My personal love for all things weather related comes from my childhood days of always wanting to know how things work. But I grew up in Arizona. Fast forward to 2005 when I moved to Arkansas. I didn't know a thing about a single component to severe weather. My first spring here, a couple of rowdy storms and tornado warnings caught my attention. I was hooked.
Since then, a near obsession has taken place. I have learned a few things. The last couple years I've harnessed the power of the internet and modern programming languages (OK, Python, really) to download and analyze data based on my knowledge of how things work. It's been fun. And eye-opening.
I've received tons of help from many people, especially professionals from both government and media meteorologists as well as others on social media.
And then, it hit me. Last night really. Weather forecasting maps are a problem. It's been said before, argued about on social media. Don't trust this map or something of the sort. But why? Every meteorologist uses them.
Here is why weather maps are a problem. Weather maps plot a single value on a single point. Be it snowfall or temperature or rain, almost every map enables one to zoom in and see the value of the point.
I have learned over time that the actual value of a single point on a weather map, and you can pick any of the parameters you like, or any model, isn't a number but, and this is key, a range of numbers.
A single number is the function output, yes, but it's true value is a range.
As models have improved, in both resolution and physics, this range has decreased.
But the number given is still somewhere inside a range which increases with time.
I'll still make and tweet these maps. But know this: They are almost never right. They are only what they are. A dart throw somewhere towards the dartboard.
Meanwhile, pay attention to the National Weather Service. They are really good at outlining the range of scenarios you will encounter with the weather. But you have to actually read the forecast.
Cheers!
Steve
Wednesday, February 20, 2019
Tuesday, May 23, 2017
Details About My GFS Forecast Graph
I created this graph as a tool to help forecast precipitation for Fayetteville, AR. It displays some common items found elsewhere like temperature and dew points every three hours as forecasted by the GFS model. Both temperature plots are 980MB temps, not surface temperature forecasts. The data is parsed from BUFKIT files generously provided by Penn State University. This data is a point forecast for my favorite airport, Drake Field, or KFYV. I tweet this thing out occasionally when it seems worth it.
Starting from the bottom up, the next seven days are listed with times adjusted to local time. Above that is the GFS precipitation type, R=rain, S=snow, P=sleet and hopefully never, Z=freezing rain.
The yellow line is a crude wind shear measure, simply wind speed at 3KM - wind speed at surface.
The green dots display accumulated precipitation in units that vary depending upon seasons and events. This is data from the most recent model.
Red stars indicate CAPE as provided by the GFS. It is displayed as CAPE/50 for better resolution on a 0 to 100 Y axis. A red star at 40 is equal to 40*50, or 2000 j/kg.
Finally, the black triangles boldly claim to forecast percent chance of precipitation. The GFS model either says it will rain or not. It doesn't list a percentage. So I came up with my own formula, and frankly, it works pretty well. Many times there will be trace outputs of rain from the GFS. Including these trivial amounts weighted the chance of rain unrealistically high. So I arbitrarily chose a small threshold as "actual rain".
The next step was to save the last 12 GFS runs, and compare the hourly results. The thinking is trying to gauge consistency. Initially I compared 16 runs, but the results became "noisy".
Finally, I sum the comparisons, weighted to give preference to the most recent run, this way:
Most Recent Run =50
Run Before That =25
Run Before That =12.5
Run Before That =12.5/2
Run Before That =Half of the last Run
and so on.
What you'll see, is when the graph shows a POP at 50% or above, the most recent run shows rain.
If the last 2 show rain, its 75%.
If every run shows it, its nearly 100%.
If every run except the last one shows it, it is below 50%.
Although this method is far from flawless, at least now you have an idea of what you are looking at.
Hope you enjoy it!
Steve
Monday, October 31, 2016
2016-2017 Winter Outlook
Dateline: Halloween 2016
This will probably lead to some controversy, as the truth sometimes does. I am compelled to write about the upcoming winter season and what kind of weather it will bring.
The data I use for this forecast varies, and includes several maps of things most people don't care about or understand, So they will not be included. I also research and track several parameters with fancy acronyms like PDO, AMO, ENSO, SST's and others. I know, blah blah blah.
I have read and considered the various folklore, almanacs and official forecasts, some of which carry more weight than others. I've looked back at recent forecasts and compared skill levels. I've spent considerable time this month on research and study.
