Recommended Flow Regime for the Cache la Poudre River Through Fort Collins, Colorado John Bartholow 1 DRAFT 9/18/2008 Abstract I combined four methods to develop and refine month-by-month flow recommendations to protect and partially restore the ecological integrity of the Cache la Poudre River through Fort Collins, Colorado. The Range of Variability Approach was used to summarize the river's historic flow regime and set provisional monthly flow targets at the lower end of each month's flow range – specifically at levels that were historically exceeded 75 percent of the time. These preliminary monthly targets were fine-tuned using results from three Poudre-specific studies. A substrate maintenance flow model was used to better define the high flows needed to flush accumulated sediment out of the river's channel and help maintain a healthy and sustainable riparian zone in this snowmelt-dominated river. A hydraulic/habitat model and a water temperature model were both used to better define the minimum flows necessary to maintain a healthy cool water fishery. The result is a range of recommended monthly flows depending on the year type: wet, average or dry. Without exception, the recommended monthly flow ranges are all above what has been recently measured in the Poudre River, signaling that the river is currently overallocated with respect to its ecological well-being. I also provide additional guidance to minimize large and disruptive short term flow changes and offer a few thoughts on non-flow related topics. Introduction and Background on the Methods A river's flow regime (the annual runoff distributed over time) is the key driver to successfully maintain river, floodplain, and stream margin wetland ecosystems (Poff and others, 1997; Bunn and Arthington, 2002; Nilsson and Svedmark, 2002; Whiting, 2002). The major mechanisms linking the flow regime to ecosystem consequences are well understood, although predicting and quantifying the exact responses of any given flow alteration in any specific river reach is not yet within the grasp of the scientific community. For this reason, developing flow recommendations for a river reach can be complicated, costly, and rife with technical concerns (Annear and others, 2004). In some cases, millions of dollars have been spent over decades and controversy still lingers regarding the accuracy, appropriateness and comprehensiveness of the methods used, the equity of the allocation decisions reached, and whether further adjustments will be necessary as more is learned. It is therefore unlikely that sufficient time, money and expertise are all available to develop a scientifically unassailable flow recommendation for the Cache la Poudre River through Fort Collins, Colorado. Even if time, money and expertise were not 1

U.S. Geological Survey, retired. 5402 Old Mill Rd., Fort Collins, CO 80528, 970-223-6488 This paper is available at http://fossilcreeksoft.googlepages.com/poudreriverflowrecommendation

1

obstacles, it is doubtful that a single 'best' objective flow prescription could be identified. However, there are certain well-accepted statistical techniques that can be used to develop a reasonable initial recommendation given sufficient data (Annear and others, 2004). The objective of these approaches is not to achieve a fully 'natural' or 'pristine' river, but rather, in the words of Colorado's instream flow legislation, "to preserve or improve the natural environment to a reasonable degree" (Colo. Rev. Stat. § 37-92102(3)). My specific objective is to recommend flows sufficient to maintain key environmental processes and services indefinitely. Range of Variability (RVA) Several sets of authors have argued convincingly that when faced with little other information, 'drawing the line' for streamflow at 25% of the native or 'natural' flow is appropriate. Richter and others (1996) first proposed establishing flow targets by defining high and low flow 'pulses' at the 75th and 25th percentile 2 for all pre-impact daily flows, respectively. Their intent was to capture the full range of seasonal and interannual hydrologic variation to sustain native biodiversity for aquatic, riparian and wetland ecosystems. Then Richter and yet others (1997) expanded on that reasoning by detailing a method they called the "Range of Variability Approach". This method establishes initial flow management targets at the 25th to 75th percentile range around the median flows for a reference time period. The assumptions are (1) that prealteration streamflow patterns establish the context for and provide proper guidance to manage ecological systems today; and (2) that flow variability is a vital attribute of ecological systems (Landres and others, 1999). Since Richter and others (1997) proposal, more scientists have advocated approaches that focus on percentile thresholds (Whittaker and Shelby, 2000; Black and others, 2005; Arthington and others, 2006; Richter and others, 2006). The National Hydrologic Assessment Tool (HAT) recently developed by the U.S. Geological Survey (USGS) is an upgrade of the Indicators of Hydrologic Alteration (IHA) software originally put together for The Nature Conservancy by Richter and others (1996). Both IHA and HAT are designed to analyze flow regimes, identify the hydrologic alterations of human activities in a flow time series, and aid in setting environmental flow standards. They characterize and compare hydrologic regimes in terms that are traceable to direct ecological and geomorphological responses based on five fundamental attributes of any hydrologic regime. The HAT software users' manual (Henriksen and others, 2006) discusses the basis for using the natural flow paradigm, which relies on: … "the full range of natural intra- and interannual variation of hydrological regimes, and associated characteristics of magnitude, frequency, duration, timing, and rate of change, are critical in sustaining the full diversity and 2

Note that Richter and others (1996) and Henriksen and others (2006) use the term percentile rather than exceedence. The 25th percentile flow is essentially the same as that flow that is equaled or exceeded 75% of the time. Percentiles can be calculated on a variety of time scales; I use mean monthly flows in this paper.

