Forum Index > Trail Talk > Can't buy me rain - record dry September?
 Reply to topic
Previous :: Next Topic
Author Message
gb
Member
Member


Joined: 01 Jul 2010
Posts: 5916 | TRs | Pics
gb
Member
PostTue Sep 20, 2022 7:38 am 
We need rain in Western Washington to finally get rid of the existing fires and fire threat, but we haven't gotten piddle this month and don't look to get more than that amount through maybe the first week of October. Many mountain areas west of the crest have probably not had even one inch of rain since mid-July. Seattle averages 3" in August and September; or rather it used to average that. Oregon has done better further south and California recently considerably better, but not Washington (save for thunderstorms along and east of the crest). Even near Mt. Baker and last week in the Soleduck, vegetation is tinder-dried, all crinkled-up. Amazingly in some of these areas there is a good blueberry crop, but mostly in areas with favorable drainage.

Back to top Reply to topic Reply with quote Send private message
gb
Member
Member


Joined: 01 Jul 2010
Posts: 5916 | TRs | Pics
gb
Member
PostTue Sep 20, 2022 7:50 am 
Sea-Tacs all-time record dry period was just a few years ago when it rained just .03" in 100 days, but I believe that started earlier than this year in June. But on that year the Baker area got .4" a few times during that period. Rainier missed out, then.

Back to top Reply to topic Reply with quote Send private message
Cyclopath
Faster than light



Joined: 20 Mar 2012
Posts: 5994 | TRs | Pics
Location: Seattle
Cyclopath
Faster than light
PostTue Sep 20, 2022 9:40 am 
I used to live in Queen Anne. I remember how much I hated the traffic Hempfest created, in late August, was always keenly aware when it happened. It used to get rained out fairly regularly. It used to rain in the summer here, that doesn't happen anymore.

Back to top Reply to topic Reply with quote Send private message
altasnob
Member
Member


Joined: 29 Aug 2007
Posts: 909 | TRs | Pics
Location: Tacoma
altasnob
Member
PostTue Sep 20, 2022 9:59 am 
I wish that rather than people just saying, I remember it used to rain here and now it doesn't, people would post the actual amount of rain in Seattle in August/September for the last 100 years. It is hard for me to find this data in graph form. From my understanding reading Cliff Mass, it actually isn't raining much less in Seattle in the summer, or annually. Maybe a very slight downward trend with a very slight upward trend in average temperatures. It's also my understanding that Seattle has always been remarkably dry from mid-July to mid September. So even if we are, say 50% drier in August one year than average, that has very little effect on our total water year because we never get much rain during this period. Anecdotal evidence is worthless. Don't the global warming models show that Seattle will actually become wetter, and warmer, in the future because of climate change?

breadcrumb, mike, rossb
Back to top Reply to topic Reply with quote Send private message
Hutch
Member
Member


Joined: 18 Jun 2009
Posts: 615 | TRs | Pics
Hutch
Member
PostTue Sep 20, 2022 10:00 am 
Don't worry, I'm sure it's just a coincidence.

Back to top Reply to topic Reply with quote Send private message
neek
Member
Member


Joined: 12 Sep 2011
Posts: 2120 | TRs | Pics
Location: Seattle, WA
neek
Member
PostTue Sep 20, 2022 10:25 am 
altasnob wrote:
I wish that rather than people just saying, I remember it used to rain here and now it doesn't, people would post the actual amount of rain in Seattle in August/September for the last 100 years. It is hard for me to find this data in graph form.
I'm pathetic at finding and analyzing weather data, but here's a first take, using data grabbed from https://www.weather.gov/wrh/Climate?wfo=sew and pasted into a spreadsheet.
Seattle area, August rainfall totals, inches
Seattle area, August rainfall totals, inches

breadcrumb, thunderhead, awilsondc
Back to top Reply to topic Reply with quote Send private message
altasnob
Member
Member


Joined: 29 Aug 2007
Posts: 909 | TRs | Pics
Location: Tacoma
altasnob
Member
PostTue Sep 20, 2022 10:32 am 
Does that show what Cyclopath says, that it is raining way less in August now in Seattle than it used to? Looks pretty constant to me with some anomalies both more than typical, and less than typical.

rossb
Back to top Reply to topic Reply with quote Send private message
philfort
Member
Member


Joined: 02 Sep 2003
Posts: 288 | TRs | Pics
Location: seattle
philfort
Member
PostTue Sep 20, 2022 10:40 am 
I wish there were an easy way to download that data instead of just having to use the website. btw, the last 3 years (2019-2021) have had significantly above average precipitation for September (which surprised me).

