raindancewx

Will be interesting to see how some of my pet theories play out in March. Nino 1.2 & Nino 3.4 conditions can be re-created for Nov-Jan using two sets of analogs (will change if Nino 1.2 or 3.4 warms rapidly this week). Major Hurricanes Hitting TX Analogs: 1967, 1970, 1980, 1999. OR Low Solar Analogs: 1934, 1934, 1934, 1964, 2007, 2007. First blend and second blend are each at -0.80C and -1.10C in Nino 3.4 and Nino 1.2 respectively, against 1951-2000 data according to this data, which is what NDJ looks like to me. https://www.esrl.noaa.gov/psd/gcos_wgsp/Timeseries/Data/nino12.long.anom.data (change 12 to 34 for Nino 3.4) First blend has high solar activity (opposite to now), but features years after a major hurricane (correct) hit Texas. Second blend has low solar activity (correct), but features years without a major hurricane (opposite to now) hitting Texas. I couldn't re-create the NDJ conditions in Nino 3.4 and Nino 1.2 with low-solar condition AND a major hurricane hitting Texas. The first set of years is pretty wet in March, second set is quite dry. So, I went to tie breakers - The first set of years is very wet in August+October, a good indicator of a wet March - we were dry those months this year. The second set of years is dry in those years. However, August + Sept 27-Oct 26 was very wet in 2017 (2.2" Sept 27-30 here) like the major hurricane years, but the non-major hurricane years were very dry in that period.

Looking through the MJO forecasts and comparing them with history, not a lot of January MJO propagation like this year, when factoring in phase and magnitude. I like January 1986 best, easily the closest propagation with 2-3-4-5-6-7 observed, we began in two this year. Doesn't appear we'll have any neutral MJO conditions in January. The magnitude in 1986 was greater than this year though. January 1978, 1989, 1990 are decent. I'd go 1986 x7, the others x1 each. 1988/1989 start in 1 and 3, but blended together, they start in two at lower magnitude than 1986. 1977 started in neutral, so its behind. Timing is closest to 1989/1986, but the MJO has been moving through the phases about a week later overall. Black line is January observed MJO conditions. Red line that I drew is the Euro forecast for the rest of January 2018. February 1986 is one of my all-time snowy months, and very wet too - would be nice to get it. City had 10.2" snow I believe, near the end of the month. Here is December 2017, v. the MJO blend I outlined above, plus one week.

The 3 km-NAM is honestly awesome at short range - it had essentially no rain/snow for the city and that seems to be verifying other than 20 minutes of trace-level rain. Euro and GFS both had much more, with some snow. Will be impressed if we end up with even 0.01" from this event in the city. On to the next storm I suppose. The non-Nino winters here after major hurricanes hitting TX lean heavily toward dry January, and we're ~2/3 through with 0.03". Not sure why the local NWS kept the weather advisory for the city for this event, it was 52F at 10 pm when they renewed it. My rule is <50F on a warm day by 10 pm if it is to convert to snow over night.

The GFS, Euro, 3-km NAM show a fairly healthy line of showers associated with a Pacific cold front moving through NM, bringing rain and then maybe snow even to valley locations. Going to have to be one incredible cold front if we go from 55-60F on Saturday to snow by Saturday Night, although it does happen. GFS thinks 0.10-0.20", Euro and 3-km NAM think 0.05" or so. I'd expect 0.05"-0.10" precipitation, with up to the final third of it as snow...0.1"-0.5" in the city?

^^ A lot of time they will ignore the lowest low or highest high if it only lasts for five minutes. Don't know if it is rounding, an error, or some kind of calibration.

MJO is now coherently just barely in phase four, many models have it in phase five in a week or less. Snap to a much warmer pattern for the US east of the Rockies should be coming. Will be interesting to see how much of this gets wiped out - Areas of North Dakota were +2 to +4 in December, so a wipe out of the (relatively) minor cold anomalies in January there would put them near normal, not cold for winter. Suspect the blue and purple areas on the map have a colder than normal month, I think the green areas end up near average though. Those reds in the West should trend down too. Would expect most areas to be +8 to -8 by the end of the month instead of +12 to -12. Even here, largely without precipitation, it looks much colder than December now. West Texas is fairly cold in phase five, so don't think the cold is particularly threatened in Texas.

I'll repeat my earlier complaint - its amazing that Monterrey, Brownsville, Houston, New Orleans, Savannah, Charleston, San Antonio and other cites have had snow this winter while it has barely rained here since September. The good news is...it is very cold today. So that's nice. I will say: One of my analogs, 1943, had multiple snow events into the deep South and TX, like down to the Gulf Coast, so that is actually verifying pretty well. You had a super warm Atlantic, neutral PDO, low solar activity, and an east-based "cold Neutral" look to ENSO that year. 1943 is like a colder version of 2008 nationally. Even the warm patch of waters north of the cold Nino 3/1.2 existed in 1943-44.