2

integrity of aquatic ecosystems" (Poff and others, 1997; Power and others, 1995; Resh and others, 1988). Streamflow is strongly related to many critical physiochemical components of rivers, such as dissolved oxygen, channel geomorphology, and water temperature, and can be considered a "master variable" that limits the [distribution], abundance, and diversity of many aquatic plant and animal species (Resh and others, 1988; Poff and others, 1997). Though there is little doubt that at least some components of all five of the critical hydrologic attributes (magnitude, frequency, duration, timing, and rate of change) have been substantially altered on the Poudre River, I concentrate here primarily on flow magnitude because it is the most understandable – and most diagnostic – for the task at hand, but I do address the other components to varying degrees. Substrate Maintenance Flows It is well established that annual peak flows dominate many river processes, especially channel morphology and riverbed community health. High pulse flows shape the physical character of the river channel including pool and riffle distribution, bank structure, and channel width; determine the size distribution of stream bed substrates (sand, gravel, cobble); prevent riparian vegetation from encroaching into the channel; tend to prevent the establishment of non-native invasive plants and animals; restore normal water quality conditions after prolonged low flows, flushing away waste products and pollutants and generally providing for the cycling of nutrients; maintain pore space that helps aerate eggs and remove metabolic waste in spawning gravels as well as a healthy macroinvertebrate community in the hyporheic zone below the streambed; scour silt along river margins that otherwise would become prime habitat for the invertebrate hosts of the whirling disease parasite; provide riparian wetlands for waterfowl and amphibians, often by raising near-stream groundwater levels; strongly influence distribution and abundance of large woody debris that creates habitat for many organisms; and provide important dispersal and reproductive triggers or cues to the aquatic community (Bunn and Arthington, 2002; R. Milhous, USGS retired, personal communication; Nilsson and Svedmark, 2002; Richter and others, 2003). It is also reasonable to expect that reductions in the frequency and magnitude of flushing flows enable more robust attached filamentous and green algal communities to persist through the warm summer period. Such an enhanced algal community would be expected to reduce dissolved oxygen in the river at certain times, potentially reducing trout survival under adverse conditions (Thurston and others, 1981). Since channel substrate maintenance flows have been evaluated and quantified on the Poudre River near Fort Collins (Milhous, 2007), I rely on that work here to fine tune the RVA high-flow thresholds. However, I note that the sediment flushing flows identified by Milhous may be insufficient to fully protect the long-term ecological values achieved through true high-velocity channel forming flows.

3

Hydraulic Model – Habitat Assessment Nelson (1987) performed a relatively intensive study on the Poudre River using the Physical Habitat Simulation component (PHABSIM) of the Instream Flow Incremental Methodology (Stalnaker and others, 1995). This method involves using detailed hydraulic measurements in conjunction with 'suitability criteria' that characterize the relative utility of river depth, water velocity, and substrate/cover for a variety of aquatic objectives such as fish habitat or recreation. Nelson developed his model for two locations along the Poudre River, one at Martinez Park upstream from College Avenue representing a more narrowly confined channel, and one near Riverbend Ponds representing wider morphology. One of Nelson's products was a set of graphs depicting the amount of 'microhabitat' available at different flow levels for several lifestages of cold and warm water fish as well as several recreational activities. I draw selectively from his figures to double-check the RVA low-flow analysis. Water Temperature I previously studied the summer thermal regime in the Poudre River from the mouth of the canyon to Interstate-25 (Bartholow, 1991). The river's thermal regime is discontinuous due to the infusion of relatively cold water intermittently released from Horsetooth Reservoir. When Horsetooth is releasing to the Poudre River and/or on days with relatively cool meteorological characteristics, the thermal regime is sufficient to support a sustainable (if hatchery supplemented) population of rainbow and brown trout 3 (Oncorhynchus mykiss and Salmo trutta) and the food base on which these species exist through most of the study area. However, when Horsetooth releases are low and/or on otherwise 'hot' days, the thermal regime can be highly stressful to these fishes in terms of growth and condition. In this analysis, I rely on my conclusions from that 1991 study to further refine the RVA low-flow standards.

Methodology Range of Variability Data for the RVA analysis were available from two sources. First, estimates of native daily flows were available from the City of Fort Collins water utility (Donnie Dustin, personal communication), obtained by taking the USGS gage data at the mouth of the canyon (# 06752000) and adding back all upstream diversions while subtracting all outof-basin imports. As supplied, these data covered a 20-year period, water years (October through September) 1976 to 1995. I noted some problems with these data, usually small negative flows for a few days in several years, that I adjusted by averaging the estimates on either side. The second source was simply the daily USGS gage record for the Lincoln Street gage (#06752260) in downtown Fort Collins, Colorado, downloaded from the Internet and covering the same water years. [Note that 3

Though rainbow and brown trout are not natives of Colorado, they are used here as surrogates for the native trout they have replaced, both ecologically and economically.

4

representative photos of a variety of flow levels taken just upstream from the Lincoln St. gage may be found in the appendix.] I used the National HAT software (Version 3.0) to import, process and display the two daily records. Though the HAT software has many useful functions, the only capability I relied on for this analysis was to summarize the daily flow record for each of the two stations, the estimated native flow at the canyon mouth and the Lincoln Street gauged flow. The software calculates the each year's mean monthly flow for the 20-year record and computes the median as well as the 25th and 75th percentile values for each month. Comparing these two data sets side by side assumes that any natural losses or accretions between the mouth of the canyon and the Lincoln Street gage are minimal. Applying the HAT software specifically for the required monthly flow magnitudes resulted in Figure 1 and Table 1, adapted from the software's output solely for ease of interpretation here. Referring first to the hydrograph shown in Figure 1, the contrast is stark between estimated native flows and recent measured flows at the Lincoln Street gage. The magnitude of each month's flow range has been diminished relative to the estimated native flow, in many cases quite dramatically, especially the highest flow months of May through August. Generally speaking, the upper bound (75th percentile) of the recent measured flows at the Lincoln Street gage is less than (or barely equal to) the 25th percentile native flow range.

