Big Cities & Tornadoes

[Jeremy Den Hartog]

This may or may not have been already discussed here, but my attempts of searching for it were unsuccessful - so hopefully I haven't overlooked it...

Not this past Spring but the Spring before while attending the annual storm spotting classes in my local area, a NWS employee doing the seminar brought up something I found to be interesting. He said the NWS is currently researching what role large metropolitan areas play on the development of severe weather and tornadoes. More specifically, what difference does the temperature change between in a metro and not in a metro make (metros are generally a few degrees warmer because of all the blacktop, black roofs, etc). The reason this was brought up was because here in Southern Minnesota there is the town of St Peter (its about 50 miles south and west of Minneapolis/St Paul). This town has had 3-4 tornadoes either hit the town or nearly hit the town in just over 10 years. At least 2 of these tornadoes were F3 (of course this is before they changed the scale) . One of the tornadoes hit St Peter in 1998 (an F3 tornado) doing extensive damage to the town. Interestingly, in 2006 another F3 tornado followed almost the EXACT same path as the 1998 one but at the last second it shifted a little further south and east and spared the town by a few hundred yards. Anyhow, the NWS employee stated that early indications of the research has shown there seems to be an increase in tornado producing storms in roughly a 50-60 mile radius around large metros. What do you guys think about this?

 

[David Draun]

The observation network has become more dense, allowing more tornadoes to be seen by the public. Also metro areas have been growing, so its more likely for a tornado to occur closer to a said area.

The fact is, we aren't seeing a huge year-to-year increase in tornadoes near metro areas, just more tornadoes being seen due to better spotting and detection.

 

[Scott Weberpal]

I'm not sure how St. Peter is part of this discussion as its a relatively tiny town...its just unlucky! I can certainly see an urban head island having some effect on tornadogenesis, but I'll leave it up to the study to come up with conclusions.

 

[Scott Weberpal]

The fact is, we aren't seeing a huge year-to-year increase in tornadoes near metro areas, just more tornadoes being seen due to better spotting and detection.

Are you sure that's a fact? A tornado in a metro area is going to be seen and going to do damage whether or not there's a large spotter network in the area.

 

[Jeremy Den Hartog]

I'm not sure how St. Peter is part of this discussion as its a relatively tiny town...its just unlucky! I can certainly see an urban head island having some effect on tornadogenesis, but I'll leave it up to the study to come up with conclusions.

The reason St Peter was a part of the discussion was simply because it was a town near where the spotter training classes were held and the NWS employee used it as an example of what they are of looking at...

I guess it could be St Peter is just unlucky too - like you said

 

[scott r currens]

I think it is possible that baroclinic boundaries associated with heat islands may enhance tornado potential. But how would you prove that? Our tornado database has a massive population bias.

 

[Jeremy Den Hartog]

I think it is possible that baroclinic boundaries associated with heat islands may enhance tornado potential. But how would you prove that? Our tornado database has a massive population bias.

That would be difficult for sure and is of course part of the problem with making it factual. However, it seems to be worth considering while you are chasing and trying to decide on which storm to go after if you happen to be near a large city - storms 50 or 60 miles outside of a large city may have a 'hidden' element that other storms in the area may not have...

Also keep in mind, while this sort of thing may have been going on for a long time, technological improvements over the last few decades have made it possible to 'see' or know about more tornadoes then ever before. Consequently, the NWS is able to research how/why tornadoes are forming more so then ever before. Unfortunately they can't really rely on data prior to these technological advances (for example prior to doppler radar) simply because there were probably ALOT of tornadoes that went undetected....just figured I'd point this out (even if it is probably obvious to most here) in case anyone was wondering why they don't just go back and look at the tornado statistics over the last 100 years and see if there is a higher percentage of tornadoes 50 or 60 miles outside of a metro.

