5. Water Quantity Evaluation

There are over seventy USGS gauging stations on unregulated rivers and streams in Puget Sound, which are continuously collecting streamflow data. There are over 170 specific metrics that can be used to evaluate different aspects of streamflow. In order to determine which of these is most suitable for Puget Sound, we performed a review of the literature to determine salient management and scientific issues. The management issues of concern and potential indicators are listed below:

These indicators and others were evaluated as described above. A summary of results is shown Table 27, Table 28, and Table 29. There are many possible indicators of Water Quantity that meet the evaluation criteria.

Table 27. Summary of Freshwater Quantity - Surface Water Hydrologic Regime indicator evaluations. The numerical value that appears under each of the considerations represents the number of evaluation criteria supported by peer-reviewed literature. For example, Frequency of flood events has peer-reviewed literature supporting 4 out of 5 Primary Considerations criteria. Details can be found in the accompanying spreadsheets.

Table 28. Summary of Freshwater Quantity – Groundwater Levels and Flow indicator evaluations. The numerical value that appears under each of the considerations represents the number of evaluation criteria supported by peer-reviewed literature. For example, Annual 7-day low flow has peer-reviewed literature supporting 3 out of 5 Primary Considerations criteria. Details can be found in the accompanying spreadsheets.

Table 29. Summary of Freshwater Quantity – Consumptive water use and supply indicator evaluations.

The Puget Sound basin includes at least thirteen major river systems and numerous tributaries, which can be classified as rainfall-dominated, snowmelt-dominated, or transitional (Ross 2006, Cayan 1996, Bach 2002). Rainfall-dominated rivers exhibit peak flows during winter; snowmelt-dominated rivers have peak flows in late-spring and late-fall with low winter flows. Transitional rivers exhibit less pronounced high or low flows in the late-Fall and late-spring, and winter. Hydrologic flow patterns are important both ecologically and in terms of consumptive resources. Alteration of historic flow patterns may cause ecological harm and supply disruptions (Wiley and Palmer 2008, Poff et al. 1997). Hydrologic flow regimes in Puget Sound rivers have been altered through the construction of dams for flood control or power generation, or by changes in land cover and climate. Flows in the Skagit, Nisqually, Green, Skokomish, and Cedar rivers are regulated by dams (Puget Sound Partnership 2009c).

There are over seventy USGS gauging stations on unregulated rivers and streams in Puget Sound. As such, there are ample data available for flow analysis and it is possible to use this data to evaluate streamflow patterns in many different ways. In order to determine which is the best way to analyze the data it is important to consider what are the most significant ecological and management concerns of the region. The bulk of this section presents a literature review that is intended to determine the important management and ecological issues of Puget Sound.

The surface water hydrologic regime of a river or stream can be characterized through measures of magnitude, frequency, duration, timing, and rate of change (Johnson et al. 2007). At least 170 specific metrics have been used to describe specific aspects of the hydrologic regime resulting in the potential for considerable redundancy (Olden and Poff 2003). The most suitable metric, or suite of metrics, is dependant on the specific nature of the question being addressed or the issues that are of greatest management concern (Levin et al. 2008, Westemeier et al. 1998, Walsh et al. 2006).

The Puget Sound Partnership (PSP) has identified the following issues of potential concern related to water quantity in Puget Sound:

• Consumptive use of surface and groundwater;
• Changes in hydrology related to land use;
• Climate change;
• Modification to stream and floodplain habitats (Puget Sound Partnership 2008b)

A stated goal of the management of water quantity in Puget Sound is:

• In-stream flows directly support individual species and food webs, and the habitats on which they depend (Puget Sound Partnership 2008a).

The intent of this section is to describe the process of determining an appropriate set of indicators of hydrologic alteration, which are relevant to management concerns. Indicators will also be screened according to the criteria discussed elsewhere in this Puget Sound Science Update.

The following sections describe a review of the recent literature with geographic focus on Puget Sound. There were two objectives of the literature review: 1) determine which of the indicators of hydraulic alteration would be most appropriate based on the predicted or observed alternations related to land use change and climate change, and 2) determine which aspects of the flow regime are known to be most relevant to the aquatic species in Puget Sound streams and rivers.

Discussions of consumptive water use and habitat alterations are elsewhere.