The official "word on the street" is warmer and dryer than average due to a weak La Nina. This is where the controversy begins.
First off all, the La Nina/El Nino effect on Northwest Arkansas is virtually zero, as I have pointed out before with 50 years of daily temperature and precipitation reports. There are certain trends elsewhere, but NWA is in a belt of weather that does not significantly trend in either direction for either scenario. That is a fact.
There are other far more important global conditions that influence the weather here. But how they interact, and what the overall effect is, is not only complicated but anecdotal; For instance it snows a lot here when x and y are this and that. But nobody claims to really know when the jet stream will move, it just does it when x and y are this and that. This is an oversimplification of the fact that most of our official forecasts are based on history with little real skill at predicting the future.
Here is the kicker. At no point in any of you or your ancestors lives (pretend forever) has the level of CO2 in the puny layer of air we breath been as high as it is right now. Like it or not, CO2 traps some heat on earth from escaping the planet. Fortunately, oceans absorb some of both. The point is this: There exists no record of any kind for forecasting weather with the current atmosphere.
The good news is that if this planet wasn't really good at keeping things balanced, we would not be having this little blog about it. So short term catastrophe seems not super likely.
So my forecast? You still want it? It very well might not snow a single time for the next 8 months in Northwest Arkansas. Clearly, a staggering warming trend in under way. Another possibility is a brutally harsh and snowy winter that suddenly shows up "out of nowhere". What is more certain, and this is the bad news, is that we are going to be rewriting a whole bunch of weather records.
I'm not a big fan of "surprises". So I'm doing what I can, and encouraging others, to plan ahead by trying to keep the climate of this planet close to where it is today. It's really the most important job you have, to keep the environment safe for the future. Just because it demands sacrifice, requires some deep thinking, and is kind of gloomy doesn't mean Climate Change is a hoax. Take it from those, like me, that are trying to sell absolutely nothing.
This will probably lead to some controversy, as the truth sometimes does. I am compelled to write about the upcoming winter season and what kind of weather it will bring.
The data I use for this forecast varies, and includes several maps of things most people don't care about or understand, So they will not be included. I also research and track several parameters with fancy acronyms like PDO, AMO, ENSO, SST's and others. I know, blah blah blah.
I have read and considered the various folklore, almanacs and official forecasts, some of which carry more weight than others. I've looked back at recent forecasts and compared skill levels. I've spent considerable time this month on research and study.
The official "word on the street" is warmer and dryer than average due to a weak La Nina. This is where the controversy begins.
First off all, the La Nina/El Nino effect on Northwest Arkansas is virtually zero, as I have pointed out before with 50 years of daily temperature and precipitation reports. There are certain trends elsewhere, but NWA is in a belt of weather that does not significantly trend in either direction for either scenario. That is a fact.
There are other far more important global conditions that influence the weather here. But how they interact, and what the overall effect is, is not only complicated but anecdotal; For instance it snows a lot here when x and y are this and that. But nobody claims to really know when the jet stream will move, it just does it when x and y are this and that. This is an oversimplification of the fact that most of our official forecasts are based on history with little real skill at predicting the future.
Here is the kicker. At no point in any of you or your ancestors lives (pretend forever) has the level of CO2 in the puny layer of air we breath been as high as it is right now. Like it or not, CO2 traps some heat on earth from escaping the planet. Fortunately, oceans absorb some of both. The point is this: There exists no record of any kind for forecasting weather with the current atmosphere.
The good news is that if this planet wasn't really good at keeping things balanced, we would not be having this little blog about it. So short term catastrophe seems not super likely.
So my forecast? You still want it? It very well might not snow a single time for the next 8 months in Northwest Arkansas. Clearly, a staggering warming trend in under way. Another possibility is a brutally harsh and snowy winter that suddenly shows up "out of nowhere". What is more certain, and this is the bad news, is that we are going to be rewriting a whole bunch of weather records.
I'm not a big fan of "surprises". So I'm doing what I can, and encouraging others, to plan ahead by trying to keep the climate of this planet close to where it is today. It's really the most important job you have, to keep the environment safe for the future. Just because it demands sacrifice, requires some deep thinking, and is kind of gloomy doesn't mean Climate Change is a hoax. Take it from those, like me, that are trying to sell absolutely nothing.
Subscribe to:
Posts (Atom)