Water Year 2018: Average supply but Above-average Reservoir Carryover

Photo of Brown Trout.
  • Water-year 2018 ended up close to average, at 102% of average precipitation and 105% of average streamflow.
  • Upper Henry's Fork subwatershed was below average in precipitation and streamflow, vs. above-average values in Fall and Teton rivers.
  • However, Upper Henry's Fork water supply improved from 70% of average in 2016 to 91% in 2017 and 94% in 2018, indicating recovery of deep aquifers from 2013-2016 drought.
  • May and June rain compensated for early snowmelt and resulted in below-average irrigation diversion.
  • Despite very dry conditions and below-average streamflow during July, August and September, Island Park Reservoir ended the water year at 73% full, compared with 43% full on average, thanks to careful and precise water management.

This blog is even longer and more detailed than my usual long, detailed blogs, but there was a lot to learn from water-year 2018. This blog is a slightly edited compilation of the four daily water reports I wrote during the first week of the new water year. The four reports appear in four different sections in this blog:

  1. Climate
  2. Natural flow and diversion
  3. Streamflow and water management in lower watershed
  4. Streamflow and reservoir management in upper watershed

Tables and graphs appear at the end of each section.

If you would like to receive my year-round daily reports, send an email to rob@henrysfork.org.

Before diving into the content, I'm including photos from water-year transitions in each of the past three years, each taken at the same location. The weather conditions in the three photos are strikingly similar, moisture streaming into the watershed as autumn takes hold.

Photo of rainbow over Henry's Fork

October 1, 2016

Photo of lower Henry's Fork

September 30, 2017

Photo of lower Henry's Fork

September 30, 2018

Climate

Overall, water year 2018 turned out to be pretty close to average, over a period of record that goes back about 30 years with temperature and 35-40 years with precipitation.

Watershed-mean temperature ended up 1 degree F above average, despite extended warm periods from mid-November through mid-February and mid-April through mid-June. These warm periods were more-or-less offset by extended cool periods in October, between mid-February and mid-April, and during late June. The summer turned out to be warmer than average, especially in July and September. August was a little cooler than average due in part to cool weather during the second half of the month and in part to smoke-filled skies, which kept afternoon high temperatures a few degrees cooler than they would have been otherwise.

The water year started out with a dry October, although moisture was good going into the water year due to cool, wet weather during the last two weeks of water year 2017. November was on the wet side, while December was fairly dry. Water-year precipitation was below average until mid-February, when a long period of cold, wet weather finally arrived, bringing precipitation back up to average in late March. Heavy rain around Memorial Day and again during the second half of June brought water-year precipitation up to 111% of average in late June. After that, precipitation was very light and widely scattered in space and time, except for a widespread rain event on August 27 that brought one-half inch to the whole watershed. After the otherwise dry summer, water-year total precipitation ended up at 102% of average.

By subwatershed, Fall River consistently received more precipitation than upper Henry’s Fork and Teton through most of the winter and spring. Among the high-elevation areas, the Teton subwatershed received more summertime precipitation, as southerly monsoonal moisture occasionally moved far enough north to hit the Big Hole Mountains, Snake River Range, and southern half of the Teton Range. The very weak and disappointing monsoon season left the upper Henry’s Fork very dry during the summer. By the end of the water year, Fall River and Teton River ended up with 102% and 103% of average precipitation, respectively, while upper Henry’s Fork came in at only 97% of average.

Snow-water-equivalent (SWE) accumulation generally reflected patterns of temperature and precipitation. The water year started out with quite a bit of snow already on the ground at higher elevations from September’s wet, cold weather. Accumulation was quite a bit above average in November and tracked average very closely from mid-December through mid-March. The biggest snowstorms of the year happened in late March and early April, leading to a snowpack that peaked at 117% of average: 120% of average in the Teton watershed, 121% of average in Fall River, and 111% of average in upper Henry’s Fork. Total SWE accumulation peaked on April 18, 8 days later than average. Unfortunately, very warm weather in May and early June, along with heavy, warm rain around Memorial Day melted this promising snowpack quickly. SWE at all nine SnoTel sites in the watershed zeroed out by June 25, 30 days earlier than average.