Back to top Reply to topic Reply with quote Send private message
neek
Member
Member


Joined: 12 Sep 2011
Posts: 2120 | TRs | Pics
Location: Seattle, WA
neek
Member
PostTue Sep 20, 2022 10:50 am 
There's gotta be a better way to download data, curious if anyone has any tips. Hope the European model is right for the next 10 days.
ECMWF 10-day total rainfall forecast
ECMWF 10-day total rainfall forecast
GFS 10-day total rainfall forecast
GFS 10-day total rainfall forecast

Back to top Reply to topic Reply with quote Send private message
philfort
Member
Member


Joined: 02 Sep 2003
Posts: 288 | TRs | Pics
Location: seattle
philfort
Member
PostTue Sep 20, 2022 11:04 am 
I managed to copy past the monthly data into a google sheet. Here are the August and September 10 year trailing averages from 1920 until now. (I tried adding linear trendlines, and August's was flat, and September's was rising, but I really don't know enough about statistics or what calculations google is using to know how relevant that was...) It looks like there is a period of high precipitation in the 70s/80s, which may be why a lot of people feel it's getting drier.

breadcrumb, thunderhead
Back to top Reply to topic Reply with quote Send private message
rossb
Member
Member


Joined: 23 Sep 2002
Posts: 1568 | TRs | Pics
rossb
Member
PostTue Sep 20, 2022 11:06 am 
I think we all remember wet Septembers, but that doesn't mean it was the norm. There is also a huge difference between early and late September, in terms of average precipitation. Most graphs only show the month, which is why they show September as being similar to August (only a little bit wetter). It is possible we will get a storm late this month (with an inch or so of rain over several days) and be average. My guess is even at this point, this is not an exceptional month or period. Sometimes it is dry from midsummer to early autumn, then we get plenty of rain.

Back to top Reply to topic Reply with quote Send private message
jaysway
Member
Member


Joined: 16 Jul 2020
Posts: 216 | TRs | Pics
jaysway
Member
PostTue Sep 20, 2022 12:25 pm 
I'm waiting for a large model to finish running at work so I have a bit of downtime (hence this post and my one in the lazy reports thread) and decided to paste the data into Excel and then do a quick trend analysis in Python. Caveats: I'm not an expert on time series, hydrology, or climate change. My preferred method for trend detection is the Mann-Kendall test which is non-parametric and detects for increasing or decreasing monotonic trends compared to the null hypothesis of no trend. This test assumes that the data do not exhibit serial autocorrelation. I summed June, July, and August precipitation values as a proxy for summer precipitation totals from the Seattle/Tacoma airport and present it below in a time series from 1945-present. I'm assuming in the dataset that "T" means trace, so I replaced those values with 0s, and that "M" means missing, so I imputed those by month using monthly means. On the plots I included the trend test results as well as the Theil-Sen best-fit line which is more robust to outliers than linear regression. Here are the results:
The p-value on the Mann-Kendall test is 0.16, which basically means that if we assume that there is no trend in our dataset, we would detect a trend as or more extreme as the one we did about 16% of the time. Since our prefined p-value threshold was 5%, that means that we no not detect any increasing or decreasing trend in summer precipitation totals for the Seattle/Tacoma airport from 1945-2021. The lower the p-value, the more confident we can be that we ought to reject our null hypothesis of no trend, i.e. our data is incompatible with our null model of no trend. In practice, p-values should be weighed holistically with effect sizes and other evidence when conducting hypothesis testing - see the American Statistical Association's statement on p-values for more information. I did a quick autocorrelation plot to confirm that the time series does not exhibit serial autocorrelation, if it did we could use Hamed and Rao's variance correction in the Mann-Kendall test.
Out of curiosity, I also repeated the above analysis for just June, July, August, and September precipitation totals:
Note that I simply repeated my above analysis and did not correct for multiple comparisons. One key thing to note with trend analysis is that it is sensitive to the time period, thus one of the easiest ways to lie with statistics when it comes to trends is changing the starting time period. Since 1945 is the beginning of the dataset this is where I began all of my analyses. Another way to lie with statistics here is to p-value hack by trying lots of different comparisons and hoping that eventually we will find ones that give us the results we are hoping for, ex. for this analysis trying different months, different ways of defining "summer precipitation," trying different weather stations. My takeaway from the above analysis: while a bit interesting, the above is woefully incomplete and cannot tell us very much (or anything, really) about how climate change has affected summer precipitation in Seattle. Let's leave that to the climate scientists smile.gif. A quick note on climate change, which again is far from my area of expertise: human-caused climate change is an undeniable scientific fact. There is a lot of variability in how climate change affects different geographies. It rarely makes sense to point to specific anecdotes or single datapoints as evidence for or against climate change. While it's less sexy, trend analysis and other techniques that shed light on long-term patterns are key for understanding the effects of climate change. I typically put no more stock in someone saying "this heat/lack of rain is evidence for climate change" than someone saying "this higher-than-normal snowfall is evidence that climate change doesn't exist." Neither are helpful. It's best to think about climate change probabilistically and consider that natural variability and climate change work in concert to affect weather patterns.

