May 30, 2019

More on the Washington Drought Situation

In my earlier blog, I examined Washington precipitation and suggested that a "drought emergency" is really not a good description of what is going on now in Washington State.  Language and accurate communication of the current situation is important to ensure long-term credibility of both my science and of the political leadership of our state.

As noted above, a large proportion of the state has been declared in drought status, with a recent press release describing a "drought emergency".   The definition of drought is found in Washington law, as noted below.  Note that the definition is not directly connected with the actual snowpack or precipitation, but rather the "water supply" forecast, with drought being called when the water supply is predicted to be below 75% of normal and "likely to create undue hardship."     So there is a substantial subjective element to the definition.

Thus, although below-normal precipitation over the water year  (since October 1) is found over western WA (see below),  there is a huge hole of no drought in the figure above, because the Tacoma, Seattle, and Everett systems have plenty of water in their reservoirs.   Thus, there is no water issues for that huge block of WA State population. 

There have been several stories in local media and talk by some politicians that extreme low snowpack is being observed, with numbers in the single or low-double digits being thrown around.  Much of this talk is  ill-informed.   Let me explain.

The snowpack map from yesterday shows the Olympics only being at 6%.

But there really is very little problem and plenty of water.  Why?  Because, the key issue is that we had a warm spell and the snow is melting out a week or two early, giving nutty numbers during the final days.   

Want proof?  Here are the observations at the Dungeness SNOTEL site, location at 4010 ft in the NE Olympics.  Cumulative precipitation as of this AM (black line) was just slightly below normal (gray line).  The snowpack (blue line) was slightly below normal (red line) for the winter, but melted out a week early.  No big deal.  No emergency.  But is produced extreme low numbers for % of normal for a short period.

Let's take a look at the October through April precipitation for Washington State as a whole since 1930 (red line below).  The mean (cyan) and 75% of the mean (blue line) are also shown.  Yes, our last water year was on the low side, but more than 75% of normal.   There are many other years with similar winter precipitation as our past year (16 years were drier).

What about the coastal zone of Washington, which is in the center of the driest conditions?

Similar story.  This year was drier than normal, but nothing special.  And even the coast was above 75% of normal precipitation.

The bottom line of all this is that this year was drier than normal over the western third of the state, but nothing exceptional.

Modestly drier than normal years happen.  Most reservoirs are in good shape, the snowpack is melting a bit faster than normal, and as I will show below, more precipitation is on the way.   No emergency.

As described in my previous blog, the latest seasonal forecasts are predicting a wetter than normal summer (the Seattle Times had a headline that said the opposite, but it was in error).  But one must note that such forecasts are not reliable nor particularly accurate.   But shorter-term forecasts are much more useful and they show precipitation next week.

The European Center ensemble (many model runs) for the accumulated precipitation at Seattle over the next ten days shows most runs bringing rain back next week.

And the totals through next Saturday for the European Center high-resolution system will be appreciable in the mountains and in BC.  Wetting down BC is good for us, since that will work against early fires there.  Lots of precipitation in Oregon and CA.  Again, this will work against early fires.  

Why the precipitation over our region?  Because the models are emphatic that we will return to a moist, wet upper level flow pattern, with a trough of low pressure off our coast and strong SW flow moving in the Northwest (see below).

In short, we have had a winter with modestly below normal precipitation and an April 1 snowpack of about 75% of normal over the State.  Not unexpected in an El Nino year.  Our reservoir and dam systems are sufficient to provide normal water availability for nearly every user in such a circumstance-- in fact, it would be scandalous if they couldn't.   If we couldn't easily handle such a year, folks would need to question our Governor and State legislators about why they weren't ensuring a robust water system.

And there is no reason to expect a dry summer and  healthy dose of precipitation is in the forecast.

Clearly, the State legislature and Governor should revisit the drought definition being used and perhaps come up with new terminology (e.g., dry condition advisory).  The term "emergency" is a powerful one and should only be used in truly dire situations.


  1. But if fear isn't force fed to the people how will they keep their jobs? As always, thanks be to Cliff for being the voice of reason rather than another talking head pushing impending doom.