Indicators of Hydrologic Alteration – Climate Change – Summary

• Analysis of historic streamflow data in the Western United States suggest that spring snowpack is decreasing and streamflow timing is getting earlier in the water year. These trends are apparent despite significant annual and systematic variation associated with the El Niño/Southern Oscillation and the Pacific Decadal Oscillation.
• Temperatures in the Puget Sound region are projected to increase an average of approximately 0.3°C per decade over the 21st century due to climate change.
• Increasing temperatures may lead to decreased spring snowpack, earlier spring runoff, and lower summer flows.
• Climate change associated hydrologic alterations may lead from snowmelt or transition (snow-rain) flow patterns to rainfall dominated flow patterns.
• Decline in snowpack may be problematic for regional water supplies as most systems have been developed base on historic flow patterns (Mote et al. 1999)

Indicators of Hydrologic Alteration – Climate Change – Literature Review

Puget Sound river hydrology may be affected by climate change. Precipitation in the region occurs predominately in the winter months. The accumulation of snow in the mountains is a primary storage mechanism particularly for the snowmelt-dominated and transitional river systems. It has been estimated that upwards of 70% of total stream discharge in the Western United States is from melting snowpack (Cayan 1996). An estimated 27% of summer streamflow of the Nooksack river originates from high-elevation snowshed and glacier-derived meltwater (Bach 2002). Climate change assessments have predicted increased winter and spring temperatures resulting in decreased snowpack storage in the mountains, increased winter runoff as more precipitation falls as rain, and lower summer flows (Barnett et al. 2008, Cayan 1996, Stewart et al. 2004, Stewart et al. 2005, Luce and Holden 2009, Elsner et al. 2009). Climate change may force rivers with snowmelt-dominated and transitional hydrological flow patterns toward rainfall-dominated hydrology (Mote et al. 1999).

Prediction of the regional impacts of climate change on river and stream hydrology can be confounded by typical variation in rainfall patterns, high geographic variability, and land use changes. There are at least two large-scale systems that affect the annual climate variations in the Pacific Northwest (“Washington State County Growth Management” 2007). The El Niño/Southern Oscillation, with a period of 2 to 7 years, and the Pacific Decadal Oscillation (PDO), with an estimated half-period of 20 to 30 years. Warm and cool phases of the El Niño/Southern Oscillation and/or Pacific Decadal Oscillation may result in variations on the order of 1°C for temperature, and 20% for precipitation (“Washington State County Growth Management” 2007). Hamlet et al. (2005) utilized a Variable Infiltration Capacity model to discern long-term trends in spring snowpack from temperature and precipitation variability. They found that downward trends in snowpack associated with temperature were related to widespread warming. Trends of snowpack associated with precipitation were largely controlled by decadal oscillations; climate change effects on precipitation have not been detected (Mote et al. 2008).

Mote et al. (2008) concluded that the primary factor in decreasing snowpack in the Washington Cascades was rising temperatures, consistent with the global warming. The long-term snowpack trends were unrelated to the variability brought about by Pacific oscillations (e.g., PDO).

Casola et al. (2009) investigated the potential impacts of climate change on snowpack by combining future temperature predictions with the estimated temperature sensitivity of spring snowpack (Stover and Montgomery 2001). They utilized four distinct methods to estimate sensitivity and all four converged on a result of approximately 20% loss in spring snowpack per 1°C temperature rise. Analysis of historic and projected temperature data indicated that snowpack reductions over the past 30 years ranged from 8%-16% while future temperature change would result in an 11%-21% reduction in spring snowpack by 2050. However, future trends may not be statistically detectable due to a high level of interannual variability.

Barnett et al. (2008) utilized a multivariate analysis to evaluate the simultaneous changes in average winter temperature, snow pack, and runoff timing in the Western United States (including the Washington Cascades) for the period from 1950 – 1999. They found significant increasing trends in winter temperature, and decreasing trends in snow pack and runoff timing (indicating earlier snowmelt). In order to distinguish natural variation from anthropogenic forcing they evaluated the observations against two separate climate models and found that the hydrologic changes were both detectable and attributable to anthropogenic forcing.