Figure 1. Native and recent mean monthly flows representative of the Lincoln Street USGS gage site captured directly from HAT software output but adapted for ease of interpretation. Colored bars depict the 25-75th percentile range of monthly flows while the black dots indicate the median for the 20-year daily flow record covering water years 1973-1995.

5

Table 1. Monthly flow summary depicted in Figure 1. Note that the 25th percentile flows for the estimated native flow regime is the starting point for the flow recommendation for the Poudre River through Fort Collins. Estimated Native Flow (cfs)

Lincoln St. Gage Flow (cfs)

Month 25th %-ile Median 75th %-ile 25th %-ile Median 75th %-ile Jan 48 64 89 4 12 30 Feb 57 65 75 4 8 32 Mar 64 96 120 3 5 42 Apr 131 198 334 4 10 140 May 675 1086 1328 90 172 337 Jun 1125 1974 2571 275 742 1107 Jul 428 673 1077 86 110 284 Aug 166 228 395 31 45 59 Sep 94 132 176 13 20 34 Oct 87 104 135 5 8 32 Nov 71 92 112 5 8 33 Dec 56 74 94 4 6 26 As mentioned, Table 1 presents the same flow values depicted in Figure 1. According to Richter and others (1997), the 25th percentile native flow is recommended as a default lower flow target, particularly in the absence of supporting ecological information. Such a target avoids the most extreme low flows and accomodates water use in an altered flow regime, yet is assumed to maintain essentially the same ecological function that the unimpacted flow regime maintained. To estimate a revised 25-50-75th percentile flow regime essentially following Richter's guidance yet giving a nod to the 'working river' paradigm, I constructed a monthly flow range that maintains the monthly median at the 25th percentile of the estimated native flows, but scaled the accompanying 25-75th percentile values around the new median proportionatly with the unimpacted flow pattern. Taking May for example, 675 cfs is about 62% of 1086 cfs and 1328 cfs is about 122% of 1086 cfs; these percentages multiplied times the May 25th percentile flow (675 cfs) resulted in the scaled values 420 cfs and 825 cfs that become the new 25th and 75th percentile flow targets. Applying this same method for all months resulted in the initial monthly flow recommendations given in Table 2.

6

Table 2. Initial monthly flow regime (cfs) developed solely using the established RVA standards (Richter and others, 1997). The median flows are the same as the 25th percentile of the estimated native flows given in Table 1. 25th 75th Month Percentile Median Percentile Jan 36 48 67 Feb 50 57 66 Mar 43 64 80 Apr 87 131 221 May 420 675 825 Jun 641 1125 1465 Jul 272 428 685 Aug 121 166 288 Sep 67 94 125 Oct 73 87 113 Nov 55 71 86 Dec 42 56 71

Substrate Maintenance Flow Requirements Using data primarily from USGS gage 06752280 above Boxelder Creek and in-channel substrate measurements, Milhous (2007) estimated that the flow necessary to adequately scour and flush sediment from the Poudre River's bed between Ft. Collins and Greeley is approximately 2050 cfs. Milhous (2007) did not specify exactly how often (how many days or how many years) flushing should occur. Milhous reports that from 32 recent years of record (1975–2006) at the closest USGS gage (Boxelder Creek), flows of this magnitude occurred during 12 years, but there was also a 7-year period during the recent drought (2000–2006) when no flushing would have occurred. I also note, in a related vein, that peak flows on the Poudre River seem to have been decreasing through time as shown in Figure 2.

7

Poudre River Peak Flows @ Mouth of Canyon Source: USGS Gage 6752000 (1891 dam failure omitted)

12000

Annual Peak Flow (cfs)

10000 8000 6000 4000 2000 0 1880

1900

1920

1940

1960

1980

2000

Year

Figure 2. Decline in Poudre River peak flows at the mouth of the canyon gage (6752000) omitting the 1891 dam failure. The dashed trend line is the 90% quantile and the solid line is the standard linear regression, both indicating the decline in snowmeltdriven peak flow events through time. (Acknowledgements to Bob Milhous, USGS.) A minimum number of days of peak channel maintenance flows must be provided in high water years to maintain channel integrity, dislodge more established vegetation, and prevent long-term sediment aggradation. By contrast, average to low water years can generally sustain some diversion of peak flows. In the absence of more specific guidance from Milhous (2007), I rely on Richter and others (2003) who recommend certain peak flow-specific metrics for channel flushing. Hydraulic Modeling - Habitat Assessment A selection of Nelson's (1987) results is presented in Figure 3, where the Y-axis represents the amount of 'suitable habitat' for lifestages of two fish and recreational tubing. I chose brown trout spawning and carp (Cyprinus carpio) fry as representatives for what I perceive as the limiting lifestages for cold water and unwanted exotic fish, respectively (other than water temperature). Though carp fry appear in the spring or summer and brown trout spawn in the fall but emerge during spring high flows, both have specific flow needs. [Not shown in Figure 3, the habitat vs. flow relationship for rainbow trout spawning in the spring is essentially identical to the brown trout fall

8

Weighted Usable Area (sq ft / ft)

spawning curve.] I chose recreational tubing because it well represents all of the recreational activities Nelson modeled, including rafting and canoeing, in the sense that the functional relationship with discharge is virtually the same.

50 40 30 20 10 0 0

100

200

300

400

500

Discharge (cfs) Brown Trout Spawning

Carp Fry

Tubing

Figure 3. Selected results from Nelson (1987) representing the Lee Martinez study site. Values shown here were approximated from his published graphs representing the amount of suitable 'habitat' for the species lifestage or activity. It is immediately apparent that habitat availability for brown trout spawning and carp fry differ markedly in response to both low and high flows. Slow and shallow flows below 50 cfs are best for carp fry; any reduction in peak spring or summer flows will tend to benefit carp in the Poudre River. In contrast, deeper and faster flows above 100 cfs are best for brown trout in their fall spawning activity. Recreational tubing follows an almost uniformly linear relationship, increasing from zero flow. Water Temperature Water temperature is critical in controlling the distribution and abundance of cold water fish by triggering movement and spawning behavior, mediating growth and survival rates, influencing competitive interactions, and strongly influencing other water quality attributes. For water temperature information, I relied on the conclusions of my previous study (Bartholow, 1991). At that time, I demonstrated that a combination of non-flow

9

and flow management alternatives could maintain suitable water temperatures for a sustainable trout fishery throughout most of the study area. Though some non-flow alternatives, such as reducing the channel width, may prove beneficial (if they can be shown to be sustainably cost-effective), my central conclusion was that summer supplemental flows of about 105 cfs would be required on an average of 30 days a year to maintain a suitable thermal regime, keeping water temperatures below about 23.3°C (74°F), a value chosen to represent the approximate daily maximum temperature limit which, if exceeded, decidedly lowers the probability of a healthy fishery 4 . The average 'base' flow at the Lincoln Street gage during July and August of the year I examined was about 43 cfs, meaning that flows would need to be about 148 cfs on the hottest days of the summer. It is important to note that my conclusion was based on supplemental releases into the Poudre River from Horsetooth Reservoir. Releases from the thenproposed mainstem Grey Mountain dam and reservoir were less effective at maintaining a suitable thermal regime simply due to the added upstream distance. It is also important to note that my objective here is not to provide cold water all the way to Greeley. Though there is little doubt that the Poudre River once supported a thriving trout fishery throughout its entire length, at least in many years (Bartholow, 1991), the Poudre River also supports a unique thermally transitional fauna as it flows into the plains.

Combining the Results I began developing the flow recommendation by using the Range of Variability Approach's identified floor, i.e., the 25th percentile flow for each month using reconstructed native flows supplied by the HAT software (leftmost column of flows in Table 1). Richter and others (1997) recommend that the mean of the altered monthly flows remain in the 25-75th percentile range of unimpacted flows. Therefore I used the monthly flow ranges given in Table 2 as the low flow target. So, how can the RVA flow recommendation be improved by bringing in information from the other three techniques that have been developed explicitly for the Poudre River? Looking first at substrate maintenance flows, Milhous (2007) estimated that the flow needed to scour fines from the Poudre River's bed was 2050 cfs. This flow is slightly above June's estimated median native flow of 1974 cfs but well below the 75th percentile native flow for the same month. Because it is highly likely that a maximum daily flow of 2050 cfs would accompany a median monthly flow of 1974 cfs (though this should be investigated), it would seem that flushing once took place at least every other year on average. This is consistent with the common assumption that bankfull flows that occur every 1.5–2 years are "effective" in maintaining a proper sediment balance and maintaining the channel (Andrews, 1980). To me this means that having median monthly June flows range only up to 1465 cfs may not be sufficient to sustainably maintain the aquatic system. An allowance must be made to guarantee delivery of 4

It is reasonably forseeable that the Colorado daily maximum water temperature standard along the Front Range, including portions of the Poudre River just above Ft. Collins, will be adjusted to 23.8°C. See http://www.cdphe.state.co.us/regulations/wqccregs/wqccreg31basicstandardsforsurfacewater.pdf

10

flows around 2000 cfs for channel flushing. But for how long should these substrate maintenance flows last? Though they were writing about another basin, Richter and others (2003) further defined their stance on just that issue by more clearly stating that one should exceed the minimum annual 1-day maxima in all years, exceed the 25th percentile of the 1-day maxima in 3 out of 4 years, and exceed the median of the 1-day maxima in half of the years. Though the HAT software has no equivalent metrics, making these calculations was straightforward. Specifically, the minimum annual 1-day maxima to be exceeded in all years is 1396 cfs; the 25th percentile of the 1-day maxima to be exceeded in 3 out of 4 years is 2405 cfs, and the median of the 1-day maxima to be exceeded in half of the years is 3018 cfs. Further, among other metrics, Richter et al. (2003) suggest annual high flow durations exceed the 25th percentile in 3 out of 4 years, but they offer no guidance on exactly how to define the duration period. I calculated the 25th percentile for the duration (number of days/year) for flows greater than several percentages of the peak daily 'natural' flow in each of the 20 years I examined. The results showed that if the high flow period were defined as the number of days flows exceeded 90% of the peak one-day event, the flush should be 1.75 days long. If on the other hand, if the definition included days with flows greater than 75% of the peak, the flush should be almost 6 days long. Interestingly, there was no significant relation between each year's peak flow and the durations of high flows in that year, regardless of the percentage chosen. This means, I believe, that a separate duration need not be specified for wet, dry, or average years. In an attempt to strike a middle ground between Milhous (2007) and Richter and others (2003), and in the absence of more refined physical modeling, I estimate that requiring (a) an annual 1-day maximum of 1400 cfs in all years and (b) a 2-day maximum of 2000 cfs in 3 out of 4 years would adequately achieve substrate maintenance objectives. I note that flow duration recommendations for substrate maintenance have generally ranged from 2 to 7 days, but have been as long as 14 days (Tennant, 1976, as cited in Whiting, 2002. Are there any adjustments to the recommended flows that appear necessary after looking at Nelson's (1987) hydraulic/habitat modeling results? Nelson's results clearly show that spring and fall flows below about 50 cfs will tend to benefit unwanted exotic fish (such as the common carp and other species) to the detriment of cool water and native species. Because the 25th percentile monthly native flows are all above this level (except for a slight discrepancy for January), the 25th percentile floor seems reasonable, but I did raise the low flow recommendation for January, March and December to be no lower than 50 cfs to benefit native and cool water species. These monthly values cannot be improved much using Nelson's recreational tubing results except to say that the more water the better, at least up to some safety level beyond the scope of this paper. Finally, does the water temperature analysis help further refine the flow recommendation? The 25th percentile native flows for July and August are 428 and 166 cfs respectively, exceeding my earlier estimate of 148 cfs on at least the hottest days. Therefore having mean monthly flows at or near the lower end of that range, 160 cfs,

11

seems appropriate. However, summer daily flows must be ≥150 cfs on the hottest days with coldwater releases from Horsetooth Reservoir to make up any difference if we want to sustain any trout in the reach. To varying degrees, I have touched on most of the critical elements defining a flow regime – magnitude, timing, frequency, and duration – but not rate of change. In fact, dealing primarily with mean (or median) monthly flows completely overlooks some potentially critical inter-day phenomena so common on the Poudre River. Figure 4 is but one example of how markedly daily flows can vary depending on starting and stopping irrigation diversions and Horsetooth Reservoir releases.

Figure 4. Example of highly variable Poudre River flows recorded in 2005 at the Lincoln Street gage. Note that flows rose from about 30 cfs to 500 cfs on October 26, eventually rising to over 1100 cfs by October 31, and then dropped to 1 cfs over the next 2-3 days. Such erratic and high volume flow variations are considered quite detrimental to the aquatic system if outside their 'natural' occurrence window (likely disrupting fall spawning in this case), reducing species richness and standing crop, interfering with seedling establishment, and causing fish and invertebrate stranding (Bunn and Arthington, 2002; Nilsson and Svedmark, 2002). Down-ramping (abrupt termination of

12

high discharges) may also contribute to bank slumping -- leading to increased fine sediment accumulation in the channel. Very rapid rises in flow can also be a public safety hazard. Using 30-minute flow values from the USGS and Colorado Department of Water Resources instantaneous data archives supplied by the City of Fort Collins (Donnie Dustin, personal communication), I calculated the absolute and percent variation in daily [(maximum minus minimum)/maximum] flows at both the canyon mouth and Lincoln Street gages for March through October of water year 2007. Variations were large and quite erratic through the year for both stations using both metrics, especially during peak runoff and after August. Mean daily variation was 24% at the mouth of the canyon and 49% at Lincoln Street, though the range and standard deviation were large at both stations. Though I cannot distinguish natural diurnal variation and manmade changes at the canyon gage using these data, I assume that the canyon variation better reflects ecologically relevant conditions. Following the RVA philosophy, I calculated 75th percentile of the relative variation in daily canyon mouth flow variation -- 31% for the limited data set available. However, this begs the question of about the number of days in a row that a 30% flow decline could be tolerated and answers nothing about very short-term deviations. For this reason, using judgment, I suggest the following criterion: No managed changes in streamflow greater than 30% from day to day, none greater than 15% in a six-hour period, and none greater than 40% over any continuous 7-day period at any time of year.

Conclusions The final recommended monthly flow ranges are given in Table 3. The 25th percentile level would be representative of low flow years, the median (50th percentile) level would be representative of average flow years, and the 75th percentile level would be representative of high flow years. Table 3. Monthly Flow Range Recommendation (cfs) 75th 25th Month Percentile Median Percentile Additional Guidance Jan 50 >50 67 Feb 50 57 66 Mar 50 64 80 Apr 87 131 221 May† 420 675 825 †Annual 1-day maximum ≥ 1400 cfs in all years; † Jun 641 1125 1465 3 out of 4 year 2-day maximum ≥ 2000 cfs. Jul 272 428 685 * Aug 121 166 288 *Flows ≥150 cfs required on 'hot' days. Sep 67 94 125 No managed changes in streamflow greater than Oct 73 87 113 30% from day to day, none greater than 15% in Nov 55 71 86 a six-hour period, and none greater than 40% over Dec 50 56 71 any continuous 7-day period at any time of year.

13

To put this flow recommendation in perspective, Figure 5 compares three monthly flow regimes for the Poudre River: historical (native) flows, recent (current) flows, and the recommended flow regime developed here. As can be seen, the modest level of protection envisioned by this flow recommendation will require more water each month than has been in the Poudre River in the recent past.

Figure 5. Comparison of estimated historical flows, recently measured flows at the Lincoln Street USGS gage site in downtown Fort Collins, and the recommended flow regime developed in this document (left to right respectively). Colored bars delimit the 25-75th percentile range of monthly flows while the black dots indicate the median for each data set. Discussion As a rule, the full range of seasonal and inter-annual hydrologic variation is necessary to completely maintain native biodiversity for aquatic, riparian and near-stream wetland ecosystems. As we have simplified the Poudre and Platte River's governing flow regimes, we have pruned away that diversity and encouraged unwanted exotic species instead of natives. If we simplify further, we will lose more diversity. Yet we do not know, and likely will never know, exactly how much seasonal and inter-annual hydrologic variation must be maintained to achieve an ecologically sustainable aquatic system for the Poudre River through Fort Collins. As a first step, however, I have applied the Range of Variability technique defining the 25th percentile of the monthly native flows as the foundation for an instream flow recommendation. Additional Poudre River-specific criteria were used to fine-tune this flow template in order to adequately address substrate maintenance flow requirements and set minimum flows to buffer cool water fish from domination by exotics. Note that using 25th percentile flows to define the bottom of the recommended flow range still allows low flow excursions in exceptionally water-short years.

14

Against this backdrop, the recent Poudre River flow regime is far from adequate to fully protect stream ecosystems and potential recreational uses (Figures 1 and 5). The median flow for all months, especially May, June and July, fall well outside their target range. Additional diversion of water upstream of Fort Collins is therefore likely to further jeopardize the river's health in the reach from the canyon mouth to Interstate-25 except under exceptionally high June flow conditions. But why should the native flow regime set the benchmark? After all, the Corps of Engineers will presumably assume that the status quo (or alternately the future without the project) of this already impacted river is the 'baseline' against which they will judge the permitability of any proposed project. I believe the answer must be that this logic has no foreseeable happy ending. That is, if project A is approved with the current baseline, then project B could be approved with a new reduced baseline, and so on until there is zero water in the river. I recognize that this is not literally likely for a variety of reasons, but the unfortunate logical conclusion based on the information presented here is that the river is currently so depleted that no further reductions should be allowed. Anyone proposing otherwise must present a more defensible case than the one outlined here. Though some aspects of the natural environment of the Poudre River have been extensively studied, to the best of my knowledge no one has tried to catalog the full range of impacts that have already occurred on the river. A reasonable template for environmental effects on the Poudre River, however, may be gleaned from what is already well known on the South Platte River. Strange and others (1999) summarized what is understood about the South Platte River and reported that: (1) alterations to riparian habitat have been so extensive that at least four species of birds have been lost through hybridization with non-natives; (2) the S. Platte had the highest contamination of ammonia and nitrate and the second highest level of phosphorus among 20 major rivers sampled by USGS, all leading to water quality violations; (3) loss of native riparian vegetation was associated with increased algal abundance and loss of benthic macroinvertebrates; and (4) modification of the flow regime has resulted in declines of six fish species native to the basin and under consideration for T&E listing, and establishment of 18 exotics. Dennehey and others (1998) pointedly amplify on the water quality and stream channel conditions in the South Platte basin, specifically that (1) subsurface irrigation return flow is a major nonpoint source of nitrate, dissolved solids, and pesticides (atrazine and prometon) in the lower reaches of the South Platte River; (2) wastewater treatment plant effluent also contributes massive loads of phosphorous, nitrate and ammonia; (3) large diversions from streams and rivers in the basin result in less water to dilute these and other measured contaminants to the point of violating EPA's aquatic life criteria; (4) alteration of the natural flow regime has seriously degraded native aquatic habitat along streams; and (5) surface and ground water reuse for irrigation has resulted in increased salinity in the lower South Platte River and surrounding alluvial aquifer to the point where the salinity is detrimental to both irrigation and drinking-water supplies. Collectively, using multiple lines of evidence, Dennehey and others (1998) conclude that large portions of the S. Platte

15

basin, including some of the lower portions of the Poudre River are moderately to significantly degraded. The City of Greeley has already had to do some dredging in the Poudre River because sediment entering the river or contributed by bank erosion is no longer flushed out of the system due to lack of peak flows. This sediment aggradation (build up) has reduced channel conveyance leading to an increased risk of flooding. Further, the reconnaissance study for this section of the river (USACOE, 2004) shows that the important riparian zones and stream corridor need costly restoration; that wildlife migration has been disrupted; that narrowing of the channel has removed connections to historically adjacent wetlands, oxbows, side channels, etc.; and native plant, fish, and wildlife species have been lost while non-native species such as Russian olive and tamarisk have invaded. Known sediment problems are not confined to Greeley. After the 1997 flood on the Poudre River, Bob Shields wrote an editorial in a Fort Collins newspaper (Shields, 1997) explaining how water withdrawals from the river and other land use practices had resulted in considerable aggradation of the river. He specifically mentioned that he was once able to drive his tractor under the I-25 bridge over the Poudre River when he owned that land in 1965. Now, according to Shields, during high water events the river is so plugged at this location that it is very close to overtopping the interstate highway. That was in 1997 and the situation may have deteriorated since that time. The data and results presented here could certainly be improved. I used 20 years of daily flow data based on the reconstructed native flow data available to me; a longer record would likely make flow recommendations more accurate – at least based on the climatic period the data represented. Looking back at the flow record that has been reconstructed for the Poudre River by Woodhouse and others (2004), the 20-year block beginning in 1976 represents about the 83rd percentile of all other non-overlapping 20year blocks back to the year 1626. However, the blocked data have little variance overall -- the 50th percentile flow for the same data set is only about 10% lower than the mean 1976-1995 flow. In short, the time period I examined is only slightly wetter than the long-term average. There are, in addition, many caveats to this flow recommendation: (a) The study by Milhous (2007) had a relatively wide margin of uncertainty surrounding his substrate maintenance flow estimate. More comprehensive studies would be required to reduce that uncertainty and evaluate the degree to which exceptionally high flows might or might not negatively influence trout population dynamics over multi-year periods. (b) The fish habitat modeling techniques used by Nelson (1987) have evolved considerably since that time. Though it is unlikely that his results would change substantially, another modeling study would need to be done to confirm or adjust his results. (c) Colorado has recently adopted a revised water temperature standard, at least in some basins. I do not know exactly how new standards might be applied on the Poudre River through Fort Collins, but it is always possible that my results (Bartholow 1991) might need to be modified to accommodate the new state standards. (d) Ramping rates typically are derived for specific rivers based on their unique channel morphology. A ramping rate

