Variability—Elephants in the Room

Yesterday morning at 3:15 I awoke to rain, accompanied by occasional faint rumbles of thunder. My dream had ignored the thunder, but had not been able to reconcile the sound of the rain pelting down outside.
The PBS/NPR stations in Tucson spent the day reporting that, National climate experts are still waiting to see whether an El Niño weather pattern will develop. I don't have any idea who may be their source for that statement, but there are multiple problems with it. First of all, national climate experts have been expecting for at least a month now (see links in my previous post) greater chances for above normal precipitation in Tucson this winter, based mostly on the high chance of El Niño being present this winter. The Climate Prediction Center's outlooks are couched in terms of probability over a certain period of time (as are, or should be, almost all weather forecasts of precipitation). It is not a matter of waiting to see what happens. It's true that an official call on the actual presence of El Niño won't be made until a monthly update; the next monthly update will be on December 13. But El Niño being present (as defined by ocean temperatures) is not the same as a particular weather pattern being present. In every season, in every winter, in every El Niño winter there are a variety of weather patterns. One pattern gives way to another. Sometimes a pattern persists for weeks or even a month; sometimes a pattern lasts for less than a week. Over the course of a five-month winter there might be only a dozen or so pattern shifts. When it looks like an established pattern will never change to something else, it seems that's just when it does. Some patterns tend to return more frequently or last longer during an El Niño winter, and those patterns, averaged with the patterns that return less frequently or last for a shorter period of time in those years, result in what is a departure from an average winter. But every El Niño is different. There is no single El Niño weather pattern that develops and then lasts the rest of winter. As surely as some expert will eventually decree that such a pattern has arrived this winter, that decree will be followed by a shift to a dry pattern for a few weeks.
So every El Niño winter is different, and every non-El Niño winter is different. But intra-seasonal variability is not the whole story behind the scatter. The figure below simply repeats the data points plotted in the previous post. But now the color labels are different for the plotted points.

The entire period, beginning with the winter of 1949-1950, amounts to 69 years. In this figure, unlike the one in the previous post, for an approximate middle half of the period, the winters 1965-1966 through 1997-1998, amounting to 33 years, are plotted as green circles. There is a different distribution for the green years compared to the earlier and later periods. In 7 of the 33 green years the January ONI was greater than 1; 4 of those 7 were whopper winters in terms of precipitation. But also, of the 20 green winters that were not in El Niño territory in January, 4 of those winters were also whoppers. (I've always been skeptical of studies that purport to establish a lagged correlation between El Niño and rainfall. The lag works, until it doesn't. It is interesting to notice that 79 and 93 lagged 78 and 92 by one year, and 68 and 85 lagged 66 and 83 by two years. But most of the time a lag fails miserably, even in the green years.)
The end of the period of green years coincides with a shift from a warm to a cool phase of the Pacific Decadal Oscillation. But the start of the green years was near the middle of a cool phase of the PDO.
Although the El Niño winter three years ago was disappointing for Tucson, we've had a couple of winters this decade that nudged into green-year territory. It will be interesting to see what this winter may bring. We'll have to wait and see what will develop.

El Niño and Tucson Winters

The scatterplot below is an update of a similar figure presented in my post three years ago.

The Oceanic Niño Index (ONI) is defined and calculated by the Climate Prediction Center [CPC] as a climate-adjusted three-month average of the sea surface temperature anomaly for the 3.4 region of the tropical Pacific. Thus the January ONIs are already averages over the three months of December, January and February. Historical calculations now use version 5 of the CPC's reconstructed sea surface temperature dataset. The figure that I presented three years ago was from version 4. In some cases the plotted points on this figure have shifted a few tenths of a degree left or right from where they were three years ago. But the overall scatter remains the same. This updated figure includes the last three winters. The last two winters would be labeled 17 and 18, but those two points are packed (unremarkably) among similar years having negative ONI with below average precipitation. (For all 69 years, the mean of the five-month total precipitation is about 3.7 inches; the median is about 3.3 inches; the upper one-third starts at 4.1 inches.) The winter of 2015-2016 (labeled 16) is remarkable because though it is the rightmost point, it is also below average.

