This study investigated how reservoir performance varied across different hydrogeologic settings and under plausible future climate scenarios. The study was conducted in the Santiam River basin, OR, USA, comparing the North Santiam basin (NSB), with high permeability and extensive groundwater storage, and the South Santiam basin (SSB), with low permeability, little groundwater storage, and rapid runoff response. We applied projections of future temperature and precipitation from global climate models to a rainfall-runoff model, coupled with a formal Bayesian uncertainty analysis, to project future inflow hydrographs as inputs to a reservoir operations model. The performance of reservoir operations was evaluated as the reliability in meeting flood management, spring and summer environmental flows, and hydropower generation objectives. Despite projected increases in winter flows and decreases in summer flows, results suggested little evidence of a response in reservoir operation performance to a warming climate, with the exception of summer flow targets in the SSB. Independent of climate impacts, historical prioritization of reservoir operations appeared to impact reliability, suggesting areas where operation performance may be improved. Results also highlighted how hydrologic uncertainty is likely to complicate planning for climate change in basins with substantial groundwater interactions.
In addition to long-standing uncertainties related to variable inflows and the market price of power, reservoir operators face a number of new uncertainties related to hydrologic nonstationarity, changing environmental regulations, and increasing water and energy demands. Anticipated air temperature increases in the Pacific Northwest (PNW) region are projected to generate changes in the timing and quantity of streamflow associated with more precipitation falling as rain rather than snow, shorter winter runoff periods, earlier spring runoff, and longer and drier summers (Chang and Jung, 2010; Tague and Grant, 2009). Longer periods of drought during the summer and shorter wet seasons as a result of climate change may impact the ability of reservoirs to meet water storage needs, environmental flows, and water supply obligations (Payne et al., 2004). The earlier runoff associated with transition from spring snowmelt to winter rainfall may make it necessary to refill reservoirs earlier in the season to ensure adequate releases will be available for summer water supply, potentially increasing flood risk if adequate flood storage is not available in the reservoir when a spring flood arrives.
Climate change is likely to affect basins and reservoir operations differently based on an individual basin's hydrogeology and elevation. Changes in precipitation and temperature patterns for the Mediterranean climate of the PNW are projected to have a limited effect on low flows in surface-water (SW) systems because they already experience very low summer flows (Nolin, 2012) (Nolin and Daly, 2006; Safeeq et al., 2013). On the other hand, mixed SW-GW systems with rain-snow transitions and GW systems that depend on snow accumulation for sustaining base flow are likely to experience greater magnitudes of change in summer low flows due to their dependence on snowpack accumulation and the projected shifts of streamflow to earlier in the season (Safeeq et al., 2013; Tague and Grant, 2009). In the Cascade range of the Pacific Northwest, basins located at the rain and snow transitional elevations, are likely the most sensitive to warming climate due to large projected changes in snow accumulation (Jefferson et al., 2008; Tague et al., 2008) compared to areas at higher elevations characterized by snow precipitation. Such differences in hydrogeology and elevation may impact differently a water resource system's reliability, defined as the probability of failure to achieve some target demand or level of flood protection (Watkins and McKinney, 1995).
In this study, we investigated the impact and importance of climate-related uncertainties and hydrologic variability on reliability and sensitivity of reservoir operations for meeting water resources objectives, given current operating procedures, across basins with two different hydrogeologic settings in the Santiam River Basin (SRB), Oregon. We assessed whether changes in the timing and quantity of water resources could affect the reliability of reservoir systems located in the North Santiam Basin (NSB), with high permeability and large groundwater storage, and the South Santiam Basin (SSB), characterized by low permeability, little groundwater storage and rapid runoff response. More specifically, we evaluated: (1) how the current reservoir operation performance for flood management, hydropower production, water supply, and environmental flows changes under future 2.5, 50 and 97.5 percentile streamflow projections for the two hydrologic regimes; (2) which operating system (NSB or SSB reservoirs) is more sensitive to hydrologic variability, and; (3) the sensitivity of different elements of reservoir operations to climate variability. We evaluated and compared hydrosystem reliability for: (a) Simulated Historical (SH) time period (1960–2000), (b) Near Future (NF) time period (2030–2060), and Far Future (FF) time period (2070–2100). This analysis of the reliability, sensitivity, and uncertainty of two hydrologic regime systems under a changing climate was undertaken to provide water managers information about plausible future water resource system capacities and limitations when developing adaptive and responsive water management and water allocations.