I have a theory that certain events repeat at relatively equal levels of sunlight, so been starting to look for some kind of "Harvey" effect in Spring 2018, since it hit in late August. To me, there are 59-non El Ninos for the 1931-32 to 2016-17 July-June years, if you separate out the ten non-El Nino years in that time-frame when TX was hit by a major hurricane from the remaining 49 years, one of the effects that pops out is a higher than normal chance of a very wet March in the Southwest: >=0.8" in non-El Nino, MH years in ABQ: 3/10 Marches >=0.8" in non-El Nino, other years in ABQ: 4/49 Marches You can say pretty safely that a "very wet" March is more likely in a non-El Nino after a Major Hurricane hits Texas. If you did it for >=1", its 2/10 v. 1/49 - quite literally 10x more likely than usual. https://mathcracker.com/z-test-for-two-proportions.php#results

It's kind of amusing see TX as one of the big winners for precipitation this winter - we're not even remotely close to Nina climatology at this point (40/90 days in). I know here, I should have ~0.63" for Dec 1-Jan 11 using 1951-2010, probably more using 1981-2010, so I'd say -0.5mm/day is about right here. The wet NW idea has absolutely not verified this winter, outside Montana.

My horrible dry streak was ended this morning - we got 0.03" of an inch of rain. First measurable precipitation since October 5, 2017, before we even knew who had a shot at the World Series or that there would be new tax policy for everyone. Flagstaff's snow-free streak ended last night as well, and they got ~1/6 of their winter precipitation in about six hours, now at 50% of normal precip out there.

I was playing with the observed SSTs in December for Nino 3.4 and Nino 1.2. Fairly strong correlations between Dec & the following March-May periods. I normalized the March-May periods against 1951-2010 means. The conversion factors from December SSTs to MAM work out to -0.27C in Nino 3.4 for MAM, and -0.37C for Nino 1.2 in MAM. That configuration can be replicated with SSTs for MAM 1961, 1981, 1981, 1984, 2006. The configuration is based on correlations, which means they'll probably be off somewhat. To offset that, I ran 13 cases, where Nino 3.4 / Nino 1.2 come in near but not exactly at -0.27C and -0.37C. I then looked at subsurface conditions, and weighted each of the 13 scenarios at 1,2,5,10,15,25 points. I re-produced all 13 scenarios using historical years with similar conditions, and then looked at the years that appeared most frequently in the higher weighted scenarios. Given the subsurface looks like this, I assumed Nino 1.2 was more likely to come in colder than the correlation implies, while Nino 3.4 was more likely to come in warmer than the correlation implies. The scenarios that had Nino 1.2 warming massively relative to the correlation were thus weighted less, as were the scenarios that had Nino 3.4 cooling massively relative to the correlation. Cases MAM Risks Points Scenario 3.4 1.2 % 1 -0.67 -0.37 5% 2 -0.47 -0.37 5% 3 -0.27 -0.37 25% 4 -0.07 -0.37 15% 5 0.13 -0.37 10% 6 -0.27 -0.77 10% 7 -0.27 -0.57 15% 8 -0.27 -0.17 5% 9 -0.27 0.03 5% 10 -0.67 -0.77 1% 11 -0.47 -0.57 2% 12 -0.07 -0.17 1% 13 0.13 0.03 1% Using those weights, I then did one final filter, adding three points to low solar years, and subtracting three points from high solar years for March-May to help sort out ties for weighting. A fair number of El Nino winters that were rapidly transitioning to La Ninas came up, will be interesting to see if we have the reverse of that. Anyway, the top 12 Springs were: 2006 (x4), 1984 (x3), 1981 (x3), 1961 (x2), 1977 (x1), 1963 (x1), 1986 (x1), 1978 (x1), 1988 (x1), 1960 (x1), 1973 (x1), 2003 (x1) That blend is relatively close to what has happened in December, especially if you use Dec 8 - Jan 6 as December. The long and the short of this is, the weighted blend is pretty wet for me in March, which hasn't happened since 2007. With my weights, solar factors, etc, the ONI would be -0.19C for Nino 3.4 in MAM, and -0.46C in MAM in Nino 1.2 Given the subsurface conditions, and the match to Dec 8 - Jan 6, plus my various simulations, fairly confident the map will look like below in MAM.

The period for mid-Dec to mid-Jan will obviously be colder for the US, but for December itself, not really a cold month nationally. With a couple of modifications, December 1980 was a good match. I threw in 1943 & 2011 to make the Upper Midwest warmer, and 2012 to make the SE/NE warmer. You have "less" of a break in the heat in my analogs than this year in a couple spots (NW, SE, TX) but its close nonetheless.

The European, NAM, and GFS all have valley rain and mountain snow for NM this week. Tues/Weds cold see the end to the absurdly long stretch of days without precipitation. Would be a good snow event if the European verifies.

The dewpoint has skyrocketed, all the way back to 30F...from 3 this time last night...but not looking like we'll get any measurable rain overnight. Close though, only missed by 20 miles or so. Of course...its 48 at 3 am as a result. Will be up to the next storm to break the dry-streak sometime Tuesday night into Wednesday.