Stewart et al. (2004) investigated historic (1948-2000) and future streamflow timing in snowmelt dominated rivers and streams in the Western United States. They found significant trends towards earlier runoff in many rivers and streams in the Pacific Northwest. Utilizing a ‘Business-as-Usual’ emissions scenario with a Parallel Climate Model, they predicted a continuation of this trend, largely due to increased winter and spring temperatures but not changes in precipitation. In a companion study they further analyzed the trends in streamflow timing with variations of the PDO (Stewart et al. 2005). While streamflow timing was partially controlled by the PDO there remained a significant part of the variation in timing that was explained by a longer-term warming trend in spring temperatures.

Luce and Holden (2009) utilized quartile regression to investigate the trends in streamflow in wet (75th percentile), dry (25th percentile), and average (50th percentile) water years in rivers in the Pacific Northwest. They reported that the highest proportion of significant decreasing trends occurred during the dry years, while there were few significant trends in the high flow years, concluding that the dry years were getting dryer in the Pacific Northwest. This aspect of the trends accounted for much of the increased variability in annual streamflow.

Recently, the Climate Impact Group, part of the Joint Institute for the Study of the Atmosphere and Ocean (JISAO) at the University of Washington performed The Washington Climate Change Impact Assessment. The assessment included analyses on hydrology and water resource management in which they utilized results from 20 global climate models and two emissions scenarios from the IPCC Special Report on Emissions Scenarios (A1B and B1) to evaluate projected changes in spring snowpack and runoff  (Elsner et al. 2009). For the rivers in the Puget Sound basin they found a dramatic decrease in spring snowpack with there being almost no April 1 snowpack by 2080. During that period, river hydrographs progressively changed from transition or snow-rain dominated to rain dominated patterns. There was little predicted change in annual precipitation.

Indicators of Hydrologic Alteration – Climate Change - Relevant Indicators

Based on the review of the literature, the following indicators of hydrologic alteration may be suitable to monitor and evaluate potential changes in the hydrologic regime brought about by climate change:

• Stream hydrographs
• Summer 7-day Annual Low Flow
• Center of Timing (CT) of Annual Flow
• Spring Snowpack (April 1 Snow-Water Equivalents)

Indicators of Hydrologic Alteration – Land Use/Urbanization – Summary

• Puget Sound region has experienced extensive development and urbanization. The population of the 12 counties surrounding Puget Sound was approximately 4.2 million in 2005; it is expected to increase to 5.5 million by 2025 (“Washington State County Growth Management” 2007).
• Land use changes associated with increases in population affect river and stream hydrology. Typical changes include reduced infiltration and increased runoff, increased flashiness, and decrease in summer flows.

Indicators of Hydrologic Alteration – Land Use/Urbanization – Literature review

Alterations in land use can affect stream and river hydrology in various ways (see Poff 1997 and references therein). Urbanization is associated with the increase of impervious surface area, which can result in increases the severity and frequency of peak stream flows by reducing infiltration and increasing runoff; overall annual stream flow volumes are generally not affected (Scott et al. 1986, Pess et al. 2002, Bilby and Mollot 2008, Paul and Meyer 2001, Montgomery et al. 1996, Schuett-Hames et al. 2000, Beamish et al. 1994). Urbanization my lead to lower base flows from reduced infiltration, though this effect can be somewhat offset by a reduction in evapotranspiration from the clearing of trees (Paul and Meyer 2001). The construction of storm drain systems has been implicated as a primary factor in the reduction a base flows (Simmons and Reynolds 1982). Logging of forested lands increases annual flow by reducing evapotranspiration in the watershed though other hydrologic changes such increasing flooding are disputed (Mathews and Olson 1980, Bauer and Ralph 2001, Matzen and Berge 2008). River basin land use alterations may lead to alterations in channel morphology which can exacerbate flooding potential without changes in stream flow (Stover and Montgomery 2001).

Burges et al. (1998) compared hydrology from a forested and a developed basin in Puget Sound lowlands. They found that surface runoff accounted for 12%-30% and 44%-48% of rainfall on forested and developed catchments, respectively, suggesting that the rate of infiltration was much higher in the forested basin. In a similar study, Leith and Whitfield (2000) found an increased streamflow in basins with the most increase in urbanization compared to basins with less development. Moscrip and Montgomery (1997) found an increased flood frequency in streams with urbanized watersheds compared to nearby control watersheds, which had not undergone development.