16

study would need to be done to refine the daily rate of change recommendations I have provided here, which are somewhat greater than others have recommended on other rivers (Whiting, 2002). (e) Richter and others (1997) and Henriksen and others (2006) recommend other metrics that have proved useful in characterizing robust flow regimes to preserve the natural environment to a reasonable degree. These other metrics could be added to what has been assembled here. (f) Finally, the flow regime recommended here for fine sediment flushing and maintenance of the riverbed surface is likely insufficient in magnitude and duration to avoid the ongoing trend of willow encroachment and loss of channel capacity (Brian Bledsoe, Colorado State University, personal communication, 2008). Through field monitoring and preliminary modeling, Dr. Bledsoe has estimated that peak flows exceeding 3000 cfs, lasting 3 to 4 days, and occurring on average every 3 or 4 years are likely necessary to maintain channel capacity and rejuvenate aquatic habitats, potentially including jurisdictional wetlands. Therefore, broadly speaking, the flow recommendations I offer here must be viewed solely as providing preliminary estimates until the critical relationships are better quantified with existing models, particularly for channel maintenance and riparian vegetation. Then, of course, active monitoring would be necessary to ensure that objectives have been achieved and identify any appropriate adjustments. One element of this flow recommendation potentially divides advocates for flow restoration. Exceptionally high flows at certain times of the year have been associated with temporarily reducing trout standing crop, presumably due to flushing out juvenile life stages (Nehring and Anderson, 1993). However, Barry Nehring has also shown that you do not need high brown or rainbow trout recruitment every year to maintain a thriving trout fishery, especially one managed as a catch-and-release (or a low bag limit) fishery. In fact, lower recruitment by flushing some of the 'surplus' may beneficially avoid density-dependent declines in trout growth (stunting). Though I do not know whether trout recruitment might be limiting in the Fort Collins reach due to poor gravel quality or other factors, I have been advised that loss of eggs by dewatering is likely a greater threat than flushing fry (Kurt Fausch, Colorado State University, personal communication, 2008), so I stand by the recommendation for high annual substrate maintenance flows during the snowmelt runoff period. After all, it is reasonable to assume that the aquatic community is well adapted to the Poudre River's annual snowmelt pulse. Having a sound and reasonable flow recommendation of course begs the critical question of who would have the responsibility, resources, and authority to oversee monitoring and make ongoing decisions to achieve partial restoration of the Cache la Poudre River. Administration of this sort is well beyond the scope of this paper but has been a recent public topic (Neil Grigg, Colorado State University, personal communication, 2008) and would need to be carefully thought out. I recognize that complete restoration of the 'working' Poudre River is not a likely option in this human-dominated watershed, especially in the face of water extraction projects such as those proposed for the River and the almost fully allocated set of existing diversion rights. It is also important to note that partially restoring the flow regime alone

17

is not likely to be sufficient to protect and enhance the ecological integrity of the river. Other water quality factors, especially nutrients and toxics, must be considered. Obsolete diversion structures should be removed, and active diversion structures must be re-engineered when replaced to improve recreational safety (Donahue and Earles, no date; Wright and others, 2004) and permit at least minimal upstream migration for aquatic species. Some provision for retaining large woody debris, currently removed from the river for bridge safety reasons, must be made. Nevertheless, having a reasonable instream flow targets is critical in furthering the goal of partially restoring the ecological integrity of the Poudre River.

Acknowledgements Many individuals have provided useful insights, challenging comments and continued encouragement for this effort. They include Brian Bledsoe, Kurt Fausch, Neil Grigg, Jim Henriksen, Bob Milhous, Barry Noon, LeRoy Poff, and John Sanderson. Not all endorse every aspect of this paper, but many endorse the general approach given the distinct lack of available empirical data for the Poudre River through Fort Collins.

References Andrews, E. D. 1980. Effective and bankfull discharges of streams in the Yampa River Basin, Colorado and Wyoming. Journal of Hydrology 46: 311-330. Annear, T., Chisholm, I., Beecher, H., Locke, A., and 12 other authors. 2004. Instream flows for riverine resources, revised edition, Cheyenne, Wyo., 268 p. Arthington, A.H., S.E. Bunn, N. L. Poff, and R.J. Naiman. 2006. The Challenge of Providing Environmental Flow Rules to Sustain River Ecosystems. Ecological Applications, 16(4):1311–1318. Bartholow, J.M. 1991. A modeling assessment of the thermal regime for an urban sport fishery. Environmental Management. 15(6):833-845. Black A.R., J.S. Rowan, R.W. Duck, O.M. Bragg and B.E. Clelland. 2005. DHRAM: a method for classifying river flow regime alterations for the EC Water Framework Directive. Aquatic Conserv: Mar. Freshw. Ecosyst. 15: 427–446. Bunn, S.E., and Arthington, A.H. 2002. Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity: Environmental Management, 30(4):492−507. Dennehy, K.F., Litke, D.W.,Tate, C.M., Qi, S.L., McMahon, P.B., Bruce, B.W., Kimbrough, R.A., and Heiny, J.S. 1998. Water quality in the South Platte River Basin, Colorado, Nebraska, and Wyoming: U.S. Geological Survey Circular 1167, 38 p. Available on the Internet at http://pubs.usgs.gov/circ/circ1167/circ1167.pdf, accessed 8/4/2008. Donahue, M.G. and T.A. Earles. No Date. Recreational Use Considerations in Planning and Permitting of Low Head Dams. Available on the Internet at http://www.stark-stark.com/attorney-lawyer-1017386.html, accessed 3/30/2008.

18

Henriksen, J.A, J. Heasley, J.G. Kennen, and S. Nieswand. 2006. Users' manual for the Hydroecological Integrity Assessment Process software (including the New Jersey Assessment Tools). Open-File Report 2006-1093. Fort Collins, CO: U.S. Geological Survey, Fort Collins Science Center. 71 p. Landres, P.B., P. Morgan, and F.J. Swanson. 1999. Overview of the use of natural variability concepts in managing ecological systems. Ecological Applications 9(4): 1179–1188. Milhous, R. T. 2007. An Adaptive assessment of the flushing flow needs of the lower Poudre River, Colorado: First evaluation. Paper presented at the Annual Rocky Mountain Hydrologic Research Center Conference on 28 September 2007 at Wild Basin Lodge, Allenspark, Colorado. 13 pp Nehring, R.B., and R.M. Anderson. 1993. Determination of population-limiting critical salmonid habitats in Colorado streams using the Physical Habitat Simulation system. Rivers 4: 1-19. Nelson, P.C. 1987. Physical microhabitat versus streamflow relationships in the Cache la Poudre River, Fort Collins, Colorado. Poudre River Corridor Fishery Plan, Phase I Final Report. 8 pp. plus figures. Nilsson, C., and Svedmark, M. 2002. Basic principles and ecological consequences of changing water regimes: riparian plant communities. Environmental Management 30(4):468-480. Poff, N.L., Allan, J.D., Bain, M.B., Karr, J.R., Prestegaard, K.L., Richter, B.D., Sparks, R.E., and Stromberg, J.C. 1997. The natural flow regime − A paradigm for conservation and restoration of river ecosystems: BioScience 47: 769−784. Power, M.E., Sun, A., Parker, M., Dietrich, W.E., and Wootton, J.T. 1995. Hydraulic food-chain models − An approach to the study of food-web dynamics in large rivers: BioScience 45:159−167. Resh, V.H., Brown, A.V., Covich, A.P., Gurtz, M.E., Li, H.W., Minshall, G.W., Reice, S.R., Sheldon, A.L., Wallace, J.B., and Wissmar, R.C. 1988. The role of disturbance in stream ecology: Journal of North American Benthological Society 7: 433–455. Richter, B.D., Baumgartner, J.V., Powell, J., and Braun, D.P. 1996. A method for assessing hydrologic alteration within ecosystems: Conservation Biology 10:1163−1174. Richter, B.D., Baumgartner, J.V., Wigington, R., and Braun, D.P. 1997. How much water does a river need: Freshwater Biology 37:231−249. Richter B.D., R. Mathews, D.L. Harrison, and R. Wigington. 2003. Ecologically sustainable water management: managing river flows for ecological integrity. Ecological Applications 13(1):206–224 Richter, B.D. A.T. Warner, J.L. Meyer and K. Lutz. 2006. A Collaborative and Adaptive Process for Developing Environmental Flow Recommendations. River Res. Applic. 22: 297–318. Shields, B. 1997. Managing flood plain needs more than maps. Fort Collins Coloradoan, 3/16/97, page E3. Stalnaker, C., B.L. Lamb, J. Henriksen, K. Bovee, and J. Bartholow. 1995. The Instream Flow Incremental Methodology. A primer for IFIM. U.S. National Biological Service Biological Science Report 29. 44 pp.

19

Strange, E. M., K. D. Fausch, and A. P. Covich. 1999. Sustaining ecosystem services in human-dominated watersheds: biohydrology and ecosystem processes in the South Platte River basin. Environmental Management 24(1):39-54. Thurston, R.V., G.R. Phillips, R.C. Russo, and S.M. Hinkins. 1981. Increased toxicity of ammonia to rainbow trout (Salmo gairdneri) resulting from reduced concentrations of dissolved oxygen. Can. J. Fish. Aquat. Sci. 38:983-988. U.S. Army Corps of Engineers. 2004. Reconnaissance Study: Section 905(b) (WRDA 86) Preliminary Analysis, Cache la Poudre River - Greeley, Colorado, Flood Damage Reduction and Environmental Restoration Study. Omaha District. 26 pp plus appendices. Whiting, P.J. 2002. Streamflow necessary for environmental maintenance. Annual Review of Earth and Planetary Sciences 30:181–206. Whittaker, D., and B. Shelby. 2000. Managed flow regimes and resource values: traditional versus alternative strategies. Rivers 7(3):233-244. Woodhouse, C.A., et al. 2004. TreeFlow Colorado Streamflow Reconstructions. IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series # 2004029. NOAA/NGDC Paleoclimatology Program, Boulder CO, USA. Wright, K.R., T.A. Earles, and J.M. Kelly. 2004. Public safety at low-head dams. Available on the Internet at http://www.wrightwater.com/wwe/wwepubs/pdfs/LowHead%20Dams%202-2004.pdf and http://www.wrightwater.com/wwe/wwepubs/pdfs/Low-Head%20Dam%20Figures.pdf, accessed 3/30/2008.

20

Photo Appendix All photos taken at the College Ave. boat chute and date/time correlated to the Lincoln St. USGS gage.

14 cfs

31 cfs 341 cfs

376 cfs - with sediment flush

433 cfs 593 cfs

21

634 cfs

904 cfs

1540 cfs

22