Precipitation in the valley areas of the watershed turned out to be one of the more interesting stories of water year 2018 and compensated for the early snowmelt and dry summer in terms of irrigation need. Precipitation in April, May and June, while above average across the whole watershed, was above average by a much larger margin in the valley areas. At the end of June, water-year precipitation in the valleys was over 130% of average, and accumulated moisture availability in the watershed was 5 inches above average. This meant that irrigation demand was very low throughout the spring and early summer, as soil moisture remained good well into July. By the end of the water year, however, moisture availability had fallen back to average, meaning that a fully irrigated crop of alfalfa needed its average application of 37.5 inches by the time all was said and done. Despite the dry summer, water-year precipitation in the valleys still ended up at 112% of average for the water year; unfortunately for late-season irrigation need and soil moisture, the above-average precipitation was simply offset by above-average evapotranspiration, which is why net moisture (precipitation minus evapotranspiration) ended right at average. 

Climate summary tableTemperature graphPrecipitation graphSnow water equivalent graphGraph of moisture availability

Natural Flow and Irrigation Diversion

Consistent with average climatic conditions in 2018, streamflow in 2018 also turned out to be pretty close to average. All comparisons in this section are made to the 40-year period 1978-2017, which is the period of “modern,” daily, computerized water-rights accounting records in the upper Snake River basin.

As mentioned above, watershed-total precipitation was 102% of average, and peak snow-water-equivalent (SWE) was 117% of average. These factors, plus good moisture and baseflows inherited from water-year 2017, combined to produce natural streamflow of 105% of average during water-year 2018, despite a very dry summer. However, the distribution of water supply across space and time was not uniform.

On the spatial side, upper Henry’s Fork (watershed upstream of Ashton) received less precipitation and SWE than Fall River and Teton River. Upper Henry’s Fork received 97% of average precipitation, compared with 102%-103% in Fall and Teton rivers. Similarly, peak SWE was 111% of average in upper Henry’s Fork, versus 120-121% in Fall and Teton rivers. As a result, natural flow was only 94% of average in upper Henry’s Fork, versus 113%-114% in Fall and Teton rivers. Another interesting way to view these statistics is as a rank out of the 41 water years between 1978 and 2018. Natural flow in Fall River ranked 10th, and that in Teton River ranked 12th, decisively in the top third of water years experienced since 1978, whereas upper Henry’s Fork came in at only 24th, in the bottom half!

Low water supply in upper Henry’s Fork resulted not only from below-average precipitation in 2018 but also from low levels in the Yellowstone Plateau aquifers that provide the majority of streamflow in the upper Henry’s Fork. Outflow from these aquifers responds to precipitation at a time lag of 3-4 years, so outflow is still reflecting drought conditions in 2015 and 2016. Nonetheless, streamflow in the upper Henry’s Fork is gradually climbing out of hole dug by the 2013-2016 drought. This year’s flow, at 97% of average, was better than those of the last two years: 91% of average in 2017 and 70% of average in 2016.

Temporally, the water year was sharply divided between its first nine months and its last three. Streamflow was consistently above average between the beginning of the water year and the end of June, due to three factors: 1) good baseflows inherited from 2017, 2) early snowmelt, and 3) heavy precipitation in April, May, and June. I should emphasize the effect of the early snowmelt here. This kept streamflow very high during April and May, but that was a period of time when irrigation demand was low and reservoirs were close to full. Thus, the high streamflow did not contribute to meeting demands on the river, although high streamflow during the spring is good for moving accumulated sediment out of stream channels and maintaining long-term stream and riparian habitat conditions. Had that same snowpack lasted longer in the summer, natural flow would have ended up a little lower during April and May and a little higher in July and August. Only heavy rains in May and June saved 2018 from being a very dry year, with high demand on reservoir storage. Once those rains were over, streamflow dropped very rapidly and stayed around 85% of average for all of July, August, and September.

Speaking of irrigation, my daily reports emphasized all summer that the May and June rains resulted in good moisture in the agricultural regions of the watershed, reducing irrigation need by at least 100,000 ac-ft during the early part of the summer. Diversion reached average on only a few individual days between April 1 and July 9, which is the average date of peak irrigation demand. This year’s peak diversion occurred on July 11 and stayed well below average for most of July, as soil moisture remained good for several weeks after the end of the June rains. However, diversion increased again in August and remained at or above average for most of August and September, as the effects of very low precipitation and high evapotranspiration were fully realized throughout the valley. In September, much of the need for irrigation was simply to keep the dust down and soil workable during and after harvest. Through September 30, total diversion was 783,386 ac-ft, only 82% of the long-term average of 952,586 ac-ft. May and June rains were primarily responsible for the decreased irrigation demand, but I should point out that total diversion has been steadily decreasing since the late 1970s due to conversion from ditch-and-flood irrigation to pipe and sprinklers. I will provide final statistics on irrigation diversion and more on irrigation demand and efficiency after administrative irrigation season ends on October 31.