 

[Jeremy Den Hartog]

Another interesting thing to consider - how much of a role (if any) does 'ground temperature' play. For example the temperature differences between large areas of differing soil types. Does 'sandier' soil make any difference? Or large bodies of water - do they make any difference due the fact its usually a few degrees cooler near them?

I guess it just goes to show we still have ALOT to learn...

 

[Scott Olson]

St. Peter is such a small sliver of a town. Now, one thing that may play a bigger role is the significant change in topography when you reach St. Peter. It's kind on a dividing line for topography in Minnesota.

 

[Jeremy Den Hartog]

St. Peter is such a small sliver of a town. Now, one thing that may play a bigger role is the significant change in topography when you reach St. Peter. It's kind on a dividing line for topography in Minnesota.

Exactly, thats why it is interesting (or very unlucky) that this town of approximately 10,000 in Minnesota of all places has had so much tornado activity in or nearby in such a relatively short period of time (10 years). I realize in places further down south such as Kansas, Oklahoma, etc this may not be as big of an oddity (for obvious reasons) but up here in Minnesota it is highly unusual...

 

[Jeremy Den Hartog]

Here is a video of the 2006 St Peter tornado if anyone is interested...
http://www.youtube.com/watch?v=adsfGUqJyjA

Here is a recent article that refers to the 'heat island' effect. I should note that this study is more focused on tornadoes and their intensity when they go THROUGH a large metro area...
http://www.meteorologynews.com/2009/03/19/study-links-tornadoes-to-urban-heat-island-effect/

I also decided to do a little 'research' myself in regards to both the 1998 and 2006 St Peter tornado. I looked at the high and low temperatures for both St Peter and Minneapolis for the day of the tornadoes and each day for 1 week prior.

First, the 1998 tornado in St Peter happened on March 29 (yes I said March and the storm not only produced an F3 tornado which hit St Peter, but also an F4 tornado a little further south and west). Here are the temps for both locations (I included a column on the far right with the difference between high temperatures in Minneapolis and St Peter)...

Minneapolis High Low Difference (High only)
Mar 23, 1998 44 30 -1
Mar 24, 1998 45 23 +5
Mar 25, 1998 59 32 +19
Mar 26, 1998 72 49 +16
Mar 27, 1998 63 49 -5
Mar 28, 1998 67 39 +11
Mar 29, 1998 59 48 +2

St Peter
Mar 23, 1998 45 30 +1
Mar 24, 1998 40 21 -5
Mar 25, 1998 40 26 -19
Mar 26, 1998 56 40 -16
Mar 27, 1998 68 47 +5
Mar 28, 1998 56 44 -11
Mar 29, 1998 57 45 -2

In 1998, Minneapolis was warmer then St Peter 5 out of the 7 days. It was warmer by 10 degrees or more on 3 of these days including the day before the tornado that hit St Peter. It seems plausible that a heat island may have had an effect. However, the storm which produced the tornado that hit St Peter had also produced an even larger (F4) tornado earlier in its lifespan further south and west. That tornado died off as it approached the Nicollet, MN area and then the storm 'cycled' and developed another tornado which eventually hit St Peter.

Below is an image of the track of the 1998 tornado (as well as the F4 further south and west)...

Next lets take a look at the 2006 tornado which happened on August 24th. Below are again the temperatures for both Minneapolis and St Peter for the day of the tornado and the week prior.

Minneapolis High Low Difference (High only)
Aug 18, 2006 78 67 -9
Aug 19, 2006 78 60 -1
Aug 20, 2006 83 59 +11
Aug 21, 2006 86 63 +5
Aug 22, 2006 83 60 +1
Aug 23, 2006 82 59 NA
Aug 24, 2006 75 64 -6

St Peter
Aug 18, 2006 87 55 +9
Aug 19, 2006 79 62 +1
Aug 20, 2006 72 55 -11
Aug 21, 2006 81 55 -5
Aug 22, 2006 82 60 -1
Aug 23, 2006 82 60 NA
Aug 24, 2006 81 60 +6

This time the differences are a little more subtle. Minneapolis was only warmer then St Peter 3 out of the 7 days with 1 day being the same temperature. It should be noted though, on the day of the tornado in St Peter (the 24th), there were severe storms with baseball sized hail that went through the Minneapolis area in the morning 10 AM or 11 AM which of course would lead to less sun and consequently lower temperatures.

