaus and Upper San Joaquin basins is hydropower generation, meanwhile the Tuolumne and Merced basins are driven by agricultural and/or urban water supply demands. Due to chronic anthropogenic effects in the past and the construction of both the rim dams and a vast array of high elevation reservoirs, the SJR system has been facing challenges in maintaining river ecosystems (Obegi and Wearn, 2016). Environmental flows in the region are prescribed generally as minimum and sometimes also maximum instreamflows, with flows varying between certain thresholds dependent upon various water year type classifications. Consequently, the State Water Resources Control Board recommended maintaining 30–50% of the natural flows in the SJR and its tributaries to reduce the environmental impacts of flow regulation (SWRCB, 2018). The high elevation reservoirs are operated independently by different utilities, mostly driven by hydropower revenue. The rim dams, on the other hand, as large storage facilities that feed into the SJR, are the focus of restoration for the SWRCB proposal that encompasses the restoration of the lower SJR as well as the Sacramento-San Joaquin Delta located downstream. In addition, the SJR is vulnerable to climate change even though hydropower production tends not to be greatly affected (Madani and Lund, 2009; Vicuña et al., 2011; Ligare et al., 2012; Rheinheimer and Viers, 2015).

This research investigated the hydrological alteration in the flow regime of the San Joaquin River basins over the past 3 decades and the future 3 decades employing ecologically relevant IHA metrics using modeled historical and projected future hydrologic data used as inputs to a water allocation simulation model. We explored the hydrological alteration due to hydropower operation as well as climate change. Though our study has limitations regarding the relatively short period of data analyzed, modeling results, and inconsistent behavior of IHA metrics, some implications from this study can be drawn.

In brief, the flow regimes downstream of hydropower facilities are likely more altered by reservoir operations rather than by climate change. Understanding such changes can help identifying strategies to conserve and restore ecosystems already stressed by human intervention under a changing climate, such as focusing on maintaining stream temperature and creating a favorable habitat downstream (Null et al., 2013; Rheinheimer and Viers, 2015; Yasarer and Sturm, 2016). We observed more stepped and wider flow duration curves at each location, what is even more pronounced under the climate changre scenarios for regulated flows, when compared to unregulated flows. Stepped flow duration curves indicate intermittent flow variation due to reservoir operations in particular the maintenance of static instreamflow requirements that might impact the movement, establishment and environmental cues for aquatic life.

We investigated the variability of IHA metrics among all ten GCMs across the study area, concentrating on the outlet of each basin, downstream of existing powerhouses. Even though the metrics greatly vary, and specific trends are difficult to discern, some notable trends are apparent in specific cases. For instance, the decreased high pulse counts and increased high pulse counts are observed in regulated and to a greater extent in unregulated flows. As such, less high pulses under regulation tend to provide less nutrients and habitat to the aquatic life along the riverbank, the opposite effect that could be promoted by the greater high pulse counts found in the unregulated flows (Xue et al., 2017). We also confirm that a low pulse count tends to increase under a warmer climate, which might cause the loss of native species, as observed elsewhere (Gao et al., 2018; Kakouei et al., 2018). Although most GCMs show decreased high pulse duration under unregulated flow regime, increased high pulse rate under the regulated flow regime leads to unstable river flow due to increased high and low pulse duration (Richter et al., 1996). There is an increase in the rise rate of unregulated flows, which are greater under regulation caused by the dam (Magilligan and Nislow, 2005). These climate change impacts likely caused by flood control rules lead to tradeoffs between environmental water users and economics (Rheinheimer and Viers, 2015). The occurrence of more extremes reservoirs tend to be fuller in the rainy season and emptier in the dry season when surface water is most needed for irrigation and hydropower generation, and at the same time minimum and maximum instreamflow requirements also need to be maintained in both cases.

We also showed that winter flows tend to increase in cooler/wetter climate scenarios, although it might not be suitable for native species (Gibson et al., 2005; Gao et al., 2018). Decreased summer flow might indicate increased stream temperature and reduced dissolved oxygen, which is also unsuitable for local species (Gibson et al., 2005; Yang et al., 2017). Finally, we also observe that warmer climate projections and river regulation lead to decrease in the magnitude of high flows, which may not be favorable conditions for creating new habitats (Poff, 2002). Further, a decrease in low flows caused by regulation may detach the flood plain from the river, resulting in loss of off-channel aquatic life (Gibson et al., 2005). Considering that, SWRCB’s proposal of setting an effective target as a percentage of natural flow might not be the best decision to overcome these challenges, as the natural hydrology tends to be already affected by climate change in the near future. Therefore, environmental water management needs to be more carefully planned as these disturbances are exacerbated under regulation. A more effective ecological conservation can be achieved through the implementation of the functional flows framework, as proposed by (Granhtam et al., 2010; Granhtam et al., 2014), for the for California’s rivers. The functional flow regime retains key components of the natural hydrograph that support biophysical processes across the riverscape (e.g., wet season flood flows and spring recession flows) that provide cues for fish migration and reproduction, for example (Yarnell et al., 2020, 2015).

