will meterologists become extint?

calvinkaskey

Guest
Feb 17, 2014
384
30
11
How are we doing on winter forecasts as compared to previous years? Latest event showed forecast radar did much better than what the NWS, TWC and all the other forecasters said. 3 inches or so in binghamton ny and surrounding area. Much less than forecast. Forecast radar showed precip 1.5hrs drive to the north in maybe 12 hr loop at the time the warning was issued. Any complaint s about the BGM meterologists? They issued a storm warning during near 40 temp at the highest elevations. Roads r slushy with only 3 inches in like 12 hrs. What in the world? But that is another issue.
 
May 1, 2011
160
150
11
38
Michigan
www.lakefx.net
It's like this. You can chuck a rock into the water and predictably there will be a splash. The splash will be proportional to the rock. But the shape of the rock when it hits the water, and whether there are waves and other factors will impact the outcome of the details of that splash. The details of that splash, every ripple and swirl are your inches of precip or exact temperature in a given location. You can 3D model this all day long and you can get a representation of the results, However you can only know every detail of that splash if you know every detail about the rock and the water it is hitting, speed, direction, orientation, etc.

So, for robots to take over meteorology to the point of exactness necessary for perfection, You would need to completely saturate the environment with sensors. Since those sensors have to consume and release heat, and by existing at all probably change the environment being sampled....The end result is a net loss for everyone, and changing the outcome of what was to happen in the first place. I want to see the pretty lake I'm chucking a rock into....Not a sea and sky full of senors and blinking lights....Just so I know whether I will get 7 inches of dry snow, versus 3 inches of wet snow. Who cares...it will all land on sensors anyway.
 

rdale

EF5
Mar 1, 2004
6,906
416
21
49
Lansing, MI
skywatch.org
To take this on a more serious note - I believe "extinct" is clearly incorrect, but in 2115 I can't imagine a country with 100+ NWS WFOs each with 20+ mets doing a daily forecast and hundreds of TV stations with mets. I bet the TV thing is done in the next 50 years at the most. You will need meteorologists to do the research and write the code for the models, but I just don't believe that in 100 years we'll still see a meteorologist putting eyes on the radar to press a button that issues a Tornado Warning. You won't have a meteorologist perusing surface maps to give you today's high. AI will be far too advanced to leave that to humans ;)
 

Todd Lemery

Supporter
Jun 2, 2014
468
449
21
53
Menominee, MI
A couple of years ago, Newsweek or Time magazine (I don't recall which one) had an interesting article on the trajectory of computing that the world is on. In a nutshell, around 35 years from now, computers will have evolved to a tipping point where they will become self learning and far surpass the ability of the human mind to figure out complex problems that elude us. The result will be dramatic advances in areas such as medicine, aging, manufacturing and I can easily imagine meteorology.
There is still the garbage in, garbage out issues that would have to be worked out obviously, but it isn't too much of a stretch to me that the time a meteorologist or chaser spends pouring over soundings, satellite loops etc..trying to choose the most likely target area for tornados could be reduced to less than a second with a more accurate outcome.
 
May 25, 2012
202
166
11
36
Albuquerque, NM
www.wxlog.com
Good points about increased sensor fidelity, machine-learning, etc. But I think there's a simpler aspect that's been overlooked. If a robot issues a tornado emergency and then takes no action to save the people in that area, wouldn't it violate the 1st Law of Robotics (i.e. through inaction, allows a human being to come to harm)?!?!?!?!
 

rdale

EF5
Mar 1, 2004
6,906
416
21
49
Lansing, MI
skywatch.org
Stephen - I'm not familiar with that "law" but I'm not clear. Right now when people get a tornado warning, there is not an actual meteorologist there telling them where to go. I'd bet many don't even know a tornado warning is NOT automated.

In addition, I have no doubts that in 25 years your smart device will not only alert you to the warning, but show you based on traffic and storm motion where the safest place to go is if you're in the car, and if you're in your home it will tell you what part of your house is most likely to survive. Humans can't do any of that...
 