Below is the track of the August 24 tornado...

Of course, my little research project is far from being conclusive (not to mention I'm no expert). I don't have dewpoints or weather conditions for the given period of time and its only 2 tornado outbreaks so the data isn't exactly great. Still, I find it interesting that in both tornado events, Minneapolis was 10+ degrees warmer then St Peter exactly 5 days before the tornado event and still 5+ degrees warmer 4 days prior to the event. However, being that we are talking several days prior to an event it is hard to say how much any of this would have on a given event - my guess is probably very little if any. It is worth noting though that the low temperature in Minneapolis was 3 and 4 degrees higher then the low in St Peter on the days of the tornadoes.

For the heck of it, I figured I would take a look at the high temperatures in Mankato, MN for the given events and compare them to St Peter. In 2006, Mankato was a whopping 7 degrees warmer then St Peter. Keep in mind, St Peter is less then 10 miles north of Mankato (Mankato is a town of about 50,000). In 1998 though, St Peter was actually a couple degrees warmer. This of course raises the question of how big does a city need to be to have a 'heat island'? If you look at the track of the 2006 tornado it almost appears to 'arch around' Mankato (if that makes sense). Makes you wonder (or it does me anyway). This is pure speculation at this point but I would guess the reason there is such a temperature difference between St Peter and Mankato in 2006 is because the passage of an outflow boundary which came from the morning storms that passed to the north and hit the Minneapolis area. However, if this is the case, does this mean a 'heat island' (from Mankato) interacted with the outflow boundary and either slowed or stopped the southward progression of the outflow boundary or did it somehow 'enhance' the storm/tornado (as suggested in the article I included above). IMO, it enhanced the storm and tornado. Look at the tornado track from 2006. As it approached the area it was having difficulty developing near New Ulm, MN. Then as it approached the Mankato/St Peter area it suddenly exploded into an F2 and then F3 tornado. As it left the area it quickly weakened then died off all together. Also, there is a significant difference between the 1998 tornado and the 2006 tornado - time of year. I would imagine that in March, heat islands probably do not have the same effect simply because its the 'cool' time of year for the area and the sun angle is lower (kind of like UV rays). Anyhow, its hard to say how much any of this has to do with 'heat islands' so the bottom line is this - who knows

 

[V. Gensini]

Wait...What?

That would be difficult for sure and is of course part of the problem with making it factual. However, it seems to be worth considering while you are chasing and trying to decide on which storm to go after if you happen to be near a large city - storms 50 or 60 miles outside of a large city may have a 'hidden' element that other storms in the area may not have...

Any "conclusions" one would come up with would be completely bogus considering the quality of our storm reporting database.

Of course large metro areas have an impact on severe weather! They alter our day to day weather, so certainly they can alter mesoscale features that would be more/less conducive to severe weather! There are many referred articles that prove land use has a significant impact on the surrounding weather from day to day...How do they alter is a question yet to be answered.Think about it from a From a theoretical perspective... Holding all else equal, a parcel of air in the same environment would have a higher updraft velocity and therefore carry the potential to produce more severe weather than a given parcel at a lower T just a few miles away. But what about the buildings? Maybe more friction at the surface decreases velocities and hampers tornado production? This of course would be impossible to prove given the fact that you would have be completely certain that the source region of the parcel is exactly over the metro area and has not mixed with the surrounding environment. In my opinion, in most synoptic/meso severe weather cases, the scale of the phenomena we are talking about is negligible. However, I think there is a valid case for enhancement of unorganized single cell induced severe weather (e.g. lightning) around large metro areas...