Being a three-month running average, the September ONI that will become available at the CPC site tomorrow will lag the latest developments. According to the weekly updates from the CPC the 3.4 anomaly has climbed rapidly in October so that for the latest week it stands at plus 1.1 degrees. The models had correctly predicted the rapid warming in October. If their prediction of a leveling off for the rest of the winter holds, then the January 2019 ONI would be close to plus one, which is the dividing line between a weak and a moderate El Niño. Accordingly the CPC's outlook for the upcoming winter (three months, December through February) issued two weeks ago expects an El Niño pattern across the country. Their outlook puts Tucson's odds into wetter-than-average territory, consistent with the plot above, where of the 24 winters in El Niño territory in January, 15 had above average precipitation, 14 of the 24 were above 4.1 inches. For the 13 weak El Niño winters alone, the two decades of the 1950's and the 2000's are tilted toward below average; the decades of the 1970's through the 1990's, along with so far the decade of the 2010's are tilted toward above average. The fact that the El Niño winter three years ago was below average is unlikely to have ushered in a new era of below average El Niño winters, not anymore than the winters of 1984-1985 or 1967-1968 ushered in new eras of above average La Niña winters. This coming winter's precipitation in Tucson might not be above average, but it more likely will be.

Fall Festival of Rain

Tucsonans don't complain when it rains. Rain makes the actual festivals more fun, more reason to celebrate while packed under the tents. Tucson so far this fall has escaped any serious flooding such as has affected points south, west and north of here.

In a post two months ago, I noted the Climate Prediction Center's odds for El Niño during the upcoming fall and winter, and revisited my rant from three years ago regarding chatter about fall rain in Tucson related to El Niño. Since this fall's rains have been so good, I decided to present a fresh scatterplot for fall. For more explanation about the reasoning behind the scatterplot and for additional sources, follow the links in the previous post.

The Oceanic Niño Index (ONI) is defined and calculated by the Climate Prediction Center [CPC], and won't be available for October until the end of November. For now I'm forecasting 2018's October ONI will be approximated by the latest (mid-October) weekly Niño 3.4 region SST departure. As for this year's fall (September plus October) rain, so far for the period from 1 September through the morning of 16 October the airport has had 3.44 inches; that running total will probably increase by the end of the month. (I've had 4.76 inches for the same period, including 0.84 overnight last night.)

Considering just the last six years (2013-2018), there is a trend toward higher precipitation totals when moving from colder to warmer ONI. The decade of the 2010's has joined the decade of the 1970's in favoring above normal fall precipitation when the ONI is in positive territory. The exceptional years (fall precip well below normal but positive ONI) for those two decades were 2012 and 1979. Fall of 1976 was close to normal.

Considering the wet years appearing in the negative territory of October ONI, with the exception of the decade of the 1950's, in each decade there is one unusually wet fall. These wet falls have followed El Niño fall/winters by one (1964, 1970 and 1983) two (1996 and 2011) or three (2000) years. Certainly it's possible that the climate system sometimes stores somewhere a lingering effect of a previous El Niño, maybe even for a few years, and that that lingering effect influences Tucson precipitation in a specific La Niña fall. But most often any lingering effect does not appear to have an influence. As a specific example, most recently the fall of 2016 does not appear to have been influenced by a lingering effect of the previous fall/winter's strong El Niño. It is as much nonsense to call the fall of 1983 an El Niño year as it would be to apply the same label to the fall of 2016.

Missing Middle of the Monsoon

Every year in Tucson people complain about the current monsoon—not like the ones they remember. That's an accurate perception; it isn't just a matter of recalling only the spectacular days. Each monsoon has a different variety. Some years feature squall lines that roll across the metro area between 3 and 5 pm, others chaotic distributions of evening storms, others organized nocturnal systems, others breaks when the monsoon completely shuts down—each of these more frequently in some years and less in others. So judging this year is not like the others is usually justified.

It's generally recognized and understood that during the shoulder seasons of the monsoon, early July and the weeks around Labor Day, rain totals can be erratic. Erratic is even better understood for the off-seasons, late June and mid-late September. Intrusions of tropical moisture if they happen in the off-seasons are appreciated; if not, oh well. But usually over a four week period near the middle of the monsoon, the last half of July and the first half of August, one way or another the idiosyncrasies of that year's monsoon generate a rain total that meets expectations. That did not happen this year. (My house is 12 miles north of the Tucson Airport.)