ALW Hiker, rossb, breadcrumb, thunderhead, neek, philfort, zimmertr
Back to top Reply to topic Reply with quote Send private message
gb
Member
Member


Joined: 01 Jul 2010
Posts: 5916 | TRs | Pics
gb
Member
PostTue Sep 20, 2022 12:42 pm 
altasnob wrote:
I wish that rather than people just saying, I remember it used to rain here and now it doesn't, people would post the actual amount of rain in Seattle in August/September for the last 100 years. It is hard for me to find this data in graph form. From my understanding reading Cliff Mass, it actually isn't raining much less in Seattle in the summer, or annually. Maybe a very slight downward trend with a very slight upward trend in average temperatures. It's also my understanding that Seattle has always been remarkably dry from mid-July to mid September. So even if we are, say 50% drier in August one year than average, that has very little effect on our total water year because we never get much rain during this period. Anecdotal evidence is worthless. Don't the global warming models show that Seattle will actually become wetter, and warmer, in the future because of climate change?
And therein lies your problem. You are listening to Cliff Mass. It isn't the total amount of rain that has changed significantly, it is how we get it. It used to be that we would get summer systems coming out of the gulf of Alaska every two or three weeks. Thus, the amount of rain we would get in the summer would be fairly consistent month to month in Seattle (or any other NW weather station). What we get now are stalled low pressure systems along the coast which bring up subtropical moisture and hence allow for thunderstorm development and periodic heavy rains.(think tropical weather patterns). Stalling weather systems reason for: https://insideclimatenews.org/news/31102018/jet-stream-climate-change-study-extreme-weather-arctic-amplification-temperature/ As to how much rain we get per month in Seattle or any other NW weather station that is easy to google if you have googling capabilities. That is why I posted on line 1 that Sea-Tac gets on average 3" of rain in August and September. Those are irrefutable statistics but if you went back 40 years you would find that "average rainfall" in Seattle is front-loaded just as "average Seattle temperatures" are with cooler and wetter times in the past and warmer, drier periods more frequent in the last decade. Perhaps Cliff Mass hasn't lived here long enough to have an understanding or doesn't know how to look at statistics. This thread was not political but you are making and effort to make it so by blind reliance on what Cliff Mass says. He is not an honest broker. More precise statistics are available through the climate portal at the NWS, but instead you are relying on Cliff Mass one-liners. https://w1.weather.gov/climate/index_nonjs.php?wfo=sew Do your own research and don't make this a Political commentary. rolleyes.gif

Cyclopath  philfort
Back to top Reply to topic Reply with quote Send private message
altasnob
Member
Member


Joined: 29 Aug 2007
Posts: 909 | TRs | Pics
Location: Tacoma
altasnob
Member
PostTue Sep 20, 2022 12:43 pm 
Thanks for all the above posts. Very interesting. I guess we would also have to plot temperature because even if it is not raining that much less in the summer here, if it is also warmer then it would exasperate the dryness for the plants.

breadcrumb
Back to top Reply to topic Reply with quote Send private message
philfort
Member
Member


Joined: 02 Sep 2003
Posts: 288 | TRs | Pics
Location: seattle
philfort
Member
PostTue Sep 20, 2022 12:50 pm 
jaysway wrote:
My preferred method for trend detection is the Mann-Kendall test which is non-parametric and detects for increasing or decreasing monotonic trends compared to the null hypothesis of no trend.
Thanks for this analysis! Are the absolute units on the trendline relevant, or just the slope?

Back to top Reply to topic Reply with quote Send private message
   All times are GMT - 8 Hours
 Reply to topic
Forum Index > Trail Talk > Can't buy me rain - record dry September?
  Happy Birthday Mt. Frank, trailtrasch!
Jump to:   
Search this topic:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum