End-of-Season Research Update

This presentation was given on Wednesday, September 23rd at the Guides and Outfitters End-of-Season Gathering at Pond’s Lodge. HFF’s research and restoration team presented a large amount of data, and for those who were unable to make it, who want to dig in to the graphs a little more, or are just generally interested in what the Foundation has been working on in terms of analyzing water quantity and water quality, here is a link to the slides, in pdf format. An explanation of each slide is given below.



  1. Rob Van Kirk: This summer, we have really listened to concerns from guides, anglers, and our members regarding high flows out of Island Park Reservoir, low flows in Fall River, and poor water quality downstream of the Island Park dam. As a result, HFF has really focused its efforts this past summer on contextualizing how flows in the watershed compare to those of the past as well as expanding and analyzing data from our water quality monitoring network.
  2. Rob: Today we are going to present on three major topics – water quantity, water quality, and fisheries management issues in different sections of the river.


Water Quantity:

  1. Rob: Christina Morrisett is going to talk to us about just how dry 2015 was. Christina is a Stanford graduate who won a competitive internship position with the Foundation, and I convinced her to stay on and continue working with us. Christina Morrisett: Today I’m going to walk through a handful of graphs to show what happened in the Henry’s Fork watershed over the past year. We’ll start at the top of the watershed and then work our way down, I’ll discuss how spring runoff can be used to predict total flow for the year, and then I’ll finish off by introducing what discharge out of Island Park Reservoir looked like this summer and then hand it off to Rob to discuss how flow out of the reservoir was managed.
  2. Christina: This first graph is a statistical summary graph showing how much water was gained, i.e., how much natural inflow came into the reach, between Henry’s Lake and Ashton Dam using data from USGS over the last 80 years or so. In general, this graph shows low flow in the winter as water is tied up in snow and streamflow is fed by groundwater. There is higher flow in the summer due to rainfall and runoff from snowmelt. Looking at this year specifically, we see that natural streamflow was low all winter – sitting right at that 25th percentile, so what you’d expect see exceeded in 25% of years. In late March/early April, that blip, that small peak, that was our runoff event. That was it. Normally, runoff peaks a month later in mid-May. This year, with a near record-low snowpack, we did not receive the amount of water we usually do from the Centennial Mountains and Yellowstone Plateau. Instead, our runoff event at the beginning of April was primarily due to snowmelt at lower elevations on the caldera floor. In fact, the amount of water we received during the runoff period between April and June was the third driest in the last 80 years, behind 1934 and 1977. Since our small runoff event in April, we’ve fallen below the 25th percentile, and watershed inflow between Henry’s Lake and Ashton Dam have sat just above the driest years on record.
  3. Christina: Now we’ll move farther down the watershed into Fall River. These graphs are again, statistical summary graphs. The top graph shows natural flow and the bottom graph shows diversions, how much water was taken out of the system. Data was taken from Idaho Department of Water Resources from 1978-2015. Looking at the top graph, we see that natural flow in Fall River has a similar shape to the last graph. Winter streamflow was actually higher than the Henry’s Lake to Ashton Dam reach, sitting at the median, and again, you can see that first runoff peak in late March/early April. That peak was runoff from lower elevations and the second peak, at the beginning of May, was runoff from higher elevations – the Tetons and Yellowstone Plateau – as well as the rainfall we got in May. However, if you look at the median and even the 1st quartile lines, you’ll see that runoff doesn’t usually peak until the beginning of June. Runoff this year peaked at the beginning of May, reaffirming the fact that runoff this year happened about four weeks earlier than usual. But as soon as the rain stopped and all the snow melted at the beginning of June, natural flow in Fall River dropped significantly and has been sitting at near-record minimums ever since. So how did that translate in terms of irrigation diversions? With an earlier runoff, we had an earlier growing season. The first peak in diversions at the beginning of May was above average for that time of year, but average in terms of its timing within the growing season (remember, it started about four weeks earlier than usual). The dip we see between May and June is due to the rainfall we received, which decreased irrigation demand. And then once the rain stopped, diversions returned to normal. However, while diversion demand may have been average, the supply was not. As a result, demand ended up taking 100% of the supply, leaving Fall River to a trickle. This is routine in below-average water years, but the reason it stuck out to us this year is because it happened a month earlier than usual. With earlier planting there was an earlier harvest and as a result, irrigation demand has been decreasing since July.