2018 Period	# days	My house, inches	Airport, inches
June 15 and 16	2	1.76	0.91
July 5 to 15	11	0.98	1.77
Jul 16 to Aug 13	29	1.01	1.15
Aug 14 to Aug 25	12	1.26	2.05

Until two weeks ago this year the usual quips about the non-soon had given way to bitter comebacks, What monsoon?!, or as someone on my community's listserve wrote sarcastically on August 11, Our trees/shrubs are just loving it!! G RRRR!

Since the current official definition of the monsoon includes June 15, as well as all of September, and since September is expected to be above normal, it's likely that the official rain total for this year's monsoon will look reasonable. But no matter how much rain arrives in September, it won't wash away the reality of a missing (less than 0.04 inch per day) middle of the monsoon this year.

Elusive Effects of El Niño

Seems only yesterday, but it's been three years since everyone was talking about expectations for the then upcoming 2015-2016 winter. That summer of 2015 the Oceanic Niño Index (ONI, defined and calculated by the Climate Prediction Center [CPC]) had already run past plus one, and was expected to (and did) exceed plus two during the winter of 2015-2016. This year the latest Diagnostic Discussion by the CPC places the odds for El Niño at 60% this fall, and 70% during the upcoming winter. The CPC's consensus of forecasts takes the ONI to around 0.6 by this October and to about 1.0 during the winter.

This time three years ago I thought that the chatter about the upcoming effects of El Niño was excessive. This summer on the contrary it seems that excessive reticence has replaced overconfidence. Three years ago in an initial rant about the chatter, I posted a scatterplot of the previous 65 years of precipitation at Tucson airport, totaled for September combined with October, versus October ONI. Here are the final numbers for the last three years—an additional three points that can now be added to that scatterplot for fall.^*

Year	October ONI	September plus October Rain, inches
2015	+2.4	4.65
2016	-0.7	1.70
2017	-0.7	0.03

I won't rehash my old rants about fall; they can be read from three years ago here, and two years ago here. The point about expectations for this fall is that to the extent that an increased level of concern about the kind of thing that can happen in the fall because El Niño was justified three years ago, that level of concern is equally justified this year.

There will be a separate post for winter, including a fresh scatterplot^* for that season, as well as wonderings about why the scatter for that season is what it is. But that post can wait for a month or two.

* The plots three years ago were made with the previous version (v4) of the CPC's ONI's SST anomalies. The current version (v5) mostly differs by no more than 0.1 ONI (degree C), but in some cases the difference is 0.2 or even 0.3. Notably: October 1983 is now -0.8 (previous -0.6); and January 2010, which previously at +1.3 was right in the middle of the "moderate" El Niño category, now at +1.5 edges into the strong category.

Joe DeNardo

The Pittsburgh Post-Gazette this morning has a story that broadcast meteorologist Joe DeNardo died. Though the story acknowledges DeNardo's time with KDKA in the 1960's, the memories in that story are mostly from the time after I left the area.
Here are some memories from the 1960's. Joe was a star on both the radio and television sides of KDKA. In the mid-60's KDKA radio was trying to blend traditional public service news/sports with top-40 music. KDKA's morning radio show Pallan and Trow I was surprised to read started only in November 1965 and ran through April 1968. Pallan and Trow were talented comedian entertainers, not just ordinary hosts of a morning news/top-40 time slot. There was a certain time when I was getting ready for school when Joe would call in with the forecast discussion. Sometimes he couldn't stop laughing at the antics of the hosts, and he would have to call back later. I was a fan of Joe as much for being a straight man for Pallan and Trow as for being a serious meteorologist.
I was slow to give up completely on KDKA's blend. In the spring of 1967 I was taught to tune-in and turn-on to Chicago's upstart WCFL, which could be picked up after dark by tuning KDKA and then adjusting the dial slightly from 1020 to 1000. WCFL's weather forecasts were as brief as possible. It might be 89 in the northwest suburbs, 93 in the southwest suburbs, and 65 downtown at the lake. And of course as a good rule of thumb Chicago's weather today would be Pittsburgh's weather tomorrow.
So I credit WCFL as much as Joe DeNardo for my interest in weather. But Joe had the serious technical details, especially on the Sunday 11 pm newscast, when he would provide an outlook for the week ahead. At mid-week he would slide across the forecast 500-mb map, copied with grease pen onto a large clear plastic overlay. "Enhanced lake-effect snow showers on Wednesday/Thursday as this upper-air trough passes." The rest of the outlook into the following Sunday clearly communicated that his was a plausible scenario extrapolating into what was then for computer forecasts terra incognita.

