Fire, Lightning, and Humidity

Following up on my June 26 post about surprisingly low fire acreage so far this season, I took up reader Gary's suggestion to look at humidity data from Fairbanks.  The idea is that higher humidity this year may have produced "wet" thunderstorms that deliver significant rainfall, rather than "dry" thunderstorms that more efficiently spark fires.

But first, a correction: I realized the chart I showed last time was incorrect, because I had some missing dates; so it turns out that cumulative lightning activity statewide this year has actually been lower than normal.  It was only about 25% below normal on June 20, but since then there has been little lightning despite some notable warmth.  Here's the corrected chart - my apologies:

This greatly reduces the mystery of low fire acreage this year, but nevertheless it's interesting to see that humidity was indeed higher than normal in June when we look above the surface to the 700mb pressure level (about 10,000') in the Fairbanks sounding data.  In the following chart, the red and purple lines show the 700mb relative humidity and dewpoint respectively, and both were well above normal in June.  Higher humidity aloft helps thunderstorms gain strength and produce more rainfall, because it reduces the "entrainment" of dry air that tends to mix away the moist updrafts.

Surface-level humidity, on the other hand, was somewhat below normal in comparison to recent decades.  Interestingly, there's little historical relationship between June fire acreage and humidity at the surface, perhaps because dry conditions mean fewer thunderstorms and fewer fire starts.  However, June fire acreage is clearly related to 700mb humidity; compare the two charts below.

In conclusion, low lightning activity is clearly a major reason for the low fire acreage so far this year (still well under 100,000 acres statewide), but above-normal humidity in the middle atmosphere may also have played a role by enhancing rainfall and minimizing so-called dry lightning.

Big Rain in the Northwest

A lot of rain has fallen above the Arctic Circle in northwestern Alaska this week, courtesy of strong westerly flow circulating around a deep upper-level low to the north of Alaska.  The high-quality CRN site near the Red Dog mine (940' elevation) reported a remarkable 4.0 inches of rain in the last 3 days of June:

NWS estimates (see below) suggest that more than 6 inches may have fallen in some spots, but I'm not sure if any ground-truth measurements came in that high.  Here's one list that the NWS provided:

Regardless of the highest amounts, this was a very significant rain event over a large area, including the interior North Slope.  Click to enlarge:

The 500mb analysis from Tuesday afternoon (see below) shows the strong Arctic low, and notice the contrast with the extremely strong ridge over southwestern Canada.  All-time heat records were shattered in Oregon, Washington, and British Columbia, including an outrageous 121°F in southern BC on Tuesday (by far Canada's highest observed temperature of all time).  There was certainly a dynamical connection between the two features - the ridge and the Arctic trough - and the large amplitude of the overall flow anomaly helps explain the extreme rain outcome in northwestern Alaska.

Low Fire Acreage So Far

The big story in the climate world this weekend is the outrageous heat wave currently affecting British Columbia and the Pacific Northwest of the lower 48 - and also far southeastern Alaska - but for today I'll highlight a bit of good news: Alaska fire acreage so far this season is surprisingly low.  As noted by Rick Thoman today, acres burned so far amount to only 60,000 - which is the lowest for the date since the exceptionally quiet year of 2008, when only 63,000 acres burned in the entire season.

Typically one-third of the season's fire acreage has accrued by this date, so it's becoming somewhat unlikely that we'll end up with an active season; there's a pretty good correlation between June 26 acreage and final acreage.  But let's not be hasty: fire activity can really take off in July, so the verdict isn't in yet.

I said the acreage is "surprisingly low" because lightning activity has been almost exactly normal so far, and temperatures have been above normal much of the time since spring, with a couple of notable warm spells this month (see below).  [Update July 6: this chart is not quite correct, see July 6 post for an update.]  It also hasn't been particularly wet, although my impression is that the thunderstorms that have occurred have been significant rain makers in localized areas.  So perhaps there has been less dry lightning than normal, and hence a reduced number of fire ignitions.

Arctic Temperature Trends

Following up on the topic of Arctic-wide temperatures (see this post from a couple of weeks ago), I've started working on some new graphics to visualize the ongoing ebb and flow of temperature anomalies across the 32 longstanding climate observing sites that I mentioned before.  But first, here's another view of monthly anomalies since 2010, but this time presented in terms of standard deviations; so this puts all seasons on a level playing field and eliminates the spike in temperature variance during winter.

The message is the same: Arctic-wide temperatures saw a jump up in winter 2015-16 and have remained generally elevated.  Since 2016, there have only been a handful of months in which any of the sites was 2 standard deviations below the 1991-2020 normal (last February at Bettles was one of them), but more often than not at least one site has been 2-3 standard deviations warmer than normal.  The community of Resolute in northern Canada saw a +3.5 SD anomaly in January last winter (it was the warmest January on record).