The Santiam River Basin (SRB) encompasses approximately
4700
We focused our study in two reservoir systems that both include coupled flood control and re-regulating dams, located in sub-basins with different hydrologic systems within the SRB: Detroit and Big Cliff located in the North Santiam Basin dominated by the High Cascade geology, and Green Peter and Foster located in the South Santiam Basin dominated by the Western Cascade geology. While the primary operating objective for both dams is to reduce flooding during winter and spring, the reservoirs also provide hydropower, recreation, and regulate water quality (Risley et al., 2012).
The North Santiam sub-basin drains approximately 2000
Detroit dam is located at river km 98 on the North Santiam River. It
maintains 561
The South Santiam sub-basin drains 2700
Green Peter dam, with inflows from Quartzville Creek and the Middle Santiam
River (MSR), and Foster dam, with inflows from the South Santiam River, are
located in the SSB. Both Green Peter and Foster dams provide flood control,
power generation, water quality, and recreation benefits. Green Peter dam is
located at river km 9 on the Middle Santiam River, with a storage capacity of
528
We applied streamflow projections (Hamlet et al., 2010; Surfleet and Tullos, 2013Estimates of future water supply) as inputs to a reservoir operation model (HEC-ResSim) to analyze reservoir system reliability under future climate (Fig. 2). We evaluated reservoir performance sensitivity to hydrologic variability as the change in the ability of a reservoir to (a) store a flood of a certain magnitude, (b) maintain downstream control points below bankfull, (c) refill to the top of Conservation pool, (d) meet environmental flow targets, and (e) produce maximum hydropower capacity. A system is considered to be sensitive to changes in climate when reservoir performance is projected to increases or decreases in the future.
To assess the effects of climate change on various objectives of reservoir operations, we applied streamflow projections from two hydrologic models as inputs in HEC-ResSim (USACE, 2013), a reservoir operation model. Inflows for the SRB were obtained from GSFLOW, a coupled groundwater-surface water flow model. Inflows for the other reservoirs in the WRB were obtained from Variable Infiltration Capacity (VIC), a spatial-distributed surface water model (Liang, 1994). Climate change projections for the basin were simulated within GSFLOW (Surfleet and Tullos, 2013) for the SRB and within VIC (Hamlet et al., 2010) for the WRB using the same eight GCMs, GHG emission (A1B and B1), and downscaling method (Delta-Hybrid method). We applied GSFLOW simulations for the SRB because these simulations, available only for the SRB, include a groundwater component and distributions of streamflows to represent the uncertainty attributed to hydrologic modeling parameters in GSFLOW simulations. The groundwater model within GSFLOW was applied only for the sub-basins in the High Cascades and the alluvial geology (Fig. 1) due to the substantial groundwater interactions that occur in those areas. For computational efficiency, only the surface water model was simulated for sub-basins draining the Western Cascades due to the limited groundwater interactions there. Subsurface flows were not transferred as surface water flow to lower sections in the basin based on the assumption that the groundwater flow is stored in deep storage and did not appreciably contribute to streamflow in the Western Cascades. Across the three hydrogeologic settings of the SRB, posterior distributions of hydrologic model parameters were developed for both dry summer and wet winter seasons using a formal Bayesian parameter approach, the Differential Evolution Adaptive Metropolis (DREAM). Five hundred of the parameter combinations with the best fit for each GCM and GHG emission scenario were used to obtain the 2.5, 50 and 97.5 % daily values. Please see Surfleet and Tullos (2013) for further details on GSFLOW modeling and DREAM analysis. The ensemble mean of all the GCMs, as the lower, median and upper confidence interval from 1960 to 2100, are used as inflows in HEC-ResSim for the SRB. For the rest of the WRB we used the median ensemble mean of all the GCMs from VIC projections. The scenarios evaluated include A1B and B1 GHG emission scenarios for Simulated Historic (SH) time period (1960–2000), Near Future (NF) time period (2030–2060), and Far Future (FF) time period (2070–2100).