Brian G

EF2
Sep 25, 2014
105
25
11
St. Louis, MO
The 1st law of robotics is in reference to Isaac Asimov's three laws of robotics. I took it in gest which is what I think it was intended to be. But yeah, it wouldn't surprise me if many tornado warnings were automated within 25 years. The Warn-on-Forecast project is a step in that direction already. It would still be a big leap to completely remove the human element, but at least it is not out of the realm of possibility even if it is slim at best.
 
Last edited by a moderator:

calvinkaskey

Guest
Feb 17, 2014
384
30
11
so no real comment about winter storm warning in place with no precip and 40 degree t

I WOULD HAVE DONE MUCH BETTER JUST LOOKING AT FUTURE RADAR
The 1st law of robotics is in reference to Isaac Asimov's three laws of robotics. I took it in gest which is what I think it was intended to be. But yeah, it wouldn't surprise me if many tornado warnings were automated within 25 years. The Warn-on-Forecast project is a step in that direction already. It would still be a big leap to completely remove the human element, but at least it is not out of the realm of possibility even if it is slim at best.
 
May 25, 2012
202
166
11
36
Albuquerque, NM
www.wxlog.com
The 1st law of robotics is in reference to Isaac Asimov's three laws of robotics. I took it in gest which is what I think it was intended to be.
Haha yea sorry, a little snark.

All that aside, it is interesting to consider what evolution of the human element in forecasting will be over the next 100 years. I think rdale could be right about the shrinking need for regional WFO offices (and I would be curious if anyone has a counter-argument).

But the idea that "future radar" could currently replace humans in the loop is definitely scifi. And speaking of scifi, this topic makes me think of a great short story that imagines what human science would look like once we aren't capable of understanding the cutting-edge anymore. Calvin maybe you will like this story (no snark, I genuinely love this author).
 
Theoretically I could see it happening when models actually are able to resolve even the smallest of eddies or pressure protuberances on the microscale but that it still a very long ways off. It would also require a critically thinking computer to completely replace a human, which is the entire crux of the matter.
 

Jeff Duda

Resident meteorological expert
Staff member
Oct 7, 2008
2,986
1,498
21
Westminster, CO
www.meteor.iastate.edu
The day has already come and gone where a human can no longer outforecast a computer below the mesoscale. Sure, humans can still forecast the s-it out of the synoptic scales, but NWP models are getting better and better, while human forecasting really isn't improving that much. The amount of information available for forecasting at the mesoscale and below is simply too much for a human mind to manage.
 
Last edited by a moderator:
Jan 18, 2015
232
58
11
Tucson, AZ
They day has already come and gone where a human can no longer outforecast a computer below the mesoscale. Sure, humans can still forecast the s-it out of the synoptic scales, but NWP models are getting better and better, while human forecasting really isn't improving that much. The amount of information available for forecasting at the mesoscale and below is simply too much for a human mind to manage.
I think as a whole of society though we are losing our touch with nature which can also be the reason why humans aren't improving like NWP models. I still believe that a machine can never replicate human intuition and we'll always have that
 

rdale

EF5
Mar 1, 2004
6,906
416
21
49
Lansing, MI
skywatch.org
Intuition is just a fancy word for "This matches something in my memory bank"... If your memory bank is great (i.e. you are a meteorologist, or have chased a similar setup in the past) you probably have a good intuition. A computer can beat us all when it comes to that. Intuition can't replace education / experience. Too often I see that term used by social mediarologists with no real value.
 
