Tim Vasquez
EF5
- Joined
- Dec 4, 2003
- Messages
- 3,411
I thought I would make this post to demonstrate how an experienced forecaster sees a situation from scratch.
First of all while noticing gusty winds outside this morning in OKC and browsing the forecast, I saw our rain chances skyrocket tonight with storms in the forecast:
I pulled up a surface chart, which sort of shows the game play for the day. I did a quick analysis, sketching on patterns and processes that I noted mentally. Looks like a nice cyclone in western Kansas, suggesting a powerful surface system, though I had some indecision about the warm frontal position. This is a classic case where a speedy analysis sinks the diagnosis as we shall see a few steps further. I'm leaving this panel to show how this problem appeared and how it gets resolved later.
I noticed the cold front extended southwestward and the question was whether there was a system further west in the mountains. Warm weather and light winds in NM/AZ would be a "no". Cool weather, west winds, and showers would be a "yes". So what did we see? Strong winds, showers, and definite evidence of a front. Here I have outlined the two air masses and noted them using the basic Norwegian school abbreviations. You can also see how the Pacific front kind of overrides the cP air mass in Colorado, with some sort of low (probably cold core) in SW CO as evidenced by the strong showers and cyclonic appearance to the pattern in this region.
Now around this time I started getting suspicious that my frontal analysis in the Plains was wrong. The surface temperatures were kind of cool, the warm front was not well defined, so....... I turned to the soundings to see what the air masses were at play. These are the soundings stacked north to south along I-35. You can see that there is a frontal inversion (layer with moisture increasing with height) throughout the entire area, signifying an old polar air mass about 5000 ft deep.
Sure enough, I looked at the analysis further south and it does look like the true warm front is FAR to the south, in the Corpus Christi-Laredo area. So this forces me to re-inspect all the fronts further north.
The UCAR maps are too small and don't cover a large enough area so I loaded up Digital Atmosphere and did a display of theta (potential temperature) in orange, which gives us isotherms normalized adiabatically to sea level, which is better for analyzing in the Rockies. Now it was clear to me what was going on. (Click on map to enlarge) Old polar air covered much of the Plains and a Canadian intrusion was coming down the high plains. The real action was in NM/AZ. Boy, and that's quite a system... blustery northwesterly winds through Nevada and Utah feeding in the polar air.
Now I like to look at upper level conditions, which completes the game play by showing the 3D relationship of air masses and the basic thermal structure of the atmosphere. I go to the 500 mb level in the mid-troposphere, which gives the best general overview. We see a very potent upper low in the four-corners area, supporting (or indicating) the low we found in SW CO. This low and the massive troughing we see in the western US is a reflection of a very cold, deep air mass in the western US. We also note that the polar jet crosses Oklahoma, indicating Oklahoma is likely to be north of a front, and is not prominent over KS/NE/SD where we would expect to find upper level support or at least a cold core low for the big cyclonic system that made a fleeting appearance on map #1. We don't see that here.
Overall, the flow is very curved and we can't find short wave disturbances very well. However we do see a jet streak that is much stronger on the west side of the low. Most often in the southern Plains we see these jet streaks on the base of a long wave trough or east of it. Being on the west side is significant because it is (1) associated with a "digging", intensifying system (also associated with negative tilt troughs) and (2) since a jet streak indicates baroclinicity and fuel for cyclogenesis, this puts our attention on southern Arizona, southern New Mexico, and adjoining areas of old Mexico rather than further up the front in KS/NE. So now I know some specific things to look for on the models.
To see whether I'm on the right track, I look at the ETA fcst for 12 hours. Sure enough, it shows the area south of Mexico to be in an area of cyclogenesis.
Now this is where the diagnosis really affects our forecast. Shown below are the two diagnoses that we came up with. Diagnosis A has the warm front in south Texas. This suggests an isentropic lift (overrunning) warm frontal situation in Texas and suggests an east-northeast low movement with the arrow shown. Diagnosis B is our original analysis that has a strongly baroclinic Kansas system and the Mexican system working its way up the front, and much of Texas off the caprock would remain in the warm sector and be unaffected. These are two vastly different scenarios! This also shows how just glancing at the models can form very different impressions without an underlying analysis/diagnosis to support it.
