Convective Atmosphere Group

Author: migel (Page 1 of 7)

Forecast Uncertainty

I’ve spent the last few months in Sweden.  The local official forecast entity here is the Swedish Meteorological and Hydrological Institute (SMHI).  They provide local forecasts with a dash of uncertainty thrown in.  So, in the picture below, you can see that today’s forecast is “fairly certain” while Friday’s forecast is “uncertain”.  That’s pretty cool, although even to me, it’s not entirely clear what those assessments mean despite looking around SMHI’s website for definitions.

What I think it even more fun, is a product they have rolled out in a beta form.  Below, you can see that they provide different possible forecasts.  So, rather than just providing the most likely forecast (what we might traditionally think of as the best forecast), they provide a range of reasonable forecasts.  It’s probably going to be partly cloudy today, but hey, it could be mostly cloudy, or possibly cloudy.  Regardless, it’ll probably be ~7C.  Tomorrow is more interesting.  It will most likely rain, but the second most likely forecast is sunny.  Without this kind of categorical forecast, we might be tempted to split the difference and provide some kind of intermediate precipitation probability and cloud cover.  But doing so risks misrepresenting the forecast.  Maybe the possible futures are 1) 100% cloud cover and extensive rain or 2) clear skies.  Providing a forecast that weights by the probability of those outcomes might be A) 75% cloud cover and intermittent rain.  But that’s not really a likely future because it’s not 1) or 2) so it’s wrong from the star

It’s a Fun Job

Warning: this post has no real point!

We’re almost done with another quarter, and it’s been one I never anticipated.  I’ve been here in Davis since 2016.  In that time, I’ve done a lot of things I never expected to do in this adjunct position.  I’ve given a lot of interesting media interviews.  I’ve gotten to go all sorts of interesting places.  I’ve gotten to work with lots of interesting and clever students.  And, I have taught a lot more than I ever expected.  It’s a fun job.  But, it’s a job that doesn’t come with enough hours in any given week to do everything that needs to be done.  Teaching essentially full time this quarter has been good in some senses because I can compartmentalize.  I hope students have learned something useful and that I haven’t led them too far astray.  If nothing else, I’ve had fun.  If I owe you an email response, it’s coming in June.

Chilly November (2022)

Yep, November was reasonably cool in the Sacramento area.  This was felt mostly in the lows.  The median low temperature for the month was 39F while the normal median is 45F.  The median high temperature was 1 degree below the normal median of 65F.   Below are the distributions of November 2022 lows and highs.  Despite the cool weather, we never even came close to setting a record low in Sacramento*, but it is quite possible that the high temperature on the 8th of just 53F was at least within sight of the record-low high-temperature for the day.

*The official low this morning in Davis of 30F would tie the official low in Sacramento for this date.  The low at Travis AFB was 26F!

Statistics of Sacramento Temperature Statistics

So, I want to talk a little bit about Sacramento temperature climatology.  This first figure below is what you might expect to see when thinking about temperature statistics: “normal” lows and highs; and “record” lows and highs.  These will feel familiar to anyone who lives in the Central Valley.  We have mild winters with little diurnal range and rather warm record lows.  We have what you might consider a long spring with slowly increasing normal highs and lows.  We have a very hot summer followed by quickly cooling normal and record temperatures in fall.  What I’m showing below is actually the 10-day running mean of temperatures.  This helps smooth out all the bumps and wiggles in the data.

But, what I really want to talk about is the variance of the normal and record temperatures.  Below, I’ve included a figure with 4 panels.  The first is the 10-day running variance (or variability) of the normal temperatures.  Since normals are locally monotonic (in time), this panel shows us how quickly the normal high and low change.  As we might have inferred from above, normal highs vary at 10-day time scales more than normal lows and do so most strongly toward the end of the year.  Interestingly, the second panel (the 10-day variance of the records) doesn’t look at all like the first panel.  The variance in record high temperatures is wiggly but functionally constant across the year.  This is not what I expected.  I had imagined the record highs would vary strongly in the summer and weakly in the winter.  The variance of the record low peaks in the cold season and is near zero around day 225 (mid-August).

On the bottom row, I have plotted 1) the 10-day running mean of the year of the record high and low (gaps occur when the same record is set in multiple years) and 2) the 10-day variance of the year of records.  Here, we see an obvious global warming signal — much more recent mean years of high temperature records than low.  The average difference is 56 years!  That’s incredible.  Also noteworthy: the 10-day mean year of record high decreases throughout the year.  I have no idea what to make of that — Our winters are more record warm than our falls?  The 10-day variance of the year of records is also a bit surprising.  Here, we see peaks in record low occurrence year in the cold season.  It seems (from this figure and looking at the raw data) that in the winter, we’ve actually had lots of instances of contemporary record lows even if on average they primarily occurred prior to 1940.

September Rain

24 hour accumulation of precipitation to 7am on the 19th in Davis was about 1″.  Below is the record (for Sacramento, at least) daily rainfall values for the month of September.  A 1″ storm would easily be a top 10 daily accumulation.  So, this rain was in rare company but was hardly unprecedented.

Records for the 23rd, 24th, and 25th were all set in 1904.  Assuming that was just one storm, total rainfall from that single event was (at least) 3.50″!

The Unseen Things Are Scary

See the source image

The scariest things in life can be those that are imagined rather than seen.  Science, in a way, is the process of taking the seen and making it not-scary through understanding and explanation.  But, what of the unseen?  Too often in physical science we conflate the unseen with the impossible. Take the quotation below that I ran across today which, in effect, I have heard uttered many times from physical scientists:

“We didn’t think it was possible because we had never seen it before.” (quotation I have altered to anonymize)

This kind of reasoning frustrates me continually.  Science is fantastic at observing the physical world but simultaneously fantastically poor.  We can be lulled to sleep by our considerable skills.  In 2022, data pours in to scientists in deluges.  We have so much data that we don’t know what to do with it all.  Despite this, there are (at least) two reasons we might never have observed something possible.  1) Rare events happen rarely. 2) We might not be looking.  So, rather than being lulled to sleep by our data richness, I think we should consider the value in being rousted by our collective naivety.

The scientific enterprise is grounded by observations.  That grounding prevents people like me, who may otherwise be so inclined, from living in an imagined world.  But, having failed to observe something in the past is not the same thing as its being impossible or even, necessarily, uncommon.  Science requires us to be looking (and indeed it compels us to look) in order to test an idea.  A lack of observations should not preclude all attempts to consider that which is consistent with existing knowledge.

If it does, then Lisa Simpson has a Tiger-Repelling Rock to sell us all.  Unseen tigers are scary.

Fog Fall

It has been an incredibly foggy autumn here in Davis (and throughout much of the CA Central Valley).  I made a histogram of daily rainfall totals from the past 50 days of my Cocorahs gauge.  I’ve grouped totals into 0″ (clear, cloudy, or light fog), 0.01″-0.02″ (heavy fog), and >=0.03″ (rain).  So far this autumn, over half of our days have seen measurable precipitation but 60% of those days have been from fog-fall rather from rainfall.

« Older posts

© 2024 Matt Igel

Theme by Anders NorenUp ↑