  2. Thanks for the clarification!

  3. It is scandalous that the dam system that we have cannot meet demand in a mildly short water year, but that is the case in the Yakima Basin. The US Bureau of Reclamation is forecasting a 75% supply for proratable Districts, meaning that they will have to shut off in September. Without the drought declaration, the orchards cannot turn on their "emergency" wells, and the trees will die.

  4. As always, thankful for facts.....

  5. The Dungeness SNOTEL is such a classic case. I've been following that one all winter since I live in the Dungeness watershed. It's interesting how one's mindset can alter the understanding of data. Do you think the world is actively spiraling down the flush cycle? Or are things drifting back and forth within normal limits, with some concerns for increased future variation?

    I guess when you get older you get a longer-term perspective and it's harder to get excited about such things.

    The Dungeness watershed looks well within normal, as it has all winter. Certainly not the wettest year, nor the driest. And yes, earlier final meltout than some years, but not outside normal variation except for the very end. Which is a very minor event given we were down to the tail-end by that time.

    Essentially, the last snowball melted a week early. If you do math on just that, you can get all kinds of excited and generate scary headlines. But it was just the last snowball.

  6. Cliff could you comment on Eastern Washington and in particular stream flow forecasts from snowmelt later into the summer. Many of the summer salmon runs are dependent on cold water, Sockeye runs up the Wenatchee, and into the Okanagon. thx

  7. @John

    You are lucky to live in such a beautiful area. I backpacked the Upper Dungeness in June 2017, and am heading back in a couple weeks. Gorgeous!
    From recent trip reports it looks like the trails are clearing 3 or 4 weeks earlier than when I was there before, but then again 2017 had an above average snowpack.

  8. Cliff, if not for this blog no one would know the truth and I hope everyone recognizes the heat you take for making these posts. Kahones of steel.

    But I will guarantee this: next fall - starting on October, we will start hearing the same thing. You can set your watch to it.

    Also - kudos to SPU for being honest this year about the fact there was never a problem, and kudos to their engineers for being pros. All in the midst of some of the most shameful fear-selling we have seen on this subject. The public is starting to get that they have been lied to over and over again (for multiple years, going on decades), and that knowledge is a real positive in all of this.

  9. One perspective that seems to be missing repeatedly from this conversation is the fact that drought is defined a few ways and on a few timescales. From the USGS:
    "A drought is a period of drier-than-normal conditions that results in water-related problems. When rainfall is less than normal for several weeks, months, or years, the flow of streams and rivers declines, water levels in lakes and reservoirs fall, and the depth to water in wells increases. If dry weather persists and water-supply problems develop, the dry period can become a drought.

    The term "drought" can have different meanings to different people, depending on how a water deficiency affects them. Droughts have been classified into different types such as:

    meteorological drought - lack of precipitation
    agricultural drought - lack of soil moisture, or
    hydrologic drought -reduced streamflow or groundwater levels
    It is not unusual for a given period of water deficiency to represent a more severe drought of one type than another type. For example, a prolonged dry period during the summer may substantially lower the yield of crops due to a shortage of soil moisture in the plant root zone but have little effect on groundwater storage replenished the previous spring."

    So- Cliff- you are talking about meteorological drought- as I would expect given your focus and academic training.

    But what would a hydrologist think? Groundwater works on a much larger and longer scale then simply the water that falls out of the sky and is more related to the longer term issues regarding water. Here's what the USGS had about that:
    "roundwater is important during all hydrologic droughts:

    Groundwater may be an alternative or supplemental source of water during periods of surface-water drought, if sufficient groundwater resources exist.
    Reduced groundwater levels due to drought or increased pumping during drought can result in decreased water levels and flows in lakes, streams, and other water bodies. (On average, greater than 50 percent of stream flow is contributed by groundwater. Groundwater also is a major source of water to lakes and wetlands. Source: Circular 1139, p. 12)
    Changes in groundwater/surface-water exchange can result in changes in water quality.
    Decreased groundwater flow to surface waters can affect aquatic ecosystems that rely on a continuous supply of groundwater to sustain aquatic habitats and stream flow.
    Reduced heads in aquifers can result in land subsidence."

    And agricultural drought I'm not as familiar with - but it can also be related to how much water is needed vs how much is available.