We applied the same rule curves implemented in the US Army Corps of Engineers' (USACE) 2010 Willamette Basin HEC – ResSim model, which includes Biological Opinion (NMFS, 2008) operations for spring and summer flow releases for the seasonal life histories for Chinook and Steelhead, in addition to winter flood control operations from the Water Control Manuals (WCMs) for each project. The reservoirs are operated by a set of operation objectives or rule curves (Fig. 3) originally designed (USACE, 1953, 1968a, b) based on assessments of natural variability, historical streamflow records, design storage capacity and the minimum releases. Reservoir release decisions are based on a set of rule curves within a zone that schedule releases from the lowest to the highest priority. There are five zones in ResSim: top of Dam, Flood Control, Conservation, Buffer, and Inactive. Each zone is based on pool storage and elevation levels for each day of the year. HEC-ResSim calculates a reservoir's release at each time step to meet the highest priority rule called Guide Curve (GC), which is the Conservation Pool Rule Curve for the analysis presented herein. When the reservoir's pool elevation is above the GC, within the Flood Control (FC) zone (Fig. 3), the reservoir will release more water than is entering the pool. In contrast, when pool elevation is below the GC, the reservoir will release less water than is entering to the pool.
A reservoirs' storage and release schedule varies for each reservoir
(Fig. 3). Detroit reservoir starts releasing water in September to create
storage capacity for flood control, dropping the reservoir elevation from 477
to 442
Since the two reservoir systems in the SRB, Detroit/Big Cliff and Green Peter/Foster are part of the (USACE) thirteen multipurpose dams and reservoirs in the WRB (Fig. 1, right inset) they all operate as a system to maintain downstream control points (e.g. Salem) below bankfull by storing water. While bankfull stage is considered to be a non-damaging level, it is a stage where action is required (USACE, 2011). Thus, reservoir releases depend on the river stage at the downstream control point with the highest priority. For the WRB, and thus the SRB, the Salem control point on the mainstem of the Willamette River (Fig. 1, right inset) has higher priority over the upstream Harrisburg and Jefferson control points, which contribute discharge to the Salem control point. The control point at Jefferson is located below the confluence of the North Santiam and South Santiam rivers and thus is regulated by both the NSB and SSB reservoir systems. If the stage at Jefferson goes above bankfull, operators will regulate releases from the Detroit-Big Cliff complex before regulating releases from Green Peter and Foster. Flows at Jefferson are usually regulated to bankfull stage by reducing releases from Detroit long before it is necessary to control releases from Green Peter and Foster. Green Peter reservoir provides the principal flood regulation in the SSB (USACE, 1968a). Foster serves as a re-regulating reservoir for power peaking at Green Peter and has limited capacity to store high winter floods from Green Peter releases and flows from the South Santiam River at Cascadia (USACE, 1968b), resulting in historical flows at Waterloo often being at or above bankfull levels.
Hydropower is generated at all four of the dams, and the maximum power release rule curve is always the top priority rule in each of the five zones in each reservoir. Releases are prioritized through the penstocks, as opposed to the spillway and re-regulating outlets, to generate power during regulation for flood control and environmental flows.
To investigate the nature and importance of climate-related uncertainties and hydrologic variability in the context of dam operations, we evaluated the reservoirs' operational performance under the 2.5, 50, and 97.5 percentiles of streamflow projections. Reservoir performance measures were chosen based on reservoir primary functions, including flood risk, hydropower production, environmental flows and probability of refill. The uncertainty related with streamflow projections, and thus with each metric, is represented by the error bars as the range between the 2.5 and 97.5 percentile output. The 2.5, 50, and 97.5 percentile values for each metric were calculated from the outflows and reservoir elevations generated from simulations of the entire study period using the 2.5, 50, and 97.5 percentile inflows to the reservoirs.
We analyzed the reliability of flood risk reduction
using two measures, one based on the adequacy of the reservoir capacity for
storing floods of different recurrence intervals, and a second based on the
frequency of flooding at downstream control points in the systems. The
adequacy of the flood storage capacity was evaluated as the ability of the
reservoir to store a 3-day annual flood event of a 1-year (1
To evaluate the frequency of flooding at downstream control points, we
evaluated the time reliability of flood control (
We calculated reservoir refill as the percentage
of the Conservation pool elevation achieved by the beginning of the
Conservation season: 4 May for Detroit; 9 May for Green Peter, and 30 May for
Foster (Eq. 3). A reservoir was considered to be 100 % refilled if it
achieved maximum Conservation pool elevation by the beginning of the
Conservation season. The percentage of reservoir pool elevation was
calculated for each year and then averaged by decade.
To determine the frequency that the system
does not meet minimum spring and summer flow targets over a period of time,
we calculated the time reliability (Hashimoto, 1982; McMahon et al., 2006;
Milutin and Bogardi, 1997) for spring (
To analyze the ability of reservoirs to produce the maximum amount of energy the power plants are capable of producing
over the course of an average year (efficiency) and its sensitivity to climate variability, we calculated the ratio of
averaged annual power generated to generation capacity (Eq. 6) at each reservoir, where power generated is estimated
from the head and discharge at each time step (Eq. 7).