So let's look at the 18 and 24 hour forecast and see how it compares. What we see is that this surface low makes fast progress northeastward. This is a movement of about 50 mph! This fast movement clearly ties it to an upper level short wave or upper level jet streak. What makes the results kind of surprising is that the thickness fields fail to reveal the warm front in Texas, and the low moves more in consistency with Diagnosis B. How can this be?
I went ahead and pulled up the RUC panel from Earl Barker's website (for some reason UCAR does not put thickness on its surface panel) and here this is a little more suggestive of the Texas warm front. We can also see the precipitation that forms from the lift along the front. So our Diagnosis A is still correct but the warm sector is advecting rapidly northward and our low tracks a little more northerly than expected in order to follow the air mass contrasts. The yellow dashes I drew in south Texas correspond to the starting location of the warm front 12 hours earlier so we can see how it has advanced.
Anyhow my lunch is here and I have to stop, but that gives a basic idea of how the foundation for a forecast is laid. This tells me that here in OKC we will have that Texas front coming up from the south tonight. I haven't gone so far as to look at instability factors but this tells me that instead of having surface based storms or an MCS we will have mostly elevated storms, isentropic lift, and maybe a few convective clusters, and as the warm front itself approaches we would have a chance for stronger storms if instability and moisture are favorable. Then at that point we will either be in the warm sector or see passage by the cold front itself, and either of those could involve an MCS if conditions are right. So to keep tabs on the warm front we would need to look at the 00Z soundings for OUN, FWD, and neighboring sites and also track the front's progress on the surface chart. We could also look at profiler data. All of this helps determine which type of convective modes will occur, which we consider in addition to shear and instability.
Though we do prohibit experienced members from starting threads or chatting amongst one another in this forum, we do allow and encourage tutorial threads like these.
Tim
First of all while noticing gusty winds outside this morning in OKC and browsing the forecast, I saw our rain chances skyrocket tonight with storms in the forecast:
I pulled up a surface chart, which sort of shows the game play for the day. I did a quick analysis, sketching on patterns and processes that I noted mentally. Looks like a nice cyclone in western Kansas, suggesting a powerful surface system, though I had some indecision about the warm frontal position. This is a classic case where a speedy analysis sinks the diagnosis as we shall see a few steps further. I'm leaving this panel to show how this problem appeared and how it gets resolved later.
I noticed the cold front extended southwestward and the question was whether there was a system further west in the mountains. Warm weather and light winds in NM/AZ would be a "no". Cool weather, west winds, and showers would be a "yes". So what did we see? Strong winds, showers, and definite evidence of a front. Here I have outlined the two air masses and noted them using the basic Norwegian school abbreviations. You can also see how the Pacific front kind of overrides the cP air mass in Colorado, with some sort of low (probably cold core) in SW CO as evidenced by the strong showers and cyclonic appearance to the pattern in this region.
Now around this time I started getting suspicious that my frontal analysis in the Plains was wrong. The surface temperatures were kind of cool, the warm front was not well defined, so....... I turned to the soundings to see what the air masses were at play. These are the soundings stacked north to south along I-35. You can see that there is a frontal inversion (layer with moisture increasing with height) throughout the entire area, signifying an old polar air mass about 5000 ft deep.
Sure enough, I looked at the analysis further south and it does look like the true warm front is FAR to the south, in the Corpus Christi-Laredo area. So this forces me to re-inspect all the fronts further north.
The UCAR maps are too small and don't cover a large enough area so I loaded up Digital Atmosphere and did a display of theta (potential temperature) in orange, which gives us isotherms normalized adiabatically to sea level, which is better for analyzing in the Rockies. Now it was clear to me what was going on. (Click on map to enlarge) Old polar air covered much of the Plains and a Canadian intrusion was coming down the high plains. The real action was in NM/AZ. Boy, and that's quite a system... blustery northwesterly winds through Nevada and Utah feeding in the polar air.