Konrad and Booth (2002) investigated possible hydrologic effects related to urbanization by evaluating stream flow statistics from ten streams in the Puget Sound basin. They found that the fraction of the year that flow was above average annual flow (TQ,mean) and the maximum annual flow (Qmax) had significant trends in the urbanized basins compared to the rural basins and could be useful in monitoring the effects of urbanization on stream hydrology. They suggested that TQ,mean might be of more practical use. Fleming (2007) analyzed the effects of urbanization by examining stream memory (i.e. the effect of prior stream flow on current discharge) in urbanizing and rural watersheds in the Puget Sound lowlands. He reported that memory decreased in the developed basin over time but not the undeveloped basins, suggesting that flow memory would be a useful measure of development in a watershed, though may be dependent on basin size, with larger basins exhibiting a greater fidelity in memory.

Cuo et al. (2009) utilized a Distributed Hydrology-Soil-Vegetation Model in order to determine the relative effects of land cover and temperature change on the flow patterns in Puget Sound streams. They found that the relative importance of temperature and land cover differed between the upland and lowland basins. In the lowland basins land cover changes were more important and generally resulted in higher peak flows and lower summer flows primarily from increased runoff. Both land use change and climate effects were more important in the upland basins. Climate effects were more important in the transitional zones and resulted in higher winter flows, earlier spring peak flows, and lower summer flows.

Indicators of Hydrologic Alteration – Land Use/Urbanizations - Relevant Indicators

Based on the review of the literature, the following indicators of hydrologic alteration may be suitable to monitor and evaluate potential changes in the hydrologic regime brought about by land use/urbanization:

• Summer 7-day Annual Low Flow
• Peak Flow
• Flashiness (High Pulse Count)

Hydrologic Regime – Ecology 5.5.2.2.5 Hydrologic Regime – Ecology – Summary

• Aquatic species in Puget Sound rivers and streams are generally adapted to historic flow patterns.
• Salmonid species appear to be sensitive to land use changes in watersheds with streams in urban areas being associated with less robust populations of coho compared to forested areas.
• Benthic invertebrate communities appear to be negatively affected by increased flashiness of stream hydrology associated with urbanization.

Hydrologic Regime – Ecology – Literature Review

The alterations of river and stream hydrology can affect aquatic ecosystems by changing physical habitats, disrupting the natural connectivity of habitats, or by facilitating the successful invasion of exotic species (McLeod et al. 2005). Native species may have evolved according to the pressures and timing of natural flow regimes; altering flow patterns may negatively affect those species (World Health Organization 1948). However, it is not always possible to separate the biological impacts of altered river or stream hydrology from the biological impacts associated with the land-use changes that often accompany or force the alteration in hydrology.

Several studies have attempted to evaluate the ecological impacts of altered land use in stream and river watersheds in Puget Sound. Spawner survey data collected by Moscrip and Montgomery (1997) suggested a decline in salmon populations in basins that underwent urbanization, but not in a nearby control basin. Scott et al. (1986) compared fish populations in a urbanized stream with a nearby unaffected control stream and found that while overall fish biomass was similar between the two sample sites there were differences in species composition. The urbanized stream population was dominated by cutthroat trout while the control stream population consisted of a wide array of salmonids, including coho, and non-salmonids.

Pess et al. (2002) performed a broad-scale analysis over 16 years to investigate salmon abundance with land use and habitat in the Snohomish river basin. The proportion of adult coho supported by a particular stream reach was consistent over the course of the study and the median adult coho density was consistently higher in the forested areas compared to the more-developed areas.

Bilby and Mollot (2008) compared the distribution of spawning coho salmon in four Puget Sound rivers with changes in land use between 1984 and 1991. They found that, while the overall numbers of spawning coho changed at all sites, there was an approximately 75% reduction in the proportion of salmon spawning in areas of increased urban land use as well as a smaller decline in areas with increased agricultural land use activities. They suggested that the protection of spawning habitat may be important.

While these studies demonstrate relationships between urbanization and ecology, and urbanization has been shown to affect stream hydrology, there are several other factors, including an increase in contamination input from surface runoff and habitat modification, which likely influence the results (Paul and Meyer 2001). There are several are several other studies which have attempted to elucidate the specific effects of hydrologic changes on in-stream ecology, including fish and benthic invertebrates; these are discussed below.