Tab;e summarizing water supply in 2018Graph of annual natural flow 1978-2018Graph of natural streamflowGraph of irrigation diversion

Streamflow and Water Management in Lower Watershed

At the May meeting of the Henry’s Fork Watershed Council, I presented my predictions for the upcoming irrigation season, including the relationship between carryover in Island Park Reservoir and summer-time flow target in the Henry’s Fork at St. Anthony. As St. Anthony flow target increased from 800 cfs to 1,000 cfs, predicted reservoir carryover dropped by 7,050 cfs. However, as St. Anthony flow increased from 1,000 cfs to 1,200 cfs, predicted reservoir carryover dropped by 11,171 ac-ft. More importantly, the uncertainty around the predictions increased substantially with increasing flow at St. Anthony. Discussion among stakeholders at the Council meeting indicated that a flow target of 1,000 cfs provided acceptable tradeoff between summer-time flow at St. Anthony and risk of low reservoir carryover. Subsequently, Fremont-Madison Irrigation District, U.S. Bureau of Reclamation, and Henry’s Fork Foundation agreed to a summer-time target flow of 1,000 cfs daily average at St. Anthony.

Streamflow at St. Anthony dropped rapidly over the last two weeks of June and reached 1,100 cfs on July 3. Storage delivery from Island Park Reservoir began that day, in anticipation that St. Anthony flow would drop to 1,000 cfs within a day or two. Delivery from Grassy Lake began on July 9. Aside from a few days around the August 28 rain, streamflow at St. Anthony remained near or below a daily average of 1,100 cfs until September 30. Over that 90-day period, flow at St. Anthony averaged 1,081 cfs. Daily average flow dropped below 1,000 cfs on 10 days, and over those 10 days, flow averaged 980 cfs. Reasons that daily mean flow fell below 1,000 cfs included unexpected increases in irrigation diversion, unavailability of personnel to make flow changes at Island Park Dam on short notice, rating-curve shifts at both St. Anthony and Island Park, and fluctuations in outflow from Ashton Reservoir due to sensitivity of the flow control system there. Based on my early-spring model, predicted reservoir carryover was 105,683 ac-ft at a St. Anthony flow of 1,081 cfs. As it turned out, actual carryover was 98,509 ac-ft, 7% below the predicted value. This small error, of course, was due to well below-average precipitation during July, August, and September, but the error would have been much greater had the flow target been higher. Overall, the decision to target 1,000 cfs proved to be sound, the model that helped inform that decision performed well, and implementation was very precise.

Although the St. Anthony flow target was used to adjust outflow from Island Park Reservoir and Grassy Lake, two other measures serve as indicators of need for storage delivery. The first is need for delivery of water from the Henry’s Fork to the Teton River through the Crosscut Canal, which diverts water at Chester Dam. Crosscut delivery was needed to meet demand on the Teton River from July 10 through September 14, coinciding closely with the actual period of storage delivery.

Second, my “600-cfs rule of thumb,” indicates need for storage delivery when natural water supply exceeds diversion by 600 cfs or less. Where did this rule come from? About 500 cfs needs to reach the last diversion on the Henry’s Fork downstream of St. Anthony in order for that canal company to divert its full water right. During the peak of irrigation season, the intervening three canals divert around 500 cfs, so the 500-cfs need at the lowest point of diversion is minimally met when the St. Anthony gage is at 1,000 cfs. Losses due to stream channel seepage on the Teton River downstream of the Crosscut Canal average around 100 cfs, essentially reducing natural-flow supply by 100 cfs. These two factors combine to require 600 cfs of natural flow in excess of total diversion to meet the 1,000-cfs target at St. Anthony without storage delivery. The 600-cfs rule indicated need for storage delivery between July 7 and September 30, again coinciding closely with the other indicators and with actual delivery.

Graph of streamflow at St. Anthony

Graph of reservoir carryover versus flow at St. AnthonyGraph of streamflow in lower Teton RiverSupply and demand graphTable of storage indicators

Streamflow and Reservoir Operations in Upper Watershed

As for water year 2018 in the upper Henry’s Fork (watershed upstream of Ashton), precipitation was 97% of average and natural streamflow was 94% of the 1978-2017 average, whereas these figures were above average for the rest of the Henry’s Fork watershed. Low streamflow was partly the result of below-average precipitation and partly the result of continued effects of the 2013-2016 drought on the deep Yellowstone Plateau aquifers.