  4. Christina: Now I’ll discuss how we can use spring runoff to predict total flow for the entire water year. This graph is a linear regression plot showing the relationship between mean daily flow during April-June on the x-axis and mean daily flow for the total water year on the y-axis. Each green dot represents an individual water year between 1934 and 2014. The green dots clustered around this line show that there is a very strong relationship between runoff the total flow. In fact, the correlation coefficient for the relationship between x and y was 0.92 – perfect linearity is 1. This graph shows us that mean spring-time runoff is a strong predictor of total flow for the whole water year. Points above the lines are years when total water-year flow was higher than expected based on spring runoff; points below the lines are years when it was lower than expected. Variability around the line is due to the fact that the Henry’s Fork is a groundwater fed system. The river’s baseflow is dependent on precipitation from previous years. If the previous few years were wet, there is more groundwater available for the following year’s winter flow, meaning that total water-year flow would be higher than predicted by spring runoff alone. On the other hand, if the previous water years were dry, there is less groundwater available, and total water-year flow is more likely to be lower than predicted by runoff.
  5.  Christina: So where did 2015 fall on the line? As I mentioned before, this water year had the third driest runoff season in the last 80 years. In terms of total flow, 2015 was the driest in the last 70 years. We have not seen a water year drier since 1941. In the last 80 years, 2015 ranked sixth. So this year was really one for the record books. 2015 fell very close to the regression line, meaning that 2015 had both a very low runoff and followed several below-average years; the prediction of low flows based on low runoff was consistent with low groundwater levels coming into the spring of 2015.
  6. Christina: This last graph is another statistical summary graph showing discharge out of Island Park Reservoir using data from USGS over the past 75 years. Here, we see how irrigation demand started a month earlier than usual. In an average year, reservoir storage is not used for irrigation delivery until July. However, with an earlier growing season came an earlier irrigation season and delivery began in early June. In fact, had we not received the amount of rainfall we did in May, delivery would have started even earlier. With an earlier growing season came an early harvest and discharge out of Island Park dam has been decreasing since the beginning of July – the spike you see around July 20th was the Chester turbine testing event. In terms of overall discharge recorded at the gage just below Island Park Dam between mid-June to mid-September, the difference between the median and WY 2015 show us that this year was actually below average – the rates and timing of discharge were just different than in years past. And now I’ll pass it onto Rob who will talk about how this discharge was managed.
  7. Rob: As Christina pointed out, the highest irrigation delivery from Island Park Reservoir occurred from mid-June to mid-July, 3-4 weeks earlier than normal. You can also see the spike at around 2000 cfs July 20-22 for the turbine test at Chester Dam.
  8. Rob: When compared with the mean daily flow over the past 82 years, you can see that flow was below average early in June because runoff was so low and then above average from mid-June through mid-July due to the early irrigation demand. However, after the Chester turbine test, flow dropped below average and has been there ever since. The below-average late summer flows are the result of the very early irrigation demand. Most crops were planted three to four weeks earlier than usual, so by late July, demand had dropped to well below average, once grain had ripened and the second cutting of alfalfa was done.
  9. Rob: I plotted daily flow from 2001 on this graph to show a fairly recent dry year in comparison to 2015. Demand also started early in 2001, and in fact, flows during late June and early July in 2001 were nearly identical to the flows we experienced this year. So, the high irrigation delivery we experienced this year in June, while unusual, was not unprecedented. However, you can see that high delivery of water from Island Park Reservoir continued throughout the rest of the summer in 2001, unlike this year, when flows dropped dramatically at the end of July.
  10. Rob: Finally, I put a recent wet year on the graph to illustrate that during wet years, peak flows are often much higher than the 2000 cfs peak we saw in 2015 during the Chester turbine test. In wet years, runoff is passed through a full reservoir, and this natural runoff flow often exceeds 2000 cfs during late May and early June. Then you can see very low flows throughout the rest of the summer in 1999 because irrigation demand was met by the river’s natural flow, and delivery of storage was not needed. The point of this graph as a whole is to illustrate that aside from the abrupt flow increase and decrease for the Chester turbine test, which is highly unusual, flow durations and magnitudes in 2015 were not much different than in other recent years and in fact were much better than they could have been, given how dry it was this year.

