Determining Mixing Heights

[weatherwiz]

One of my friends had taught me to use mixing heights to forecast high temperatures so for the past year or so I've been working with doing that to make high temperatures.  It can be a very useful technique.  I obviously use bufkit for this b/c you can get temperatures at specific levels and lapse rates and such.  This past week I've been forecasting for Nashville in the weather challenge and I've been getting burned because of mixing way too low.  What I guess I do is look at the sounding and see where the temperature line stops becoming dry adiabatic (or slopes less than the dry adiabatic lapse rate) and use that as the mixing height).  But last week I guess this ended up being off by like at least 1000' on a day or two and I missed the high by like 4-5F.  I have googled for stuff on determining mixing heights but haven't really found information which helps.  

What are some techniques either within bufkit itself or just in general to sort of gauge how high the mixing heights will get?

Also, why does the NAM and GFS always seem to like differ by great distances on mixing height? 

[isohume]

I've never heard of derviving maxT from mixing heights. Mixing heights are calculated from maxT. The general method is fcst maxT then lift a parcel dry adiabatiacally until it reaches the sounding temp curve. That height is your mixing height, which can be well above or below the LCL. 

If you really want to fcst maxT from a sounding, I suggest using an old Air Force technique listed below. (This method has seasonal variations and isn't good near a large body of water. It also doesn't account for terrain, sfc-type albedo/absob, or adv effects.) 

If no inversion is present between 4-6 kft above the surface:

1. From the morning sounding, find the temp at 850 mb.

2. Clear/scattered clouds: Proceed from 850 temp dry adiabatically to the surface to get maxT.

3. Broken clouds/overcast: Proceed from 850 temp moist adiabatically to the surface to get maxT.

If an inversion is present between 4-6 kft above the surface:

1. Find the temperature at the warmest point (top) of the inversion.
2. From that point, proceed dry adiabatically to the surface for maxT. 

 

There are many reasons why each model has different solns, but the main reasons are what and how they parameterize various atmos variables, initialization schemes, and resolution (sigma levels).

Here are some of the variables which are parameterized by the models:

[weatherwiz]

This is great information, thank you!  

What I have been doing is this:

look at the soundings and determine where the temperature stops cooling dry adiabaically.  

Using that height level, look at the data section and find the closet level in the data section of bufkit.  Take this value (which is in feet, then divide by 3.28 to get meters, then 1000 to get km.  

Take this value and then multiply by whatever the dry adiabatic lapse rate is up to that level.

Add the temperature of whatever the temp is at the level you're mixing to.  Take this value, multiply by 1.8 then add 32 to get F.  

It works extensively well but just as long as you're on with mixing height level.  

It doesn't work well though on days with clouds of precip.   

[isohume]

That's a good procedure, but the level where temp stops cooling dry adiabatically is often not the mixing height.  BL mixing often continues into the cloud layer. 

[weatherwiz]

4 hours ago, isohume said:

That's a good procedure, but the level where temp stops cooling dry adiabatically is often not the mixing height.  BL mixing often continues into the cloud layer. 

That makes tremendous sense.  This explains why lately (as we move into the warm season) I've completely underestimated mixing height levels.  Thank you!!

[so_whats_happening]

I usually go from the top of the BL, usually but not always ends up being around 850mb (1.5km). Take the temp at this level and dry adiabatically take it to surface it really depends on the conditions below as to whether a place will fully mix and reach max temp or whether to keep it down a bit. I like the idea from isohume:

Quote

If no inversion is present between 4-6 kft above the surface:

1. From the morning sounding, find the temp at 850 mb.

2. Clear/scattered clouds: Proceed from 850 temp dry adiabatically to the surface to get maxT.

3. Broken clouds/overcast: Proceed from 850 temp moist adiabatically to the surface to get maxT.

If an inversion is present between 4-6 kft above the surface:

1. Find the temperature at the warmest point (top) of the inversion.
2. From that point, proceed dry adiabatically to the surface for maxT. 

 

Ill have to give this a try and see how it works. Nashville was a pain for sure. My biggest issue I know at times was the low temperatures in places have to get better with forecasting for that and winds I tend to go a little too high on some forecast days.

[weatherwiz]

Getting back to this topic there are a few things I've noticed and hopefully this semester I can get the time to really explore things further but attached are a bufkit profile sounding for Waterbury, CT from the 12z/13 runs of the GFS and NAM.  I've noticed (and this seems to be a constant across the country) but the GFS always seems to suggest much stronger and well-mixed boundary layers than the NAM does.  Like here the GFS looks to be well-mixed up past 5000' while the NAM is below 1000'.  Also, quite the differences with regards to temperature...21C as opposed to 28C