    So- I get that you often think that drought is overblown- but it seems like the actual issue is about more than meteorologic drought alone. And yes- a rapid snowpack melt out can dramatically reduce infiltration of water into the ground, as much more of it occurs as overland flow- so the simplistic assessment of amount of snow or amount of rain alone may not accurately reflect the actual problem.

  10. If I understand correctly, the citations of abnormally low snowpack percentages from SNOTEL data are essentially down to math illiteracy!

    By definition, in 50% of all years, snowpack will melt out earlier than normal (the other half will be later than normal). Every such spring will end with a period of very low percent-of-normal snowpack. This is nothing more than x/y as x goes to zero before y does.

    Consider a "90% of normal" snowpack at peak, followed by a meltout which perfectly mimics the average curve, so that the difference between actual and normal snowpack is constant over time:
    90/100 = 90%
    50/60 = 83%
    10/20 = 50%
    5/15 = 33%
    2/12 = 17%
    1/11 = 9%
    0.1/10.1 = 1%

    At peak, the snowpack is 90% of normal. One day before meltout, it's 1% of normal. Maybe you should be using this as a teachable moment w.r.t. math education in WA state...

  11. Ali,

    Consider this comment from Cliff's post on 5/20: Should California Be Renamed the Evergreen State?

    " Blogger GlacierBake said...

    I'm not sure what the groundwater situation was in the Whatcom County lowlands in 1958, but I've gathered the annual mean flow (streamflow) records (cfs) for the USGS Nooksack River gauge nearest to Glacier, 1938-2017, and produced a graph. (I'm a water commissioner who's lived her for more than 45 years, I've observed the ups and downs first-hand.) This graph includes the 1950's, and there was a dip around 1958. But clearly there's a gradual long term trend upward. I've put a PDF of the graph in my dropbox (link below), and I suspect that lower-B.C. (Fraser River) data is similar to what we've seen here. "Just sharing.""

    While it's not impossible, it's unlikely that a long-term downward trend in groundwater level would be correlated with a long-term upward trend in annual stream flow. This data is, of course, for one location within one watershed, but it's hard data, nonetheless, which has been compiled by an apparently reputable source. For what it's worth, since Cliff's post from 5/23 explicitly references the Palmer Drought Index, USDA soil moisture report, USGS streamflow report AND WA reservoir storage, I think it's safe to say that he has a comprehensive understanding of the components of the various conceptions of drought.

  12. Thank you so much Cliff for clarifying (with clear and undeniable science) what most of us already knew. That there is no drought emergency. Even in 2015, that spring and early summer were very dry and it looked like we were off to the races of a record breaking drought. But we got a big rainfall in late July and the Fall rains began early that year,in late August. The following water year was also much wetter than average. So much for the drought! The only drought emergency that we could ever experience in this region would be a perfect combination of low snowpack and below average winter rainfall, along with a warm dry Spring, a hot dry summer followed by a delay in the Fall rains. This is the only situation on which a true drought emergency should be declared, in my opinion, due to the widespread water shortages that would likely be the result of that situation. But to jump the gun and declare it now when the numbers prove it's not even that bad yet is premature. Is it drier than normal? YES but not that much drier. Could it become a serious situation? YES it could ,but like you said it's unlikely. Should the media and the State take a chill pill? YES!

  13. Regarding GlacierBake’s comment above:

    The trends for Whatcom County are similar to many other areas of the PNW. An increase in annual precipitation, + 0.60”/decade since 1950, likely explains the increase in annual streamflow. The slight decrease in summer precipitation, - 0.13”/decade, June - September, combined with glacier loss and rapid summer warming, + 0.4F /decade since 1950, has likely resulted in a decreasing trend in mid-to-late summer streamflow.

  14. Mark
    “Consider a "90% of normal" snowpack at peak, followed by a meltout which perfectly mimics the average curve.....”

    You seem to be making the following assumption - if a 100” snowpack decreases by 10” in a given period, a 50” snowpack will also decrease by 10”.
    Isn’t it possible the two snowpacks would decrease at a rate somewhat proportional to their starting points? I.e., if the latter decreased by 10”, the former would only decrease by 5”.

    I’m guessing snowmelt is a combination of the two ideas.