Now I like to look at upper level conditions, which completes the game play by showing the 3D relationship of air masses and the basic thermal structure of the atmosphere. I go to the 500 mb level in the mid-troposphere, which gives the best general overview. We see a very potent upper low in the four-corners area, supporting (or indicating) the low we found in SW CO. This low and the massive troughing we see in the western US is a reflection of a very cold, deep air mass in the western US. We also note that the polar jet crosses Oklahoma, indicating Oklahoma is likely to be north of a front, and is not prominent over KS/NE/SD where we would expect to find upper level support or at least a cold core low for the big cyclonic system that made a fleeting appearance on map #1. We don't see that here.
Overall, the flow is very curved and we can't find short wave disturbances very well. However we do see a jet streak that is much stronger on the west side of the low. Most often in the southern Plains we see these jet streaks on the base of a long wave trough or east of it. Being on the west side is significant because it is (1) associated with a "digging", intensifying system (also associated with negative tilt troughs) and (2) since a jet streak indicates baroclinicity and fuel for cyclogenesis, this puts our attention on southern Arizona, southern New Mexico, and adjoining areas of old Mexico rather than further up the front in KS/NE. So now I know some specific things to look for on the models.
To see whether I'm on the right track, I look at the ETA fcst for 12 hours. Sure enough, it shows the area south of Mexico to be in an area of cyclogenesis.
Now this is where the diagnosis really affects our forecast. Shown below are the two diagnoses that we came up with. Diagnosis A has the warm front in south Texas. This suggests an isentropic lift (overrunning) warm frontal situation in Texas and suggests an east-northeast low movement with the arrow shown. Diagnosis B is our original analysis that has a strongly baroclinic Kansas system and the Mexican system working its way up the front, and much of Texas off the caprock would remain in the warm sector and be unaffected. These are two vastly different scenarios! This also shows how just glancing at the models can form very different impressions without an underlying analysis/diagnosis to support it.
So let's look at the 18 and 24 hour forecast and see how it compares. What we see is that this surface low makes fast progress northeastward. This is a movement of about 50 mph! This fast movement clearly ties it to an upper level short wave or upper level jet streak. What makes the results kind of surprising is that the thickness fields fail to reveal the warm front in Texas, and the low moves more in consistency with Diagnosis B. How can this be?
I went ahead and pulled up the RUC panel from Earl Barker's website (for some reason UCAR does not put thickness on its surface panel) and here this is a little more suggestive of the Texas warm front. We can also see the precipitation that forms from the lift along the front. So our Diagnosis A is still correct but the warm sector is advecting rapidly northward and our low tracks a little more northerly than expected in order to follow the air mass contrasts. The yellow dashes I drew in south Texas correspond to the starting location of the warm front 12 hours earlier so we can see how it has advanced.
Anyhow my lunch is here and I have to stop, but that gives a basic idea of how the foundation for a forecast is laid. This tells me that here in OKC we will have that Texas front coming up from the south tonight. I haven't gone so far as to look at instability factors but this tells me that instead of having surface based storms or an MCS we will have mostly elevated storms, isentropic lift, and maybe a few convective clusters, and as the warm front itself approaches we would have a chance for stronger storms if instability and moisture are favorable. Then at that point we will either be in the warm sector or see passage by the cold front itself, and either of those could involve an MCS if conditions are right. So to keep tabs on the warm front we would need to look at the 00Z soundings for OUN, FWD, and neighboring sites and also track the front's progress on the surface chart. We could also look at profiler data. All of this helps determine which type of convective modes will occur, which we consider in addition to shear and instability.
Though we do prohibit experienced members from starting threads or chatting amongst one another in this forum, we do allow and encourage tutorial threads like these.
Tim