We first provide an overview of hydrologic projections in the SRB and then present results on the impacts and uncertainties of streamflow changes for reservoir performance measures.
Streamflow projections from GSFLOW simulations (Fig. 4) for the SRB indicated the two sub-basins will undergo similar responses to projected warming, characterized by increases in winter flows and reductions in summer flows relative to simulated historic hydrology. However, the degree of differences varied between the basins. For example, increases in December median inflows, relative to historical flows, were projected to be 17 % higher at Detroit reservoir in the NSB (Fig. 4a) than at Green Peter reservoir in the SSB (Fig. 4b). Conversely, reduction in August median runoff was projected to be 13 % higher at Green Peter reservoir than Detroit reservoir. Additionally, streamflow projections suggested that uncertainty in streamflows were higher during the winter months (Fig. 4c–d) compared to the summer months at both locations, and higher uncertainty was projected for NSB streamflows into Detroit reservoir relative to SSB inflows to Green Peter reservoir.
The change in the magnitude of floods draining into the reservoirs (Fig. 5)
was projected to vary with
The ability of Detroit and Green Peter reservoirs
to store a three day event of a particular recurrence appeared to be high
now and in the future (Fig. 6). Despite the projected changes in the size and
frequency of smaller floods entering the reservoirs (Fig. 5), impacts of
warming on the flood storage ratio were negligible. The ratio remained below
one at both Detroit and Green Peter under all time periods and scenarios,
indicating that Detroit and Green Peter reservoirs will be able to reliably
store the analyzed floods under the simulated future. The flood storage ratio
remained constant into the future, presumably because increases were only
projected for floods of small magnitude, which are generally easy to
regulate. Like the inflows (Fig. 4), uncertainty in the flood storage metric
was high for the NSB and very low for the SSB. While the range between the
2.5 and 97.5 percentile predictions for the flood storage ratio at Green
Peter was close to zero, Detroit ratios for the 2.5 and 97.5 percentile were
Under all time periods, the control point at Waterloo in the SSB was
projected to experience higher risk of winter flows exceeding bankfull stage
than other control points in the SRB (Fig. 7). Simulated river elevations at
the Jefferson control point, located on the mainstem of the Santiam River,
and the Mehama control point, located in the North Santiam River, were below
bankfull stage under all time periods and scenarios. In contrast, river
elevations at Waterloo, located in the South Santiam River, exceeded bankfull
stage during at least a few years under all time periods. When uncertainties
were considered, Waterloo bankfull target was exceeded for 18 of
40
For both the simulated historical and future inflows, the reservoirs did not reliably refill to maximum Conservation pool (Fig. 8) by their respective deadlines in May (Fig. 3), and the impact of a warmer climate appears to be negligible, particularly when uncertainty is considered. For both historical and future scenarios, while the reservoirs failed to reliably refill by their May deadlines, they often reached water levels very close to maximum Conservation pool (Fig. 9) and refilled within 15 days of the refill deadline in 90 % of the years, based on median runoff scenarios. Relative to historical, the future appeared to have an initially higher but declining refill reliability, though the differences were all within the range of uncertainty. Thus, despite not refilling by the deadline each year, the reliability of reservoirs to eventually refill, both in the past and future, was high and does not appear to be appreciably impacted by a warming climate.
Some variability between basins was observed. While Detroit reservoir in the
NSB may never refill during a dry water year (e.g. 1996 for the simulated
historical time period) (Fig. 9a), reaching only
Results indicated that the reliability of meeting
spring flow targets (Fig. 10) was generally high under both historical and
future scenarios and in both the NSB and SSB, though reliability was lower in
the NSB when uncertainties were considered. While both basins met spring flow
targets every year for the SH time period, the NSB did not meet the spring
flow targets in the NF and FF time periods for the 2.5 percentile flows for
A1B scenario and in the NF for B1 scenario. The lower reliability in the NSB
was associated with higher uncertainty for the NF_A1B scenario, where 13 out
of 30
Reservoirs' ability to meet summer flow targets and the uncertainty in those
estimates, varied across the two basins, but projections indicated that
decrease in summer flow reliability may occur into the future for both basins
(Fig. 11). From the simulated historical record, summer flow targets were met
in 100 % of days for the SSB, while both the number of days of
inadequate flows and the uncertainties in those estimates were higher in the
simulated historical NSB. With failure defined as a year in which all
confidence intervals for the number of days below a target were non-zero, the
SSB failed to meet summer targets in 2 of the 30
The impact of a warming climate on the
reliability of producing hydropower appeared as a decline in power
production, though the effect was within the uncertainty limits of the model
(Fig. 12). For the simulated historical period for the median flows, the NSB
reservoirs operated at between 40–50 % of maximum power production.