Would imagine you guys get a big snow event between the big cold snap now and the big thaw coming later in the month. I haven't seen precip since October 5th, but we've got two shots at it the next six days if you believe the European, if those systems verify with measurable precip here, it will be because they are stronger/further south than any systems in the past three months and some different areas downwind of the SW should get some snow / rain too.

GFS and Euro have a relatively similar solution for a storm around the 11th, maybe 0.2" to 0.5" for the city. Snow levels...probably above the city for most to all of the event but good for the mountains. Will change though. Hopefully the models still have it tonight. Its nice to see the GFS, which had nothing for the period, cave to the Euro which has had this for two-three days now.

I'm getting really annoyed with this pattern - Monterrey, Brownsville, Houston, New Orleans, Tallahassee, Savannah have all seen snow before me this winter...and we haven't had any precip since Oct 5. It is neat to have a fairly cold winter for lows for once, that hasn't happened in a while, but I really hope the big storm the Euro shows next week verifies or we're going to have some major water issues here in Spring and Summer. I had 2005/2008 as analogs, and it was snowy in the South in 2008, and it was very dry here in 2005-06, but really was expecting other looks to the pattern which haven't shown up yet. Oh well, still time yet.

The Canadian update is showing one of the biggest month/month trends toward wetness I've ever seen it do in the SW for February. Let's hope it shows it again next month. Might be reacting to the Nina falling apart in Nino 1.2, or the SOI collapse in December?

New Canadian has trended colder for January in the majority of the country. Lot wetter for this area of the country too.

Did an experimental blend of the SOI drop (8-16 in Nov to <0 in Dec), low precip in Oct-Dec in ABQ, no precip in Dec in ABQ (pretty rare actually), and the US temperature anomaly profile for December to produce January. This is what I got - its a blend of 17 years, with 1951 (nada Dec, dry Fall), 1955 (US temp pattern Dec), 1996 (dry Fall, US temp pattern) weighted twice. Nada Decembers are rare since 1931 here, only 1950, 1956, 1963, 1981, 1996 - so for recency and comparable rarity that is given the same weight as the dry Falls.

Closing in on three full months without precipitation here. I was pretty worried about my analog blend for the month, but anomalies went toward the analogs over time. On the 20th, when the Northern Plains were +15, my "near average" idea for ND/MT/MN near Canada looked terrible, but since the 20th, an incredible correction began toward the analogs, in both the Northern Plains & Southeast where the signs were wrong v. what I forecast through 20 days. Will be curious to see how the final anomalies look

I'm rooting hard for the ongoing SOI crash. Currently down to -3.1 for Dec 1-27. Over +10 in Nov and Oct. Think it will may end up lower than that by month end. It isn't correlated super strongly with January precipitation, but it is fairly strongly correlated with January-March. Current reading (-3.1) would be 1.32", +/-1.25" at 95% certainty using the Dec SOI readings for 1931-2016 against Jan-Mar 1932-2017.

Anyone want to guess when we'll get some weather down here? Last measurable precipitation was October 5th - longest dry streak here since 1956. European ensembles seem to be hinting at January 4th or so. It's been neat watching the cold return to the Northern Plains - the high in Bismarck for Dec 1-20 was 42.5F, but for Dec 1-26 its down to 35.4F, cold is slaughtering huge amounts of the heat, but suspect the Northern Plains end up at +4 to +6 anyway, the mean December high for Bismarck for 1951-2010 is 26.1F, so 30/31F is still pretty warm. Obviously the deeper into the month you get, the harder it is for daily anomalous heat/cold to change the monthly anomalies. We're going to end up at 52F/53F with no meaningful snow for the mountains or valleys since October in NM, going to be a horrible May-June for fires without a huge turn around. The huge heat in the Northern Plains / AZ/CA with a cold NE is often a strong wet signal down here for March, which would be nice. Hasn't been wet (+20%) in March against the means sinc 2007.

I was playing with Nino 1.2 temperature correlations earlier today - its interesting that it isn't working this December. Long-term trend, and pretty strong at that - is for the Northern Plains to be cold in December if Nino 1.2 is cold in November. It's trending colder there now at the end of the month, but Bismarck looks like it will be +5 or so by the end of the month nonetheless. The areas that are coldest in Mexico this month should also be warmest - pretty strong trend. The yellows/blues are pretty strong relationships for ~68 year period. The super hot November, centered on NM/CO makes a lot of sense looking at the composites, cold Nino 1.2 Oct = hot November for that area for 1948-2016. I was pleasantly surprised to see NM is near the 0 mark for February's correlation to March, it seems like cold Nino 1.2 has the strongest temperature impact here in November, then it fades to no impact by March, before gaining in power again.

This is just for fun...but all five Decembers here without rain/snow at at were followed by cold Aprils here, and the composite is cold nationally, literally no one above the +1F threshold for April highs against 1951-2010. I'm sure something will get screwed up, but its interesting nonetheless.