High flows can affect salmon returns by disrupting redds, increasing deposition of fine sediments and reducing dissolved oxygen transfer, reducing growth rates, or increasing downstream displacement and mortality (World Health Organization 1948). In a Puget Sound stream, egg burial depths were observed to be slightly deeper than typical scour depths caused by flooding during the incubation period suggesting an adaptation to environmental flow conditions (Montgomery et al. 1996). Increases in peak flow due to land development or other causes may then significantly contribute to embryo mortality. Schuett-Hames et al. (2000) also investigated scour depth in two locations in a Puget Sound lowland stream. They observed sediment scour during two storm events with estimated return intervals of 1 and 1.4 years and found that scour depths reached median egg pocket depths at 20% of the monitored sites during the larger storm. This suggests that scour related to high flows may be important in salmon mortality in Puget Sound.

Beamish et al. (1994) identified an inverse relation between anomalously high flows and indicies of production for coho and Chinook salmon in the Frasier River but not for chum, pink, or sockeye salmon suggesting that, at least in some cases, extreme flows may affect survival. They did not identify a causative mechanism.

Greene et al. (2005) utilized standard multiple regression analysis to evaluate correlations between various environmental factors in the freshwater, bay/delta, and ocean habitats and the return rates of Chinook salmon in the Skagit River (Suter 2007). Their results indicated that flood magnitude, as measured though the Flood Recurrence Interval of the peak flow during incubation period, was a strong predictor of the return rate for Chinook salmon; there was a negative correlation between flood magnitude and salmon returns. A bay habitat factor, which was calculated based on measures of sea level, sea level pressure, and upwelling, was also significantly correlated with Chinook return rates.

In order to evaluate the overall effects of anthropogenic changes on salmon abundance, Scheuerell et al. (2006) utilized a multistage model to incorporate population growth, habitat attributes, hatchery operations, and harvest management based on predictive relationships from the published literature (Tallis et al. 2010). Relationships between peak daily flow during incubation period to egg-to-fry survival rate for Chinook or sockeye have been reported for Puget Sound rivers (Cheung and Sumaila 2008, Selecky et al. 2006, Jennings 2005, Failing and Gregory 2003). Although the reported data generally indicate a decrease in egg-to-fry survival with increasing peak flow during incubation period, the apparent best-fit regression (i.e. negative exponential, logarithmic, or linear) varies, demonstrating the uncertainty in the relationship. Battin et al. (2007) utilized the same relationship but also considered the potential limitations on spawning capacity that could be brought about by minimum flows during the spawning period. They found that the model results were relatively insensitive to spawning capacity (and minimum flows).

Summer flows have been shown to be correlated with coho run strength in Puget Sound (Mathews and Olson 1980).

Bauer and Ralph (2001) evaluated the potential utility of incorporating aquatic habitat indicators, including those related to flow regime, into legal standards for water quality. However, they concluded that the effects of low flow on habitat availability was sufficiently well understood to only allow the development of narrative, but not numeric criteria; the relationships between peak flows and habitat were less certain.

Similarly, Poff and Zimmerman (2010) recently reviewed 165 papers to investigate the possibility of developing quantitative relationships between various types of hydrologic alteration and ecological response. While there was a general reported decline in ecological metrics in response to changes in flow metrics, including a general decline in fish abundance and diversity with alterations in flow magnitude, they were unable to support any quantitative relationships.

Matzen and Berge (2008) evaluated the relationship between urbanization and fish populations in Puget Sound lowland streams through the development of a fish index of biotic integrity (F-IBI). Due to the low species diversity characteristic of Puget Sound lowland streams, they utilized several metrics, which were specific to the region; the final F-IBI included a combination six metrics, which showed the strongest correlation to TIA. The authors cautioned against the direct comparison of individual IBI scores, or the value of short-term trends due to the likelihood of spatial or temporal variation that can occur within streams.

There are several studies that evaluate the effects of urbanization on stream condition based on a benthic index of biological integrity (B-IBI). Morley and Karr (2002) investigated the relationships between stream biological condition, as measured by the B-IBI, and the extent and distribution of urbanization, and stream flow in Puget Sound lowland streams. They reported that B-IBI was significantly correlated with urbanization, as measured by percent urban area and percent impervious area in a sub-basin. Further, they found that B-IBI was correlated with measures of flashiness though not peak flow, and relative roughness though not measures of pebble or fine diameter (e.g. D16 or D50). Based on these relationships they argued that benthic invertebrates were a key measure of stream condition, though not necessarily predictive of the condition of fish populations.

Booth et al. (2004) reported similar correlations between B-IBI and percent urbanization, percent imperviousness, and several measures of flashiness. They did not conclude that urbanization would be a good predictor of stream health but rather suggested that levels of urbanization may constrain the potential benthic diversity of a particular stream and that urbanization may affect each stream differently.

Bond and Downes (2003) performed a set of controlled studies and found that flow increases, but not changes in fine sediment transport, were sufficient to disturb benthic communities in streams, though the effects may be dependent on the availability of flow refugia. This is consistent with studies, which suggest that benthic diversity is sensitive to hydrologic alterations brought about by urbanization.

King County investigated the relationships between flow alterations and in-stream ecology in Puget Sound lowland streams through the Normative Flow Project. They used data from a set of locations representing a range of land cover conditions to evaluate the effects of land use on hydrology and biological condition, as measured through the B-IBI and other macroinvertebrate metrics. The hydrologic metrics with the strongest correlation with B-IBI included low-flow threshold pulse events and interval between pulses, high-flow threshold pulse events and total period of the year with high pulses, TQmean, percent of time above the mean two-year flow, and timing of the onset of fall flows. Although none of the hydrologic indicators were good predictor of B-IBI they were able to discriminate the difference between high and low B-IBI values.

Alberti et al. (2007) evaluated the patterns and connectivity of urbanization by performing an empirical analysis of land use intensity, land cover composition, landscape configuration, and connectivity of the impervious area, on B-IBI in Puget Sound lowland streams (Caddy 2002). Their analysis suggested that total impervious area (TIA) explained much of the variance in B-IBI across basins, but other factors such as mean patch size of urban land cover and number of roads crossing a stream could explain part of the variance not explained by TIA alone. They also reported an inverse relationship between the aggregation of forested land and B-IBI suggesting that intact forests are important to benthic diversity.

DeGasperi et al. (2009) performed a retrospective analysis to relate measures of hydrologic alteration that were sensitive with measures of urbanization and benthic diversity, but not sensitive to basin area. They found that high pulse count (the discrete number of high pulses per water year when flow is exceeds twice the average annual flow rate) and high pulse range (the number of days from the first high pulse to the last high pulse in the water year) best fit their evaluation criteria. Their analysis suggested as a basin is urbanized the number of high pulses increase in the winter and are more likely to occur in the summer increasing both the discrete number of pulses and the range. These pulses affect appear to affect B-IBI values.

Although the B-IBI score may be correlated with specific types of hydraulic alteration which specifically affect benthic communities, there is no clear relationship between B-IBI and the condition of vertebrate species (Morley and Karr 2002). Further, the natural variability of biological indices has not been well characterized; large variability may lead to inaccurate determinations of river health (Link 2005). There can be both large and small scale spatial variability as well seasonal and inter-annual variability, all of which needs to be well understood in order to correctly attribute changes in biological condition with physical alteration brought about by anthropogenic activities. Mazor et al. (2009) found fluctuating conditions at sights without obvious changing conditions suggesting that short-term bioassessments may lead to inaccurate conclusions (Link 2005).

A summary of the indicator evaluation in presented in Table 27, Table 28, and Table 29. In summary there is a wide range of possible indicators of the Surface Water Hydrologic Regime, which perform very well under both the Primary and Data Considerations. There is ample data for the region that can be parsed and evaluated in many different ways. It is, therefore, essential to understand the management concern or objective prior to indicator selection to ensure that the indicator is appropriate to the question at hand.

Only a single indicator was evaluated for groundwater levels and flows. It performed well against the Primary and Data considerations. However, owing to subsurface heterogeneity, the spatial variation is often not well understood, nor is it possible to confidently infer condition at on location from based on data collected proximally.

No indicators were completely evaluated for consumptive use and supply. However, a preliminary review suggests that there are good performing indicators, though it may be a time-consuming task to collect and compile the data on a regional scale.