Jan 18, 2015
232
58
11
Tucson, AZ
Intuition is just a fancy word for "This matches something in my memory bank"... If your memory bank is great (i.e. you are a meteorologist, or have chased a similar setup in the past) you probably have a good intuition. A computer can beat us all when it comes to that. Intuition can't replace education / experience. Too often I see that term used by social mediarologists with no real value.
Haha I kind of like that term. I was kind of assuming those with experience and education though and trusting their guts
 
May 1, 2011
160
150
11
38
Michigan
www.lakefx.net
While we certainly have a disconnect with nature.... Getting worried about 3 inches of wet snowfall instead of 10 inches of dry snowfall in a forecast at a specific elevation in Appalachia is more representative of that problem, rather than lack of ability to forecast it perfectly. What honestly is going to change your life with 3" versus 10" of snowfall? If your life is thrown into upheaval over that difference, that's a reflection of yourself or of society. Honestly it could be a blizzard outside right now, and I'm pretty sure if I had to go somewhere I know how to handle myself and I also know better than to get caught or put myself in that situation. I certainly don't blame a weather forecaster for getting it wrong. I don't worry that today's forecast temp was off by 5 degrees and it was more like mostly cloudy than partly cloudy. The person who runs to the store to buy bread and milk over snowfall has absolutely no concept of survival or the natural state of things. It's winter...Winter things happen. Be prepared for winter stuff---And by the way, speaking of what's natural, be awestruck that you can buy a strawberry in February---your great-great grandparents would have reacted like the crew of the Serenity.

I don't think humans need to worry about global forecast patterns and mesoscale features like this. We need to understand the concepts and the math, but then allow machines to handle it. It's fine and dandy that you can long division, but calculators are more accurate, faster, and it allows humans to handle more important intuitive and big picture concepts. Let's not sweat the details...We have machines for that.
 
Last edited by a moderator:

calvinkaskey

Guest
Feb 17, 2014
384
30
11
Computer models continue to overestimate cold outbreaks

FOR YEARS THAT HAS BEEN GOING ON BUT DOES ANYONE DO ANYHING ABOUT IT? No. Im not a meteorologist but I am calling we won't see a high of 0 or less at BGM (Binghamton regional airport his Sunday) and I called it days before everyone else.
While we certainly have a disconnect with nature.... Getting worried about 3 inches of wet snowfall instead of 10 inches of dry snowfall in a forecast at a specific elevation in Appalachia is more representative of that problem, rather than lack of ability to forecast it perfectly. What honestly is going to change your life with 3" versus 10" of snowfall? If your life is thrown into upheaval over that difference, that's a reflection of yourself or of society. Honestly it could be a blizzard outside right now, and I'm pretty sure if I had to go somewhere I know how to handle myself and I also know better than to get caught or put myself in that situation. I certainly don't blame a weather forecaster for getting it wrong. I don't worry that today's forecast temp was off by 5 degrees and it was more like mostly cloudy than partly cloudy. The person who runs to the store to buy bread and milk over snowfall has absolutely no concept of survival or the natural state of things. It's winter...Winter things happen. Be prepared for winter stuff---And by the way, speaking of what's natural, be awestruck that you can buy a strawberry in February---your great-great grandparents would have reacted like the crew of the Serenity.

I don't think humans need to worry about global forecast patterns and mesoscale features like this. We need to understand the concepts and the math, but then allow machines to handle it. It's fine and dandy that you can long division, but calculators are more accurate, faster, and it allows humans to handle more important intuitive and big picture concepts. Let's not sweat the details...We have machines for that.
 
There will "always" be a need for human forecasters... just maybe not as much in the traditional sense that people seem to be focusing on. Okay, so one day we could have accurate enough forecasts out to five days that no human input is needed. What about the 6-10 day? 11-15 day? Beyond that? There will seemingly always be a way for humans to edge out the computer output, even if that means having to move the goalposts.

Humans will almost always have an edge in the nowcast as well. There is no way for a computer to catch every individual storm in the exact location at the exact time... there's just not enough available observations, and obs won't be 100% accurate, so even perfected computer software wouldn't be able to produce a perfect forecast. You get something highly sensitive to that kind of stuff like a big sporting event, and you bet a meteorologist with a radar, visible satellite and other obs will be able to outperform a short term model.

As of now, model (and human) day-by-day forecasts beyond 10 days have a hard time just keeping up with a forecast based purely on climo, so I don't think we have anything to worry about for a long time. I'm excited to see how far we can progress in my lifetime, but I'm certainly not worried about getting replaced by a computer.