Category Labels vs. Distinct Things

Suppose that it would be interesting, useful, or even important to categorize hot days for cities in the Northeast. A day when the temperature at the airport reached 95F or higher could be labeled Category 5 heat. Better yet, label that day a Scorcher. The Glossary of Meteorology would define Scorcher. Climate studies might draw conclusions about the number of Scorchers. To the question there would be a ready answer, What is responsible for this heat? Meteorologists say it is caused by something called a Scorcher. A simple report, Hot yesterday, a bit hotter today could be expanded into an earnest elucidation, Yesterday was not a Scorcher, but today we have a Scorcher because ...

Too often in meteorology an attribute of a thing is cast into a newly labeled category of that thing, and then in no time at all the name of that category is communicated as a distinct, peculiar thing, even though the new thing is basically the original thing, and the familiar label for the original thing adequately, succinctly and more clearly communicates the weather situation.

The jargon bomb is a subcategory of intense area of low pressure. Bomb does not categorize intensity, but rather the time rate of change of the central pressure of the low. If the question is, What is responsible for this wind?, then part of the answer is the large, ineradicable cold area of high pressure to the north of the low. Perhaps a sexy label could be devised for the high's stubborn persistence; anti-bomb? But no need for that. Already there is a long-established, familiar term that most everyone in the Northeast recognizes as labeling the overall pattern of extreme winter weather.

There are many other local labels that succinctly communicate a familiar pattern of wind, temperature, humidity and/or sky condition; examples include Santa Ana, Bora, Blue Norther and, in Arizona, Monsoon. These locally familiar labels are effective precisely because they resist being defined precisely. Instead they are learned as a pattern—know it when you see it. Conversely, when a subcategory of a thing is defined by numerical threshold(s) of attribute(s) of the original thing, it has to be wondered what is intended when using that jargon term for communication since use of the unfamiliar term requires a redundant reiteration of the threshold(s).

The term derecho is like the term bomb in that both terms could be termed second-generation categorizations. Derecho does not categorize the magnitude of the severe winds accompanying a fast-moving line of thunderstorms, but rather the time during which the line remains severe, or equivalently the area swept out by the line during that time. Certainly it is important for forecasters to recognize and for both forecasters and the public not to dismiss an unusually long-lived line of severe thunderstorms. But why not communicate the attributes unusually long-lived for the thing and unusually high confidence for the long-lead-time warning? (What's worse, earnest elucidations about the term derecho these days invariably short-circuit the history, butchering the fact that when the Spanish word for right translates to English straight it is not as a direction but as not evasive, honest.)

Polar air is a thing. It is defined by the climate system, not by an arbitrary numeric threshold. For the same reason, polar front, polar jet stream and polar vortex are all things. Arctic air is a distinct thing, generated by absent sun during the arctic winter. The labels Polar and Arctic are arbitrary jargon, but they are long-standing and deserve to be respected and elucidated. Arctic air—whether stationary, displaced or on the move—is both Weather and Climate, and, because it is generated by minimal sun where absolute humidity is very low, it will always be with us, no matter how much manmade greenhouse gases continue to increase and no matter what schemes may be instituted to reflect a portion of incoming solar energy.

[I acknowledge that at least part of this rant was inspired by the Air Weather Service Back to Basics program of 40 years ago.]

Republican Disgrace

In a few days I'll be back to posting about weather. But first, a follow-up to my previous post (about Collusion). Yesterday Senator Dianne Feinstein released a pdf-file of testimony, putting discussion about the famous dossier (prepared by Christopher Steele, the guy who's father had been a weather forecaster) back on the table. I'll quote from Josh Marshall's summary at Talking Points Memo:

... this is a testament to what can happen when the GOP unites behind a campaign of willful disinformation at the country's expense ...

I've read the first 140 of the 312 pages of the released pdf-file, and I would say that so far the contrast in questioning by the committee's majority vs. minority staff is striking, if not disgraceful.