NOTE: After the presentation, someone asked why irrigation delivery remained so high throughout the late part of the summer in 2001, whereas it didn’t in 2015.  The answer is that often, even though Island Park Reservoir physically fills with water, Fremont-Madison Irrigation District’s storage rights account does not fill. This means that some of the water that is stored in Island Park actually belongs to water users in the Magic Valley, and this occurred in 2001, as it did again in 2013. When this happens, that water needs to be delivered to those users, even after irrigation demand in the Henry’s Fork watershed reaches its maximum earlier in the summer. In 2015, thanks to a lot of rain in May, all of the storage rights in the upper Snake River basin were filled, both physically and on paper. This meant that all of the water in Henry’s Lake and Island Park Reservoirs belonged to water users in the Henry’s Fork watershed, so no extra water had to be delivered beyond what our local irrigators needed.  Had we owed water down the river this year, Island Park Reservoir would have been nearly completely drained. Another question was how this year’s overall use of irrigation water compared with other years. It turns out that because of the rain in May, total irrigation withdrawals were near average if not even a little below average in some places, especially considering how dry the water year was. However, because natural flow was so low, average irrigation withdrawal used more net storage than usual. Island Park Reservoir was a good example; total summer-time discharge was actually below average, even though it was above average during late June and early July. But, reservoir inflow was far below average, resulting in use of 78% of the reservoir’s capacity. But, it could have been much worse. The rain in May saved us about 700,000 acre-feet of storage system-wide.

Water Quality

  1. Rob: Our $100,000 per year water quality monitoring program was established last year after listening to your concerns and the concerns of our members. Melissa Muradian is a post-graduate researcher with a Master’s degree from the University of Washington. When she arrived this summer, I handed her the entire program and said, “Here you go!” So here’s Melissa to talk about what we’ve learned so far.
  2. Melissa Muradian: If you see someone poking around in the river with an eight-foot pole with a water bottle at the end of it, that’s me taking water quality samples. Come say hi!
  3. Melissa: I want to talk to you guys about some work we’ve done to monitor water quality changes in this area of the Henry’s Fork. Here we have an aerial view of Island Park Reservoir, here’s the dam, Box Canyon boat ramp, and Buffalo River.
  4. Melissa: First, I want to talk about summer water quality inside of Island Park reservoir. We took what are called depth profiles using a sonde, shown here [bottom left photo]. I know you can’t see it very well, but this is an instrument that measures different water quality parameters. For the purposes of this talk, we were primarily concerned with looking at changes in temperature, dissolved oxygen, and turbidity at depth. We and Jack, the intern from DEQ shown here, went out on a boat into the reservoir on June 22nd with a sonde in continuous logging mode where it takes measurements from the water several times a second. We slowly lowered the sonde down to the bottom of the reservoir and then slowly brought it back up.
  5. Melissa: What we got is a “snapshot” of water quality through the entire water column for that location at that time; what we call a depth profile. Here you can see our depth profile for temperature. You’re looking at temperature on the x-axis and depth on the y-axis. Now the axes may throw some of you off – usually you expect to see the independent variable (here depth) on the x-axis, but I created the depth profile plots with the y-axis as depth, plotted as elevation above sea level so that the plots have a physically intuitive interpretation…
  6. Melissa: …Meaning the top of the plot is the top of the reservoir (the water surface) and the bottom of the plot is the reservoir bottom. I drew the surface and the dirty bottom on this slide to help you visualize it! Note that the notches are increments of 10 feet.
  7. Melissa: Now that we’ve got that straight, you can see that water temperature at the surface was a little over 18 ˚C and remained at that temperature until we went 30 feet lower where temperature started to cool. We recorded a 5 ˚C drop in temperature in the next 20 feet of the water column, between 6,250-6,270 feet. Temperature continued to cool a little more slowly as we went lower, another 3˚ in the lowest 20 feet. Therefore, we saw an 8˚C drop in temperature from the surface to the bottom of the reservoir on June 22 upstream of the gates outflow.
  8. Melissa: Then we repeated these depth profiles at two more locations within the reservoir to see whether there are changes in water quality across location.  All three IP Reservoir sampling locations are shown on this map: sample location 1 is in front of the power plant intake, location 3 is in front of the gates intake, and location 2 we called “middle” since it’s in the middle of the other two sampling locations.