  15. Of course there's an emergency. Jay Inslee is running for president on the Climate Panic ticket. Isn't that enough?

  16. I agree it's just a little drier than usual- down here. But did anyone notice the reddish sun yesterday? They said on the news that it was fires in B.C.- NORTHERN B.C. so HOW could it already be so dry up there in snow country that it has caught fire?

    This doesn't portend well. I am hoping it's just another outlier, but I'm beginning to despair for our historically crystal clear (when not cloudy) summer skies.

  17. Where I agree with Cliff is in this- the definition of "drought" is often a muddy concept when you look at the difference between the colloquial definition and the government definition.

    As Cliff said, "The definition of drought is found in Washington law, as noted below. Note that the definition is not directly connected with the actual snowpack or precipitation, but rather the "water supply" forecast, with drought being called when the water supply is predicted to be below 75% of normal and "likely to create undue hardship." So there is a substantial subjective element to the definition."

    Yup- because "Drought" is mostly relevant to how much we have vs HOW MUCH WE NEED. Hence the agricultural drought concept, etc. At least in terms of how it's being used by the state. That definition might well be different than the meteorological drought- which is how much precipitation falls from the sky, and when. Which might be different then the hydrologic drought- streamflow and groundwater situation. Which might be different from.... you get the idea.

    So.... this is readily observable on all the different maps and datasets on This isn't new, but the idea that the drought predictions or decision making is somehow persistently "wrong" because the definitions are used differently isn't a super useful construct here.

    So Cliff- instead of repeatedly writing the same blog- I would be way more interested in hearing what you think a more useful assessment would be and why- and what you think a more useful working definition is for the state and all of the users that rely on water. Disagreeing is not an issue, but often the conversation stops there and the takeaway points that your faithful readers get just seems to be the usual "bad gov't" etc. Which doesn't really help solve much of anything.

  18. It's not a "drought" if there's less water than people need. That can be caused by all kinds of non-climate, non-weather issues. It's a "drought" if soil moisture goes below a threshold on account of a persistent abnormal lack of precipitation.

  19. Ali,

    Once again, the answers you seek are contained within the content of the post. The usefulness of the assessment is contingent upon the definition of drought being employed and the accordingly associated terminology. The substantial subjective element that Cliff mentions in reference to the government definition of drought is similarly present within your own: how much we have vs. how much we need. Just how much do "we" "need"? At what level of supply will "undue hardship" occur? Undue for whom? Hardship being defined as..? If there is no definitional agreement and terminology is fast and loose, for (activist) political purposes, then objective measures become meaninglessly subjective.

  20. Ali - thank you for that very useful comment.

  21. Above normal temps and below average rainfall seems to be the OWSC forecast for June and this summer in Western WA. Am I reading this wrong?

  22. To Ali and others: The keywords that matter in the drought statute are "supply" and "hardship." In the Northwest much of the "public water supply" is groundwater which recharges during rainy months. Little credible work has been conducted to quantify groundwater supply in our water resource inventory area (WRIA 1), so the matter of "what 75% of supply" means is highly uncertain. "River water supply" forecast ensembles are constantly updated, based on a combination of streamflow data and weather models. Even when streamflow declines during dry periods, that flow may have no direct bearing on the state of overall supply. Groundwater can be ample even during dry stretches when runoff ceases and the water table drops below stream level. See this short COMET Program hydrogeology ed video (The Fate of Rainwater in a Basin) that illustrates the mechanics of rain and groundwater on streamflow:

    The chart of USGS flow statistics that I shared reflects the long-term trend upward in surface (stream~) flow locally in our headwaters location (in Whatcom County, near Glacier). The flow total is surely the sum of both runoff and baseflow. In the 45 years that I’ve lived here observing both weather and streamflow, snowpack has varied widely (world record snowfall on Baker, for example) but the general groundwater supply has remained very stable. It seems to me that, at least west of the Cascades, snow and glacier conditions may not correlate as-well to water "supply" as they do east of the mountains. Reliance on snowmelt here often seems overstated. I have yet to see minor differences in snowmelt timing (a week "early or late") to have a major effect on river or stream condition. (Mother Nature has deadlines for such things?)