The longest record of natural flow available in the Henry’s Fork watershed is inflow between Henry’s Lake and Ashton, which consists primarily of outflow from these aquifers. This record dates back to 1930. The 2013-2016 period was the driest four-year sequence of water years since 1936-1939. Recovery in 2017 and 2018 was substantial, as streamflow increased from 76% of the 1930-2018 average in 2016 to 97% in 2017 and 99% in 2018. These figures were 70%, 94%, and 97%, respectively, of the wetter 1978-2018 period. The severity of the 2013-2016 drought and the magnitude of recovery is evident in the first graphic below, which is this 89-year record of natural flow.

Starting at the very top of the watershed, inflow to Henry’s Lake was above average for most of the winter and spring, reflecting good precipitation early in the water year. Input from snowmelt was well above average during May and early June and even better than that in 2017. Spring rain helped keep inflow right at average until early August, when the effects of the dry summer began to show. During the middle of August, evaporation from the lake surface was slightly higher than inflow, resulting in a period of zero net inflow. Under natural conditions, outflow from the lake would have been zero during this time period. Henry’s Lake reached its peak volume during the late-June rain at a little over its nominal full-pool capacity. At a steady outflow of around 70 cfs most of the summer, the lake drafted to 87% full by the end of the water year, just a little above average.

Inflow to Island Park Reservoir was close to average for most of the winter and spring and a little above average during late April and May due to early snowmelt and the big Memorial Day rain event. The mid-June rain event barely brought inflow back up to average after recession from the early snowmelt. After that, inflow was well below average, due primarily to low reach gain between Henry’s Lake and Island Park. Although below average, this reach gain was higher than it was in 2017 during the middle of the summer, again reflecting recovery of the deep aquifer from the 2013-2016 drought. Evaporation loss from the reservoir surface was above average due to very dry weather all summer, consuming about 25 cfs of inflow. However, this evaporative loss was more than offset by gains from direct precipitation during the winter and spring, resulting in a net gain of about 8,000 ac-ft (6% of reservoir capacity) by the end of the water year. This fell about 2,000 ac-ft short of last year’s gain but beat average by that same margin.

Outflow from Island Park Reservoir matched natural flow very closely from the beginning of the water year through late April, as the reservoir started the water year right at its long-term April-1 average contents. The reservoir was filled with April’s snowmelt, after which outflow roughly matched natural flow again until late June, when Henry’s Lake delivery exceeded its inflow, and kept Island Park a little higher than it would have been at its outflow of around 600 cfs. As mentioned above, draft of Island Park Reservoir began on July 3—right at average timing—and ended on September 25, 12 days later than average. Over that time period, outflow from the reservoir ranged between 1,140 cfs at the peak of irrigation delivery in mid-July to around 450 cfs before the last outflow reduction was made on September 25. Mean outflow over that time period was 744 cfs, compared with a long-term average of 1,046 cfs.

No discussion of Island Park Reservoir outflow would be complete without mentioning the record shift at the USGS stream gage that occurred in 2018 due to growth of aquatic plants in the stream channel. At its peak in mid-August, plant growth had displaced a full foot of water depth in the stream channel. As a result the stream gage was reading about 300 cfs above actual flow when it was adjusted on August 17. Streamflow data from the Island Park gage are officially approved through the August 28 rating-curve adjustment. Since then, senescence of plants has moved the shift in the other direction, and by my best estimate, the water year ended with the gage reading about 200 cfs below actual flow. The inflow and outflow figures I have reported here will change a little once official data are reported by USGS.

As a result of careful management of the irrigation system in the lower watershed, Island Park Reservoir ended the water year at 73% full, compared with 43% on average. Grassy Lake and Henry’s Lake also ended the water year above their long-term averages. At the end of the water year, the three-reservoir system was 78.4% full, compared with a long-term average of 61.2% full. This equated to a net storage draft of 51,993 ac-ft from the 240,385 ac-ft reservoir system.

Since reservoir carryover is extremely important to irrigators in the Henry's Fork watershed and is the single most important factor determining winter flow, and hence survival of juvenile trout downstream of Island Park Dam, the most important observation from 2018 is that careful management provided well above-average reservoir carryover in a year with only average water supply and with below-average natural flow all summer.

Graph of annual upper Henry's Fork streamflow 1930-2018Graph of Henry's Lake inflowGraph of Henry's Lake volumeGraph of natural flow between Henry's Lake and Island ParkGraph of inflow to Island Park ReservoirGraph of gain from direct precipitaiton on reservoir surfaceGraph of Island Park Reservoir volumeGraph of Island Park Reservoir outflowGraph of regulated vs. natural flow at Island ParkTable of storage delivery and carryover