This range appeared to drop to 30–40 % for by the FF time period,
though the differences were generally within the lower confidence interval of
the simulated historical data. The SSB reservoirs operated at
By applying a reservoir operations model to distributions of simulated future runoff impacted by climate change, we found limited evidence of a response in reservoir operations performance to a warming climate. Despite projected increases in winter flow and decreases in summer low flows, only the ability to meet summer flows in one of the two study basins was conclusively impacted by the simulated future climate, suggesting that reservoir operations may adequately accommodate hydrologic changes in the Santiam River basin, without compromising the ability to meet operating objectives. However, independent of climate impacts, the results highlight areas where operations performance may be improved and how hydrologic uncertainty may impact uncertainty in evaluations of reservoir performance.
While some studies have suggested the need to modify reservoir operations to
mitigate the effects of climate changes (Watts et al., 2011) or to reduce the
impact of climate change on water systems (Vonk et al., 2014; Watts et al.,
2011), our results indicated that impacts to reservoir operations in the
Santiam River were limited. To review, projections indicated that summer
baseflow could decrease while winter runoff could increase (Fig. 4). This
modified hydrology resulted in projected future increases in the magnitude
and frequency of small floods (i.e. 1
Regarding the comparison in sensitivity between the two basins due to hydrogeology, the three distinguishing features between the basins were the sensitivity of the SSB to the hydrologic changes associated with summer low flow, differences in prioritization around flood risk reduction, and the uncertainty in streamflow in the NSB, which lead to uncertainty in several of the reservoir performance metrics. Regarding sensitivity to summer low flow, only the ability to meet summer environmental flow targets appeared to decline in the future (Fig. 11) and only for the SSB. This discrepancy between the NSB, with higher elevations and greater groundwater connectivity, and the SSB, with a more limited snow zone and more rapid runoff, is consistent with other studies (Nolin and Daly, 2006; Safeeq et al., 2013) that found summer low flows in basins at higher elevations with snow precipitation may be less sensitive to changes in climate than basin at lower elevations located along the rain-snow transition zone (Fig. 4). However, this discrepancy between the NSB and SSB summer flow target reliability may also be related to the high uncertainty in streamflow projections in the NSB, which generated higher uncertainty in the reliability of meeting summer flow targets. Regarding prioritization of flood risk reduction, existing operating priorities in the basin appeared as higher flood risk at Waterloo than at other control points in the Santiam River basin and lower hydropower production for the SSB, relative to the production capacity, at the reservoirs in the NSB. These results suggest that operating policies and priorities may need review, independent of impacts of climate change. Finally, the uncertainty in streamflows for the NSB has implications for waters managers seeking to evaluate the reliability of the water resources of the basin. Relationships between climate projection uncertainty, system reliability and system sensitivity to climate variability suggests that reservoir systems located in basins with groundwater interactions are more unpredictable than reservoir systems located in surface water basins. Higher uncertainty for groundwater basins compared to surface water basins is likely a result uncertainty associated with transfer of model parameters in the groundwater model (Rosero et al., 2010; Surfleet and Tullos, 2013). Thus, it is likely that the large uncertainty for other basins with substantial groundwater interactions and snow cover will be similarly challenged by high uncertainty in model projections.
This top-down climate change assessment was conducted to evaluate the impact and importance of climate-related uncertainties and hydrologic variability on reliability and sensitivity of reservoir operations in basins with contrasting hydrologic conditions. Key factors that may have impacted our results included modeling uncertainties around groundwater recharge and discharge. As described and justified in the Methods, groundwater was only modeled within GSFLOW where substantial groundwater interactions occur (High Cascades) and subsurface flows were not transferred as surface water flow to lower sections in the basin. Despite a generally high model fit (Surfleet and Tullos, 2012), this model configuration may have contributed to an underestimation of groundwater contributions to summer baseflow on the NSB. In addition, we acknowledge that our analytical approach assumed stationarity in relationships and interactions between climate and the landscape, as well as reservoir operations and priorities. This assumption may not be appropriate for some types of analysis, such as the design of hydraulic structures (Obeysekera and Salas, 2014). However, for the purpose of identifying key differences in the sensitivity of reservoir operations and priorities to a warmer climate, we do not believe the stationarity assumption significantly impacted our key findings.