    Weather is an observable natural phenomenon subject to scientific study, and this is blog is a great place to visit and weigh in about weather reality and weather's effects. I have a real distaste for the politicization of weather and hydrogeology. In the 20 days since Inslee expanded the drought area (May 13), we've had 12 days with measurable precipitation totalling almost two inches. Only time will tell if the official predictions play out in terms of hardship.

  23. Unkown - same for BC

    Buy a respirator and stock up on O2 bottles

  24. The definition being used to label a condition as drought is not useful for predicting wildfire potential. Wildfire potential has more to do with the fuels, temperatures, humidity, winds, and ignition hazards. A winter or spring snow pack has little effect on fire behavior later in the season. If fact, a good snow pack enables more vegetation (fuel) growth that, when it inevitably dries out, will burn hotter. All the water in the lakes (dammed, or otherwise) has no effect on the fire potential of dried out forests, lowlands, or suburban neighborhoods. Would Paradise CA have not burned if there was a heavy snow pack on the mountains, miles away? How about if the nearest lake been full?

    An experimental way of determining fire potential has no mention of "drought" conditions is being developed by the USFS called the HDW (Hot-Dry-Windy) model.

  25. H.L. Mencken said it best; “The urge to save humanity is almost always a false front for the urge to rule.”

  26. To GlacierBake -

    My interest in the topic isn't so much political - because I agree with you that politicization distorts the conversation! My background and career is in geology and for a while- climate science- and thus I think the intersection of science and how we have to make decisions is interesting. For a long time I lived in the science and research side so much that it was hard for me to understand HOW these topics could be such a political weapon. I sometimes think I should have gone into hydrogeology - so don't take my commentary as an attack on you at all. I like what you have to add and my questions for Cliff are more about trying to refine where the concepts are muddled. The gulf between what terms mean for scientists and what terms mean for governing bodies and what terms mean for the general public are obviously a problem :)

  27. Reports today are of below normal snowpack and dry conditions with above normal temps for the summer.

  28. While the current situation isn't all gloom-and-doom, I do think that some of the outlooks presented here are over-simplifying the hydrology of the region.

    First of all, for sure using % of average for snowpack gets wonky when snowmelt is early. However, you conveniently picked a station where melt-out conditions are only a week ahead of normal. If we look at other stations around the state, we can see that the melt out can vary up to as much as a month earlier than normal ( ) This is much more significant than a week, and particularly looking at the outlook for summer for snowmelt-driven rivers, the influence from the early snow melt will continue to persist, in effect leading to low flow conditions that are 1-4 weeks ahead of normal, and extending the vulnerable period later in the summer.

    A better metric for snow pack is to look at percentiles, to give us a better sense of how "unusual" the current conditions are. Currently most SNOTEL sites have been recording around the 5th-15th percentile - maybe not a completely extreme situation but certainly on the moderately infrequent side of the spectrum(

    On the river side, for sure in areas where there is reservoir storage, shifts in timing and relatively modest, if at all, long range cumulative precipitation defecits should be easily accommodated by reservoir management. Where the challenges lie is in rivers where snowpack plays a minimal role in spring/summer flows, and in areas with no reservoir storage to buffer the impacts of dry weather or early snowmelt. If we look at current streamflow across the state ( ) we can see that indeed rivers are very low, with most rivers in the 5th-10th percentile range. For most of the coastal or lowland rivers, snow isn't that important as control on the hydrology, and rivers are very flashy in that they respond more to short-term rainfall rather than long term cumulative rainfall. As a result, the wet fall-early winter that we experienced is really not in the memory of the current flow, and the dry weather over the last 4-6 weeks is more of a driver. Again, this doesn't matter if we have storage, but for rivers that don't we are seeing extremely low flow for this time of year.

    While a number of factors are stacked to increase the risk of low flows this summer (again for those rivers that don't have storage) it might not be cause for concern for well managed water suppliers, however for folks that draw water from unregulated rivers (eg irrigation) or for flow requirements for fish, the current outlook should be of concern.

    Of course summer weather is the critical driver for how streaflow plays out this summer, and as Cliff notes there isn't really a reason to expect, or not expect, it to be dry.


Please make sure your comments are civil. Name calling and personal attacks are not appropriate.