This work utilized a traditional approach to investigating the effects of climate change on river flow using the Mann-Kendall test. However, there are no clear trends in most metrics in time and space, even among GCMs. The lack of trends in flow metrics is driven by an increase in variance in most cases, and most GCMs imply decreasing trends under the regulated flow regime compared to the unregulated. Nevertheless, the flow regime differs between the regulated and unregulated hydrology under historical and future climate change scenarios. This study shows that reservoir operation’s impact on river flow regimes leads to higher DOA than changes in hydroclimatic conditions in the main tributaries of the SJR. We found that the DOA for regulated flows is more than twice as high as the ones caused by climate change. Arheimer et al. (2017) also observed greater impacts caused by hydropower installation and associated river regulation than climate change in snow-fed river basins as the SJR. However, both the future climate and reservoir operations can influence ecological processes in aquatic ecosystems and conjunctively further alter the functioning of ecosystems (Gibson et al., 2005; Rheinheimer and Viers, 2015).

This study demonstrated the practical application of IHA metrics to categorize both unregulated and reservoir-induced regulations to river flows. In addition to studying IHA metrics, we also assessed the magnitude of extremes. Such enhancement has the potential to recommend flow metrics in any river basin process, such as in-stream flow requirements, drought assessment, and of course, environmental flow recommendations that lead to ecological restoration activities within the river basin (Yarnell et al., 2020, 2015). In addition, this study is applicable elsewhere to investigate flow alteration induced by climate change and other human-induced activities such as urban development, agriculture, and beyond. Finally, such a holistic approach can be combined with ecological models to assist water and land managers in prioritizing ecological protection plans.

The present study focused on the impact of reservoir operation and climate change on river flow. Although the four basins also have agricultural and/or urban water users, this study does not implicitly address the impacts of demand-driven diversions, as they generally occur at or right below the rim dams. We also do not incorporate other ecosystem characteristics, changes in energy demand, and mechanistic relationships between abiotic conditions and biotic quality. The combined effect of dam operation and other anthropogenic activities that impact downstream water quality also need attention in future work.

This study could be improved by integrating information from other climate variables, including other river flow conditions such as water quality and annual water volume that categorize different water year types in detail. Since there are 33 IHA metrics, most are redundant (Olden and Poff, 2003). Nevertheless, they are useful in devising a simpler metric similar to water year types implemented by DWR. Future work could integrate all metrics using a robust machine learning classifier, e.g., Random Forest, support vector machine. Such may be useful to predict the flow metrics 1 year ahead by establishing transition probability matrices among integrated IHA metrics (Puente et al., 2017). Knowledge of future IHA metrics such as monthly average flow and low flow could be useful to daily downscale using a deterministic approach, e.g., Maskey et al. (2019), and opens further advancement with knowledge of daily instream flow requirements for aquatic life.

This study investigated climate change and flow regulation effects on hydrological alteration in the San Joaquin River basin. We adopted a systematic research protocol to concisely interpret the IHA metrics derived below hydropower facilities in the past and future climate. The result revealed flow alterations that can lead to ecosystem degradation caused by river regulation and future climate scenarios implied by ten GCMs. As observed in this study, hydrological alteration in both cases can affect frequency, duration, and change rate. In addition, dam construction led to decreased winter flows, as water is saved to be used later in the year, causing an increase in summer flows, which exacerbate ecosystem disturbance. In both cases, the changes in the flow regime alter the ecological behavior of the river basins. Some alterations may or may not be beneficial. For instance, a decrease in low flow may detach the floodplain from the mainstream, while increased high flow may benefit aquatic life. While hydrological alteration due to climate changes will be gradual, reservoir operations abruptly alter the flow regime.

Our results indicate significant deviations in the future inferred from the trend test and DOA. However, only some models and specific basins show some trends, therefore, the interpretation of IHA is more model-specific and site-specific, and even metric-specific. The higher range of DOA values suggests more attention should be given to impacts from reservoir operations compared to climate change. The flow design suggested by SWRCB can help mitigate the impacts of regulation, but will not avoid the impacts of climate change that will likely already be experienced in the coming decades. Therefore, we suggest the adoption of functional flows as a better approach.

Hydrological changes driven by agricultural deliveries cannot be directly estimated as diversions occur at or right below the reservoirs, impacting water storage but not environmental flows directly. In addition, the river regulation can help ameliorate some of the harmful effects of warming, such as increasing toxic compounds, lowering oxygen content, and reducing desired pH levels. Nevertheless, such adverse impact can be overcome with a more detailed scientific study integrating human and climate change impact on river flow.

We have demonstrated the usage of the IHA framework for investigating changes in river flow regimes associated with reservoir operations and climate change. Findings from this study suggest riverine ecosystem processes are sensitive to changes in IHA metrics, which augment empirical knowledge on ecological response, environmental drivers, and so on to support and establish environmental flow standards. Therefore, similar studies elsewhere offer suitable management decisions in maintaining minimum flow requirements needed to sustain downstream ecosystems below multipurpose reservoirs under different climate scenarios.