wx_statman

Yeah, the warmth in the thermal belt was unreal with that one. A couple stations were 78°F if I recall from memory in the 3,000-4,000 foot elevation zone. Some RAWS stations near Brookings were into the 80s with downsloping. The Gorge had a major windstorm with HWW criteria gusts extending into Gresham which is really rare.

Quick technical note about some of the dates I've given in this thread - the ERA5 reanalysis is set to UTC time so the days are actually 5pm-5pm PST. This might result in some of the provided dates actually reflecting the pattern on the previous day.

One of the cool things about doing this pattern matching is finding "hidden" analogs during the warm season. When the right synoptic pattern sets up - one that might cause a historic arctic blast during the winter - but doesn't produce any notable weather because it occurred at the wrong time of year. Here's a good example. A nice arctic pattern set up on May 24, 2017, with a large, deep trough well positioned to deliver major arctic air to the PNW. But... it was late May and Portland had a 66/49 day. You can definitely see that a colder airmass moved through by looking at daily PDX max/min: 21st: 83/52 22nd: 91/56 23rd: 84/52 24th: 66/49 25th: 75/49 26th: 83/52 27th: 89/56 Some of the best pattern matches to May 24, 2017 during the cold season since 1950 include: 3/3/1955 12/15/1964 1/9/1963 2/14/1956 11/26/1985

Yeah, I see how that might jump out. But January was the better match. June 2021 was extraordinary in many ways, one of which was the extreme amplification/cutoff blocking setup for a summer heat wave. Our biggest summer heat waves are traditionally open ridges that are centered further south. All of the best pattern matches I saw for June '21 were from non-summer months!

Yeah, that makes a lot of sense. BTW you might get a kick out of this - I ran a pattern match for the June 2021 heat dome and one of the best matches was 1/18/2009. It's pretty uncanny how similar those patterns were at the 500mb level.

No problem! Some of those were surprising, but others I kinda figured. Jan 13-15, 2020 is interesting. The 13th mostly matches to warm season patterns (best match is 10/24/1991 right before the big pre-Halloween blast), and the best DJF matches are 2/24/1956 and 2/23/2011. The 14th is all over the place - best match is 9/19/1988 followed by 11/8/2000. Best DJF matches are 12/18/1951 and 1/9/1950. The latter is obviously in a transition to what would become a major cold wave. Similar story for the 15th. Oddly enough, the best matches are all from the warm season (7/23/1953, 7/2/2002, 6/16/2001, 6/23/1981). The closest DJF matches are 2/28/1976 (heading into a major maritime blast), 1/23/1965, and 1/22/1954 (another epic onshore snow pattern). However, one of the best matches to 1/22/1954 is 12/2/1980 so you nailed that one, they are clearly very closely related.

No problem! That was fun to put together. I find the warm season matches really interesting. Hidden analogs so to speak, since they usually don't produce noteworthy low temperatures and go unnoticed in the records.

Some of you might enjoy this - I made a post about historical analogs to some of our recent cold waves:

So, I've been playing with ECMWF ERA5 reanalysis data for 1950-2021. I thought it would be fun to go through some recent cold waves and find their closest historical analogs. The criteria here is the spatial pattern of 500mb GPH between 40-60N, and 140-110W. The magnitude of the actual heights won't be the same especially if a similar pattern occurred in a different season. For each pattern, I will list the date which I am trying to match (in bold), followed by its closest historical matches (in order of best match): 12/27/2021 --> 1/2/1950, 2/13/1990, 2/22/2018, 11/21/2003, 5/23/2004 2/12/2021 --> 6/23/1993, 11/28/1985, 9/22/1955, 9/24/1986, 5/14/1986 10/25/2020 --> 12/29/1990, 4/19/1966, 11/24/2010, 12/12/1967, 3/26/1970 3/14/2020 --> 6/6/1980, 1/2/2017, 2/9/2019, 9/16/1954, 11/10/1985 2/26/2019 --> 4/22/2008, 11/19/1978, 12/31/1992, 10/6/1957, 1/9/2005 2/9/2019 --> 4/28/2009, 1/8/1993, 8/28/1951, 1/2/2017, 12/6/1972 1/10/2017 --> 7/29/2013, 2/13/2000, 11/11/1965, 12/13/2016, 8/10/2005 11/13/2014 --> 1/15/1957, 4/27/1958, 4/30/2009, 6/14/1969, 12/23/1983 2/6/2014 --> 6/27/1959, 5/24/2012, 5/11/2000, 5/3/1996, 3/4/2019 12/7/2013 --> 12/20/1998, 9/30/1950, 12/8/2009, 12/20/1990, 12/7/1972 1/17/2012 --> 1/3/1982, 4/1/2008, 11/9/1995, 11/4/1990, 10/26/1971 2/25/2011 --> 1/20/1962, 1/3/1974, 1/5/1993, 3/21/1954, 2/17/1993 11/23/2010 --> 1/14/1950, 10/27/1971, 12/27/2021, 1/3/1973, 1/2/1950 Some notes: -The patterns on 1/2/2017, 2/9/2019, and 3/14/2020 were all historically very similar, despite 72 years of patterns to choose from. -The November 2010 blast had two of its closest matches from January 1950! -February 12, 2021 was a weird one. 4/5 best matches were in the warm season, along with Nov. 1985 thrown in. -December 2013 was in legit company! Interesting that some of the biggest December blasts on record had very similar patterns. In fact, 4/5 best matches were from December despite having the entire calendar to choose from.

Oh interesting, I wasn't aware of those. Yeah not surprised about some of these DOT stations. I saw one of the Seattle newspapers also reported 119F at a couple stations in the Chelan Valley. Lots of candidates for the state record it seems.

Looks like NWS Spokane will investigate the 123F reading at Alpowa Creek HADS automated station in eastern WA last Tuesday (June 29th). My guess is that it's a bit overexposed, but it will be interesting to find out.

Yeah, that's the problem with Matlab. You have to be in academia, otherwise it's not really relevant (or accessible).

You nailed it, Matplotlib + Basemap for those maps. I ran the SOMs algorithm in MATLAB and then exported the data to Python for analysis and plotting. The time series chart was actually done in Seaborn. I would do everything in Python if it were up to me, but I have to jump to closed source platforms once in a while.

Hey, I'm using the 'somtoolbox' in MATLAB to create these. As far as I understand, this is still the canonical choice which most published papers have used for their SOMs analysis. There are also implementations in Python and R that users have contributed, but I don't think any of them are as stable or trusted as the MATLAB implementation.

That's a great observation about 1985. Our heat that summer had very different dynamics from the recent massive ridge-dominated summers. Also speaks to the fickle nature of trying to project how heat extremes might change along the immediate west coast in the future, places west of the Cascades and west of the Coast Range in California. So much depends on the right longwave patterns recurring, as opposed to actual warming in the means.