  9. Melissa: Now we can go back to our depth profile of temperature and add the two other locations. The power plant site, in blue, has a shallower bottom than the gates location, so the temperature profile ends about 30 feet higher; the middle location, in green, is shallower still. You can see the three temperature profiles are essentially right on top of each other; what that shows us is there is little change in the depth profile for temperature across locations, especially for middle and power plant. We did see a 1˚C difference in temperature between the gates profile and the other two at their lowest depths, which is a small difference. Then we repeated the profiles across date and location…
  10. Melissa: We went out on June 22nd, July 22nd, and August 3rd (columns) and captured depth profiles of temperature (top row), dissolved oxygen (middle row), and turbidity (bottom row). Here are the full results from our summertime reservoir depth profiles: The first thing I want you to notice is that within a frame, the three colored lines are pretty much on top of each other – with the exception of dissolved oxygen (DO) on June 22 (frame in first column, second row) and I am very tempted to chalk this up to messy data (we were learning how to best capture the depth profiles on this day). There is a longer time lag for the sonde to record DO than the other two parameters that we weren’t taking into account on this first collection day, so we were lowering and raising the sonde too quickly and the data suffered for this.The interesting comparisons are changes in the parameters across time (within a row) and there are several points I want to highlight:
    1. Notice that in all of the panels you can see the reservoir level drops over time as water is delivered downstream to supply irrigation demand
    2. We can see that surface temperatures rose over time with warmer weather (top row)
    3. DO surface levels rose over time, but DO levels at depth are very low on June 22, show an increase on July 22, and then dropped back down to dangerous levels by Aug 3rd; trout die in water with lower than 6 milligrams per liters (mg/L) DO (middle row).
    4. Finally, we see there is almost no change in turbidity across location or across date (bottom row): turbidity levels remained between 0 and 10 FNU (notice the x-axis is on a log-scale). For reference, THIS [holds up two mason jars of water] is 6FNU water and this is 0 FNU water – these are not very large changes in turbidity as far as water quality standards go. [FNU is an initialization for “Formazin Nephelometric Unit”, which is a measure of how much light is scattered by particles suspended in the water.]
  11. Melissa: In addition to seasonal weather changes driving some of the changes in reservoir water quality, I want to show you what was happening with reservoir outflow during this time to help explain the differences in water quality across time. Here is a plot of gates outflow during the summer. A max of 960 cfs exit the reservoir through the power plant and anything greater than 960 cfs exits through the dam gates. You can see an increase from 0 cfs to 700 cfs during mid-to-late June when irrigation demand rose and then a drop in gates outflow beginning in mid-July as demand decreased. You can also see the flow test performed on July 22nd when total flow jumped up to 2,000 cfs resulting in gates outflow of about 1,100 cfs. Then after the flow test outflow on the gates side dropped to zero, so there’s only outflow from the power plant.
  12. Melissa: Therefore, you can see that when outflow is greater than inflow (which we saw on July 22 for our depth profiles, middle column) the water in the reservoir gets mixed, so there’s less stratification in the water column and higher DO at depth. Regarding the latter point: you can see that the lowest levels of DO during high flow was 5 mg/L (middle panel middle row), but DO levels during lower flow went down to 2 mg/L at depth. Regarding the former point, you see that temperature on July 22 was between 17˚ to 20˚ – a 3˚ spread, but a week later on August 3rd temperatures were between 15˚ and 21˚ – a 6˚ spread, so the water column was more stratified. Once again, we didn’t see a strong signal in the turbidity data across location or date.
  13. Melissa: Now that we have an idea of what was happening in the reservoir during the summer months, I want to talk about changes in water quality downstream of the reservoir. We wanted to know how summer water quality downstream of the reservoir was (or wasn’t) impacted by reservoir outflow. Note: this picture was taken looking south from IP Dam on June 22 when discharge was approximately 200 cfs out of the dam gates.
  14. Melissa: First, I’ll talk about how we collected the data to be able to answer our research question: During the summer we had two sondes permanently housed downstream of the reservoir – locations 4 and 5 on this aerial view. These sondes recorded water quality measurements every 15 minutes from May 18th to August 4th. Location 4 shows the DEQ sonde, which was housed immediately downstream of the dam gates outflow on the west side of the river. A little further downstream of this and upstream of the Box Canyon boat ramp we had our sonde on the east side of the river, capturing water mostly flowing out of the dam power plant.
  15. Melissa: This picture gives you a view of the dam, looking north. These sondes were measuring the water on opposite sides of the river from different intake points in the reservoir. The dam gates outflow is where anything over 960 cfs exits the reservoir and the intake for this side is on the bottom of the deepest part of the reservoir. The power plant outflow is on the east side and spills 960 cfs or less. The intake for the power plant is about 30 feet higher in the water column than the dam gates intake point.
  16. Melissa: First, I’ll show you data from the IP West sonde, the one directly in front of the gates outflow. Top is the same outflow graph you’ve already seen: you see the increase in summer irrigation demand rise and fall and the flow test on the 22 of July, then the end of irrigation demand right after this. DO is in blue, Temperature is in green, and Turbidity is in Orange. First, let’s talk about the large-scale changes in water quality parameters through time focusing on DO and temperature. If we compare changes in water quality parameters to changes in outflow, beginning on the left (earliest in time and moving on in time) we see DO is steadily dropping as temperatures rise, then we let a little out of the gates side (100 cfs) and no response. Then once outflow increased to 200 cfs there’s a big response in DO and temperature: 1) DO increased from 5 mg/L to 8 mg/L – this is great news for trout who really need water with DO of or greater than 6 mg/L and 2) Temperature decreased from 13 to 11˚C. Let’s look again at what water quality was like on the other side of the dam at this time.
  17. Melissa: Looking in the first column – June 22 – we see that temperature at the gates intake was indeed around 11˚C (look at the very bottom of the top-left plot), but DO was much lower than what we saw downstream (middle plot first column), around 2 mg/L. So, what happened, where did the extra oxygen come from?
  18. Melissa: Once flow is above 200 cfs, there is enough force pushing the water to actually churn it up and that turbulence mixes air into the water. So, letting more water out of the gates side improved conditions in the water for trout by lowering temperature a couple degrees and mixing in a lot more oxygen. Now let’s talk about our turbidity data for a bit (bottom panel in orange): First, data before mid-June was far too variable (full of logging errors) to see any kind of pattern, so I removed it. These eye-catching spikes you see in the time series are also errors in the data, they are times when, say someone walked in front of the sonde and kicked up sediment, or a fish swam in front of the sonde probe, or a plant is caught on the sonde end and waving back and forth in front of the probes. We know these are errors since the data looks like “5,5,5,15,5,5,5, etc”, individual data points that are high, but a true change in water turbidity in the river would last longer than 15 minutes. So, read this turbidity data by following the dark, filled-in minimum line. At the higher outflow from 100 to 200 cfs, we didn’t see an increase in turbidity beyond an increase in the variability of the data.
  19. Melissa: During the period where outflow is above 200cfs (approximately June 15th to July 22nd), we didn’t see a large response in downstream DO (blue line). Temperatures continued to rise, but that is following the seasonal trend of hotter air temperatures, and turbidity levels increase from about 5 to about 7 which is a very small increase in turbidity, one that you would most likely not notice. I want to take a moment and point out these regular cycles in the DO series that are present when gates outflow is less than 200 cfs – these are daily cycles caused by photosynthesis of all the rooted plants in the river. As daylight nears an end, the plants have stored tons of energy and are chugging away at full capacity producing oxygen, then as they run out of energy through the night, their oxygen production slows and that oxygen is still being used by the system so we see a dip in DO. Then the sun return to the sky DO levels go back up and the cycle starts again. Once gates outflow was above 200 cfs, there was enough mixing of oxygen in the water that these cycles are dampened, their contribution being masked by the churning from the outflow. Alright, going on in time – we see a drop in DO after the gates outflow dries up and the trend returns to what it was doing prior to the outflow – displaying those daily cycles and steadily decreasing. Average turbidity dropped by a small amount after gates outflow was shut off.
  20. Melissa: I wanted to give some visuals of the difference between high flow and low flow downstream of the dam: This slide shows approximately 600 cfs coming out of the gates on June 30 (top 2 photos) and about 700 cfs outflow looking downstream on July 9 (lower photo). You can see that churning on the west side of the river and how far downstream it extends.
  21. Melissa: These photos show the same perspectives with 0 cfs out of the gates and about 900 cfs coming out of the power plant side after irrigation demand dropped off.
  22. Melissa: In this graph, I’ve added in the data from our IP East sonde in black. Recall the IP East sonde is downstream of the power plant side before the water has had much chance to mix with flow out of the gates side. Therefore, the IP East data in black shows us what downstream water would look like during that time of year if there was no increased flow out of the reservoir for irrigation. The first thing I want to talk about is this big gap right where we’d like a bunch of data (end of June to mid-July) – there was a power failure in our sonde, unfortunately we had a bad batch of batteries that ran dead far too quickly and so information during this period is lost. However, we do have information at the beginning and the end of the high-flow period which is enough to see that the trends in temperature and oxygen would have continued on without interruption if there was no gates outflow. There wouldn’t have been the drop in temperature we see from the IP West data nor this wonderful period of increases oxygen caused by the physical churning of the water from the outflow. Turbidity would have been a little lower, a couple of FNU, which is a small change in turbidity. So even with lousy batteries, the results are clear – higher gates outflow resulted in improved summertime water conditions for trout, namely lower temperatures and higher dissolved oxygen.
  23. Melissa: Very briefly, I want to mention we also take water samples every week from location 6 on this map to monitor water quality downstream of IP Reservoir – that’s when I’m out there every Tuesday poking around with my pole and bottle. We take water samples once a week at 7 locations all along the Henry’s Fork between the Flatrock fishing club and Parker-Salem (this map shows 1 of those 7 locations). We analyze the water samples for 3 things: turbidity (to ground truth our sonde turbidity data), suspended sediment, and phospohorus. The latter two parameters, sediment and phosphorus, are key water quality parameters important for river hydrology and for trout life history that our sondes cannot measure. We use these weekly samples to monitor sediment and nutrient levels in the water, which is what Rob will talk to you guys about next.
  24. Rob: These graphs show total load of suspended sediment and phosphorus coming into and out of Island Park Reservoir, and also at Pinehaven. Over the past 16 months, about the same amount of sediment has been exported into the river as has flowed into the reservoir. Sediment comes into the reservoir from tributary streams during spring runoff and then is delivered out of the reservoir during high flows during the summer.  Although the timing of outflow is different than it would be naturally, at least over the past couple of years, inflow and outflow of suspended material has roughly been a wash—output has equaled input at around 3000-3500 tons, which would have occurred even if the reservoir were not there. This may seem like a lot, but put that in comparison to the 1992 drawdown of Island Park Reservoir, when between 50,000 and 100,000 tons of sediment were delivered into the river in just two weeks. Sediment load at Pinehaven is almost twice what it is at Island Park Dam, primarily because of export of sediment that has been trapped by macrophytes between Last Chance and Pinehaven. This sediment is released primarily during the winter and spring, when macrophyte biomass is at its minimum, allowing sediment trapped during the previous summer and fall to be released into the water column and be transported out of Harriman and past Pinehaven. The bottom graph shows that about twice as much phosphorus comes out of the reservoir as comes into the reservoir from streams. This occurs for two reasons: 1) there is a lot of phosphorus sitting on the bottom of the reservoir that has been stored in sediments, and 2) there is probably a lot of direct input of phosphorus from septic tanks around the reservoir. In any case, the phosphorus loads in the river downstream are very high and potentially create a water-quality problem. I’ll talk more about that later.
  25. Rob: This summer we heard a lot of complaints about how turbid the water was coming out of Island Park. So for the month of July, we added total organic carbon analysis to our water quality monitoring program. At $65 a vial for water quality analysis, it was expensive, but from it we learned just how much of the turbidity coming out of Island Park was organic carbon versus mineral sediment. As it turns out, between 40-80% of the suspended load coming out Island Park is organic carbon, which contributes to the food web downstream. It does not degrade the quality of gravel needed for spawning or for insect habitat. The fraction of organic matter in the suspended load was somewhat dependent on flow—export of organic matter was generally higher when flow was higher. However, later in the summer, export of organic matter increased with time, regardless of flow, as growth of algae, cyanobacteria (“blue-green algae”) and other phytoplankton (tiny photosynthetic organisms suspended in the water column) reached its peak in the reservoir.


Fisheries Management Issues

  1. [transition slide]
  2. Rob: I’d like to conclude with some big-picture issues facing fisheries throughout the watershed. We are following and paying attention to all of these issues, although our ability to do anything about them varies, depending on location and issue. Starting in the upper river, major issues that affect the fishery and angling experience there are conflicts among recreational user groups, erosion and land-use issues, flows in Henry’s Lake outlet, and lack of physical habitat. We know from our research in Harriman State Park that in these low-gradient reaches of the Henry’s Fork, physical habitat structure for trout is provided by aquatic vegetation (macrophytes). Macrophyte growth is much lower in the upper Henry’s Fork than below Island Park Reservoir because nutrient concentrations are much lower above the reservoir. In addition, lower nutrient concentrations mean that there is less overall productivity to support fish growth.
  3. Rob: Phosphorous is a limiting nutrient in for productivity in aquatic systems. At Flatrock (in blue), phosphorous levels are low – which is reflected by lack of macrophyte productivity at Flatrock, when compared with that at Island Park or Pinehaven. Note that 0.10 mg/L is the recommended maximum phosphorous concentration for streams and we meet or exceed that maximum every year downstream of Island Park Dam.
  4. Rob: As a result, presence of large fish in the upper Henry’s Fork is dependent on the fishery in Island Park Reservoir and migration of those fish seasonally between the reservoir and the upper Henry’s Fork. This presents much larger challenges to maintaining a quality fishery in the upper Henry’s Fork than simply focusing on the river itself.
  5. Rob: From Island Park Dam to Riverside, we know that the single factor that limits trout population size is winter flow, and maintaining high winter flow below Island Park Dam is always our number one priority, even though it sometimes means lower flows at St. Anthony than are optimal there. There are obviously some other issues in this reach of river, including management of the Harriman Canal, the potential for a 1992-type sediment delivery event if the reservoir is drawn down too low, and the higher turbidity we have been seeing the past few summers. However, the two biggest issues are winter flow and this last item, the potential for oxygen depletion due to TOO MUCH aquatic productivity. Especially if we experience nuisance growth of plants or algae in the river during hot weather, there is potential for a fish kill due to oxygen depletion. Winter flows and a potential oxygen-depletion event are my main concerns for the fishery in this reach of river.
  6. Rob: In the lower part of the watershed, including lower Fall River and the lower Teton River, the biggest issues are low flows and high water temperatures during the summer. In addition, we saw some nuisance algae growth in Ashton Reservoir this summer, which highlights the potential for some water-quality problems there, too. Large flow fluctuations over short time periods are also problematic in the reach between Chester Dam and the Parker-Salem Road. Some of these fluctuations are caused by irrigation and power-plant operations, while others are caused naturally by macrophytes.
  7. Rob: Daily cycles of macrophyte photosynthesis and respiration cause daily cycles in flow in the river. The plants literally cause temporary damming and release of water on a daily basis. These cycles originate in the Harriman State Park reach of the river and persist all the way to Rexburg, although these cycles are most noticeable between Ashton Dam and the railroad trestle downstream of St. Anthony. They are not caused by operations at Ashton Dam as previously thought. Since HFF installed water-quality sondes at Pinehaven and the old Marysville Bridge last summer, we have been able to observe these cycles and describe their properties. This graph shows the origin of these daily cycles and their propagation downstream during the first week of September in 2014. The vertical axis scale on each graph is river depth relative to the mean so that changes in river depth are comparable across different locations along the river. The vertical axis scale is 8.5 inches, so you can see that when the cycles were at their maximum, water depth at Pinehaven was changing about 5 inches over the course of the day, corresponding to a change in flow of as much as 300 cfs.
  8. Rob: This next graph is the same data from the same week in September in 2015. You can see that the cycles have been less frequent and much lower in amplitude than they were in 2014. The largest dip we have seen in 2015 was after the decrease in flow on September 2. This decrease coincided with the decreasing part of the daily macrophyte cycle, causing a very long period of decreasing flow. At Ashton Dam, the automated reservoir-level gage sensed this long downward trend and continued lower outflow to match that trend even after the decrease in flow had reversed. As a result, the outflow adjustment overshot the inflow a little, resulting in a very low flow on the morning of September 4. Since then, flows have been relatively stable. We have talked with Rocky Mountain Power about these macrophyte cycles and flow management at Ashton Reservoir, and we think there is some possibility to smooth flows a little but at Ashton, regardless of the source of flow fluctuations there.