Here's a look at the 1980-2020 linear trend by calendar month, for the 32-site average temperature.  The absolute value of the trend is smallest in summer (a "mere" 0.4 °C/decade), but the statistical significance of the summer warming is on par with the rest of the year, and considerably exceeds the late winter minimum in significance.  All in all, the trend is most pronounced in late autumn, as we well know from discussing Utqiaġvik over the years (e.g. here).

Now on to a map view of monthly anomalies so far this year: below are pairs of maps for each month, showing departures from normal in both absolute and standardized terms.

January:

February:

March:

April:

May:

As noted in the previous post, February and March were slightly colder than (the 1991-2020) normal overall, but this was a mere blip in the context of the sustained warmth of recent years, as the top chart above shows.  Two Canadian sites - Resolute and Pond Inlet - had their warmest January on record, and the widespread warmth in Siberia in May echoed the persistent, excessive warmth of last year (see here for comments on 2020 in Siberia).

This month will also end up much warmer than normal in most of northern Siberia, because there's a major heat wave under way at present.  Here's a chart for Suhana at 118°E longitude, showing daily mean temperatures in the last 60 days as well as a forecast from today's ECMWF ensemble.  Temperatures are peaking at well over 10°C above normal for the time of year; today's high temperature was 32.7°C or 91°F.

For the 32 sites collectively, we see that temperatures have been running about 1-2°C above normal since late May.  The forecast suggests a respite towards the end of the month, with temperatures dropping back to near or just below normal Arctic-wide; we'll see if this pans out.

Summer Heat

The heat is on for the eastern half of the interior today, with widespread 80s, including 83°F so far in Fairbanks and 84°F in Eagle.  Many RAWS site are hotter, but that's typical under strong summer sunshine.  More notable is the heat across the eastern North Slope: 72°F at Barter Island, and 80°F at Deadhorse on a southwest wind before the sea breeze kicked in.  Click to enlarge:

The 80°F at Deadhorse is the first time in nearly 3 years that the temperature has exceeded 73°F.  The calendar years of 2019 and 2020 had maximum temperatures of 73°F and 71°F respectively, which is actually the coolest pair of summers (by this metric) since 1987 and 1988.  But all four years from 2015-2018 saw at least one 80°F day in Deadhorse.  Here's a visual representation of all 75+ and 80+ days since 1970 in the combined history from Deadhorse and (prior to 1999) Prudhoe Bay.

It's interesting to note the occurrence of warm days in August - remarkably late in the summer for the Arctic coast - between 1989 and 2007.  Also, the trend towards earlier hot days seems notable: prior to 1990, the earliest 75°F was July 3, and today was the earliest 80°F on record.

In other (but related) news, Fairbanks airport has been reporting distant wildfire smoke this afternoon: presumably the result of the Haystack fire 20 miles north of town.  This is the first smoke report at the airport since June 1st of last summer, and that was the only day last year that reported smoke.

Stubborn Northwest Chill

It's been a cold start to climatological summer in the Bering Strait region, with a remarkably stubborn upper-level low residing over the area for an entire week.  Cut-off low pressure features like this are often persistent, but it's unusual to see almost no movement for such a long time.  Here's a simple animation of 500mb height analyses, courtesy of Environment Canada, at 3am each day from June 1-10.

With cloud and cold air stuck overhead, Nome has seen the coldest first 10 days of June since 2006.  Another statistic: the week ending June 8 was the third coldest week entirely within June since 1977; but 2006 and 2013 both had slightly colder weeks in early June.  Like this year, 2006 and 2013 both had major disruptions to the stratospheric vortex in January; there may or may not be a connection (perhaps more likely not), but I'm always on the lookout for these things.

Some fresh snow has been seen at various locations around the Seward Peninsula: here's a photo from Shishmaref on Tuesday (June 8), courtesy of Alaska DOT&PF on Twitter.

And here's a video of snow in Nome on June 5.

@AlaskaWx
Current weather Nome,Ak pic.twitter.com/nGKk28eswC

— Hawkeye 4 Life (@mdiaak54) June 5, 2021

As for the larger context of Nome's temperatures so far this year: it was a chilly end to winter, but spring has been considerably warmer than normal most of the time since mid-April.

 

Arctic-Wide Temperatures

Several years ago I spent some time looking at Arctic-wide air temperature trends based on measurements at climate observing sites around the basin, and it's probably worth returning to that idea for an update.  See here (and links therein) for previous work.  In my fresh look at this, I found 32 sites - distributed fairly evenly around the Arctic land areas - that have nearly complete data since 1980 and also report reliably on a daily and monthly basis.  Here's a map of the sites:

This set of locations is designed to sample both inland and coastal environments above the Arctic Circle (neglecting interior Greenland); previously I was focused on coastal sites clustered around the central Arctic Ocean.