Given that reservoir systems' sensitivity to climate variability can be influenced by basin hydrogeology, operating rules, and available storage, we assessed the impact, sensitivity, and uncertainty of changing hydrology on hydrosystem performance across different hydrogeologic settings. We evaluated the changes in future performance of reservoirs in the Santiam River basin (SRB), including a case study in the North Santiam Basin (NSB), with high permeability and extensive groundwater storage, and the South Santiam Basin (SSB), with low permeability, little groundwater storage and rapid runoff response. Key findings included: (1) Projected reductions in summer flows and increases in winter flows for both basins, but at levels small enough that reservoir performance did not appear to be impacted, except in summer flow targets for the SSB; (2) The hydrologic uncertainty in the NSB resulted in uncertainty in the reliability of reservoir refill, spring and summer flow targets, and hydropower production, indicating that water resources may be less predictable in basins with substantial groundwater interactions; and (3) Irrespective of climate change, historical prioritization of reservoir operations appeared to impact reliability, suggesting review of operations may be warranted to consider how flood risk could be reduced at Waterloo and power production could be prioritized on the NSB. Results highlighted how summer flows may be vulnerable to climate change in surface water basins, but that large changes may be required for other operating objectives to be impacted. In addition, hydrologic uncertainty is likely to complicate planning for climate change in basins with substantial groundwater interactions. Finally, assessment of climate change impacts may support the identification and modification of existing inefficiencies in system operations that are independent of a warming climate.
The study was conceived and designed by Cristina Mateus, with substantial input from Desiree Tullos, and all modeling and data processing analysis was conducted by Cristina Mateus. Both authors contributed equally to interpreting results and writing the manuscript.
This material is based upon work supported by the National Science Foundation under Grant No. 0846360. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. The research was supported in part by the National Science Foundation through TeraGrid resources provided by Purdue University under grant number TG-ECS100006. Funding was also provided by Secretaria Nacional de Educación Superior, Ciencia, Tecnología e Innovación SENESCYT, ECUADOR.
Reservoir characteristics.
Left inset: santiam River Basin (SRB), reservoirs and geology. Right inset: willamette River Basin Reservoir Network. Thirteen multipurpose dams and reservoirs (in bold) work as a system to meet downstream flow targets at control points (in italic). The arrows indicate the direction of the flow, the black dots represent stream nodes in the stream alignment, the black dots with gray circles represent computational points where streamflow projections are added to ResSim model, and the black dots with gray boxes represent control computational points for reservoir operation.
Study approach.
Santiam Basin reservoir rule curves.
GSFLOW streamflow inputs at Detroit reservoir and Green Peter
reservoir. Figures
Percent change from historic in the size and frequency of
peak daily inflows (median) of 1, 2, 5, 25,
50, 100 and 200
Flood to storage ratio represented as the ability of a reservoir, on any given day to store a three day event of a particular recurrence interval was calculated for Detroit, and Green Peter reservoirs for the Simulated Historical (SH), Near Future (NF), and Far Future (FF) time periods under A1B and B1 GHG emission scenarios. A higher ratio means a potentially larger failure to store high flood events.
Time reliability of flood control at downstream control points represented as the number of days flood exceeded at Jefferson in the mainstem of the Santiam River, Mehama in the North Santiam River, and Waterloo in the South Santiam River for the Simulated Historical (SH), Near Future (NF), and Far Future (FF) time periods under A1B and B1 GHG emission scenarios. Error bars represent the upper and lower confidence interval.
Reservoirs ability to refill by decade to maximum conservation pool showed as percentage of water stored by 4 May at Detroit, 9 May at Green Peter and 30 May at Foster during the Simulated Historical (SH), Near Future (NF), and Far Future (FF) time periods under A1B and B1 GHG emission scenarios. Error bars represent the upper and lower confidence interval.
Reservoir (median) pool elevation and storage for a dry (left column) and wet (right column) water years during the Simulated Historical (SH) time period for Detroit, Green Peter, and Foster reservoirs.
Spring flow target reliability. This figure shows the number
of days (
Summer flow target reliability at Mehama in the North Santiam
basin and Waterloo in the South Santiam basin represented as the
number of days (
Hydropower production represented as reservoirs' ability to
produce the total power capability in a given year under the A1B GHG
emission scenario. Error bars represent the upper and lower
confidence interval. Scale for the