Friday, June 2, 2017

Daily Melting Seen in Local Stream Flows

David New of Jones Engineers in Bellingham sent me an interesting email about the flow in some of our local streams.  Specifically, their flow is heavily modulated by the outside temperatures, with warm temperatures causing snow melt that surges downstream.   You can see the daily (diurnal) variations as snow melts during the day and surges downstream later in the day, producing a midnight maximum of flow.

We start with an unusually high snowpack for this time of the year (see graphic), with some regions at 200% of more of normal.

 And our temperatures have varied substantially (Sea-Tac temperatures shown below), with a a particularly warm period from May 26-29th.   So there was a lot of melt during that period, with less during the near-normal temperatures of the past few days.
Let's consider Pilchuck Creek, found along the Snohomish - Skagit County line and includes Lake Cavanaugh. It drains 75 sq. mi. upstream of I-5, with the highest point in the basin just under 4000 feet. It is just under 20 miles from the remaining snowfields to I-5. The gauge data is from a Dept of Ecology site near I5 (see map)
The streamflow at that site from May 26 through June 1 (below) shows large flow and large daily variations during the warm days (May 26th-May 29th), with a maximum value near midnight (the vertical lines are at 5 PM).   But as soon as the weather cooled, the streamflow declined, as did the magnitude of the daily modulation.

The temperature of Pilchuck Creek also varied substantially (see below), with large swings of temperature during the warm days.   The lowest temperatures matched those of the largest flows near midnight, with temperature peaking during the afternoon.  On the the warmest day (May 28th), the stream's temperature varied from 14.2 to 18.2 C, a change of 4C (or roughly 7F)

The streamflow at Pilchuck Creek over the entire water year (since October 1) is shown below.  Lots of variation forced mainly by rain events, but melting also produces variability.  The snowmelt variability is small potatoes compared to the storm events.

What about the temperature of the water over a longer period?  Here is the water year plot. The coldest temperatures were in early January (around 4C), with water temperatures zooming up to around 16C today.  Keep in mind that bigger rivers can be much cooler.



Brian Pernick said...

Cool stuff Cliff!

I run white water rafting trips on the Sauk river, and have always thought the water was cooler on the days with the biggest diurnal spikes, or after the hottest days. I have never had proof though, until now!

As you can imagine, we get may of the same questions from our guests. How deep is the river? Are we gonna get wet? At what elevation do the deer become elk... of course, we have some "creative" answers, but I try to be informative.

One question is, "How cold is the water?" My answer is, "It was snow 12 hours ago." I base this on the signal from the gauge(downstream of our run) which peaks at midnight to 1am. If peek melt is during midday, then 12 hours seems right.
What do you think Cliff, guide tales or reasonable hypothesis?

I have to mention that, this can be observed in person on one of our trips. You can find us at

Ellen Baker said...

Those graphs may look convincing, and might bear out in some places, but in our area (headwaters of the Nooksack River) I've been observing the weather and the river for more than 40 years, and daytime temperatures don't follow a regular 12-hour cycle. In fact, length of day varies quite a bit. In addition, with the exception of just a few days we've had measurable rain daily. Surface runoff/precipitation likely has had more of an effect on changes in river flow (cfs) than snowmelt has, by a long shot. At this instant (about midnight), it's a chilly 46'F at about 1000' in our valley, in Glacier, and supposedly it's a very cold 26'F on Mt. Baker (peak). What snow remains on the surrounding mountains does not contain a lot of water/SWE (snow water equivalent); much of that snow remains in shady contours well into July. There are a limited number of instantaneously-reporting stream gages that correlate to temperatures at higher elevations, or the depth of snow where snow remains. Bottom line - I think hypothesizing about the melt cycle is best-left to hydrogeologists. and even for those fellows, in the absence of carefully coordinated (truly scientific) observation/measurement - I think presuming that surface water "was snow 12 hours ago" doesn't quite fit. Stream and river water are always very cold here, with a suprisingly narrow temperature range/variation 9-10 months of the year.

Unknown said...

I spent Memorial Day weekend camped on the Stillaguamish. During the heat of the day, I was in and out of the river about every half hour to cool off. The water was cold, but not bone chilling. While I didn't notice any variation in water temperature, the volume and pace of the water increased markedly throughout the weekend. By the end, I was getting quite a ride in the current!

Stephen Murdock said...

Cliff's analysis of the data may or may not be correct but, I agree, the graphs do seem to support his hypothesis convincingly. Even if his inferences are incorrect, however, it is certainly no more wrongheaded to attempt to directly compare the upper North Fork Nooksack (length: 75 miles, basin area: 786 square miles, average discharge at mouth: 3800 cubic feet/second) with the lower reaches of the Pilchuck (length: 40 miles, basin area: 127 square miles, average discharge at mouth: 470 cubic feet/second). Clearly these two streams are not spatiotemporally equivalent not to mention that direct comparison of the upper reaches of one stream with the headwaters of another is entirely inappropriate even if the two streams were similar in terms of the physical parameters of their basins in addition to discharge.

Earthwater said...

One small error here. The plot that the narrative describes as stream flow is actually water temperature.
I did pull down the flow from that period, and it's true that the peak flows occur in the early morning. But other patterns are puzzling. The highest flows are on the 26th and 27th, but looking at data from Arlington Airport, the warmest days were the 27th and 28th. The highest flows actually occurred before the days got warm. Looking at the Finney Creek RAWS (in the Pilchuck headwaters), the daily maximum air temperatures increase every day from the 26th to the 29th, then plummet. The flows flatten out with the cool weather, but again, curiously, the peak flows drop every day even as peak temperature increase. The Elbow Lake SnoTel site is at an elevation similar to the Pilchuck headwaters and slightly to the north. Looking at Snow water Equivalent data, a clue to the mystery is found. There was 2 feet of snow at Elbow Lake on May 20th, 1 foot by May 27th, and it had melted off by June 5. So it appears that Cliff is spot on about the surging snowmelt causing peaks delayed until the early morning. But the rest of the story is that the snow was disappearing during that time.

Ellen Baker said...

At the risk of splitting too many hairs, the point I tried to make the other day was that many geologic/ hydrogeologic factors come into play. A foot of snow yields surprisingly little water, and some of that water very likely infiltrates, (some even evaporates), and my intent was to caution against over simplification in the absence of other essential details. Cheers.