The chart below presents a multi-decadal view of monthly mean temperatures at these locations; the black line shows the 32-station average of the monthly departure from normal, with "normal" being defined as 1991-2020.  The red and blue lines show the highest and lowest monthly anomalies within the set (with the largest anomalies occurring in winter, when variance is highest).  Click to enlarge:

The overall warming trend amounts to +0.76°C per decade (using linear regression), or a little over +3°C in the 41-year period.  This is much more warming than has occurred across the globe as a whole, illustrating the reality of "Arctic amplification".  I should note, however, that this analysis is based on the unadjusted ("unhomogenized") data from GHCN-M, so it's the "raw" data without corrections for changes in station siting, instrumentation, or observing practices; read more here.

Looking at the last decade or so, it's remarkable to see the jump up that occurred in winter 2015-2016, when the super El Niño occurred; temperatures seem to have remained at a higher baseline since then.  Since 2016, it has become fairly uncommon to have any station with a -5°C anomaly (even relative to the warmer 1991-2020 average), whereas most months in the past few years have had at least one site with a +5°C anomaly.

Interestingly, November of last year was the second warmest month in this period of record, with a +3.7°C average anomaly for the 32 sites.  Only January 2016 was warmer, at +4.3°C.  Again, this is relative to the new 1991-2020 baseline.

However, there was quite a drop-off in Arctic-wide temperatures as last winter progressed, and both February and March were colder than the new normal.  Here's a chart of the daily progression (with a few missing days just before the turn of the year).  For a little over two months in late winter, the Arctic was generally colder than normal, but it warmed up considerably in April, reaching 4.7°C above normal by late April.  The year-to-date temperature anomaly so far (January through May) is +0.47°C.

Now that I've identified this more comprehensive set of stations, I'll be able to slice and dice the historical data in various ways, and keep on top of Arctic temperature variations going forward.  If readers have any suggestions for interesting analysis, feel free to leave a comment.

Winter Lives

The calendar may say June, but the weather has been more than a little wintry in the past few days at Howard Pass in the western Brooks Range.  Readers of this blog are well acquainted with the extreme wind chills that occur regularly up there in the winter, but it turns out that the site can put on a good show even in "summer".

As the chart below illustrates, two multi-day episodes of wind chill below +10°F have occurred in the past few days, and early yesterday morning the wind chill dropped below 0°F for a few hours.  This is the first time a sub-zero wind chill has been measured at the site in the summer months; previously the latest was on May 19, 2013, and the earliest was on September 9, 2012.  (However, there are only 5 complete June's in the station's history since 2012, because the anemometer had been disabled by winter winds and was not yet repaired in June 2015, 2018, and 2020).

An interesting difference from the winter wind chill setup is that the cold air blowing through Howard Pass this week is not being drawn from a locally reinforced cold pool on the North Slope.  In the cold season the extreme wind chill occurs when severe cold develops across the interior North Slope under a surface-based inversion, and then a strong pressure gradient forces that cold air up and over the shallow pass.  But in this case it's just an early summer Arctic air mass; there's no serious radiative cooling at the surface in the continuous daylight of Arctic summer.

To document the nature of the air mass, here are the surface observations at 5am yesterday, and the 3am Utqiaġvik sounding (click to enlarge).

Wild Temperatures

The weather of the past week in mainland Alaska is a tale of a remarkable ridge of high pressure that intensified and migrated northward across the west coast and up into the Arctic.  At its onset the northward surge of warm air brought very warm conditions to southern and western Alaska on Tuesday, including 71°F in Anchorage, 78°F in Talkeetna, 75°F in Bethel (tied for 2nd earliest 75°F on record), and 80°F at Sleetmute on the Kuskokwim River.  The high-quality CRN site at Port Alsworth on Lake Clark reached 77°F, after a morning low temperature of 28°F; that's the largest daily temperature range outside of winter at that site since the sudden heat wave of late May 2013.

The heat extended farther north too: Nome reached 71°F on Tuesday, and as in Bethel, this was tied for second earliest on record for such warmth (72°F occurred in May 22, 2002).

Below are 500mb height charts from 3am on Monday, Tuesday, and Wednesday mornings: the ridge ballooned up into the Arctic by Wednesday, with a remarkable peak geopotential height of 5700m.  This is the highest on record for the month of May over the Pacific side of the Arctic Ocean (based on ERA5 data since 1950).

With a strong anticyclonic circulation taking shape around the ridge as it became cut off from the mid-latitude jet stream, the flow turned to the north over central and eastern Alaska - and the temperature started to drop.  Arctic air began to intrude, and on Friday an upper-level low pressure system dropped in from the northeast, forced down to the south and west by the flow around the ridge.  Here are Thursday, Friday, and Saturday mornings:

And so by Friday morning wintry conditions had returned to northeastern Alaska: Toolik Lake dropped to 17°F with fresh snow.  Freezes occurred widely in the eastern interior; and to cap it all off, snow fell yesterday in parts of the eastern interior and at low elevations in the Alaska Range.  Eagle reported an inch of snow - the latest on record for the town, according to Rick Thoman.  Here are low temperatures ending at 8am Saturday morning: click to enlarge.

The sequence of events is a bit reminiscent of what happened back in April, when a record-breaking cold air mass dropped south in response to a strong ridge to the west (but over the Aleutians in that case).  The ebb and flow of these circulation features has been quite striking for some months now, with notable cold outbreaks causing problems in the middle latitudes (Texas in February, spring freezes in Europe and the eastern US); but I haven't seen any evidence of an increase in volatility or extremes on a hemispheric scale.

Here's a simple animation of the 500mb heights at 12-hourly intervals from last Sunday afternoon through yesterday afternoon.  A fascinating sequence for meteorologists!

Late Snow in Bettles

I was too busy to post on this at the time, but last Wednesday a very late snowfall occurred at Bettles, on the south side of the Brooks Range.  Rain turned to snow in the early morning hours, and even though the temperature remained above freezing, 2.4" accumulated by late morning before melting later in the day.  This is actually the largest snow event on record for so late in the season, with data back to 1951 (and some data back to 1944).  The only other snowfalls of 1.5" or more after May 15 were May 19-20, 1996 (1.6") and May 22, 1956 (2.0").

Say it isn't snow?! Moderate to heavy snow fell at Bettles this morning. Around 1.5 inches fell by 10 AM, however, it is melting due to temperatures above freezing. We are expecting this system to move northeast and continue to bring snow, especially for higher elevations. #AKwx pic.twitter.com/QHTV2OsTUJ

— NWS Fairbanks (@NWSFairbanks) May 19, 2021

 

Here's a visual depiction of all 1" or greater calendar-day snowfalls in May through September since 1951, for both Bettles and Fairbanks.  Click to enlarge:

A couple of comments are in order.  First, that August 9 snowfall in Bettles is real: 2.6" fell in the middle of the day on August 9, 1969, and 2" was still on the ground the next morning.  The next day 3" was reported far to the southeast in Northway, so obviously it was an extraordinary cold snap for the time of year.

But the most striking feature of the chart is the increase in mid-September snows at Bettles in recent decades.  The change is remarkable: in the 40 years from 1951-1990, only 2 years saw a 1" snowfall on or before September 20, but in the latest 30-year climate "normal" there are 9 such years.  Digging into the reasons for this would make for an interesting study.

Warm Season Precip Normals

I mentioned in the last post that July retains the top spot for monthly mean precipitation in Fairbanks, according to the new 1991-2020 climate normals, and that historically this was not true for most of Fairbanks' modern climate history; August used to be wetter than July.  A July rainfall peak also differs from most of the state, as August is considerably wetter than July in all but the eastern interior and the North Slope.

To look at the shifting climate normal in more detail, here's a chart showing the overlapping 30-year normals since 1931-1960 for the warm season months.  A spate of wet July's from 2001-2010 brought a significant increase to the July average in the 1981-2010 normal, but the other months were largely unchanged until more recent years.  However, the 1991-2020 normal is wetter in each of June through September (but interestingly there's a slight downward trend in May).

 

Here's a time series view of the June-September totals (blue columns) and the fraction that is attributable to July (gray columns).

 

The change in July is largely attributable to a handful of very wet months, but the consistency of the high seasonal totals since 2014 is quite remarkable.  Here are a few blog posts from last summer about this:

https://ak-wx.blogspot.com/2020/07/wet-summers.html

https://ak-wx.blogspot.com/2020/07/rainfall-trends.html

https://ak-wx.blogspot.com/2020/08/back-to-rain.html

Another interesting aspect of the Fairbanks warm season climate history is that the early decades were also relatively wet, but the excess was seen mostly in August back then; August used to be distinctly wetter than July.  But this change isn't just about a shift in rainfall timing from one month to another; it also reflects a major change in the character of summer rain events.

Consider the chart below, showing the total precipitation in categories of daily totals, i.e. how much of the long-term normal is attributable to light versus heavy amounts.  The first two decades in this analysis (1931-1950) saw a higher fraction of precipitation from relatively light daily amounts, but over time there's been an upward trend in the contribution of high daily amounts.

The change in the distribution of daily amounts is connected to the July-August shift, because July rainfall is typically much more convective (brief, heavier events), whereas August rains tend to be stratiform in nature (lighter, more extended rains).  In recent years Fairbanks has seen a decreased frequency of August rains, but very heavy rains in July - while still quite rare - have begun to occur frequently enough to make a significant difference in the long-term rainfall totals.

For more on this topic, see this post from a few years ago:

https://ak-wx.blogspot.com/2017/07/wetter-in-july.html