Maximum p_top_requested setting for HRRR?

Discuss advection schemes, timestepping, pressure gradients, coordinates, etc.

Maximum p_top_requested setting for HRRR?

Postby Tntaviwx7 » Tue Oct 05, 2021 2:36 pm


I'm trying to run a simulation with the highest model top (p_top_requested) setting possible when initializing with HRRR data for a high altitude modeling test case.

Based on some documentation I've read, there is a discrepancy of what the actual model top is for the HRRR. Some sources state it is 15 hPa ( ... v4aab.pdf; ... chMemo.pdf), while other sources state it is 50 hPa (NOMADS).

When using HRRR isobaric data from NOMADS, I ran a test case to see if I could run a simulation with a p_top_requested setting of 1500. Upon testing, real.exe failed stating that 15 hPa or p_top_requested = 1500 is not available in the input data. Upon further investigation, the highest level (physically) in these input data is 5000 Pa (50mb/50hPa) which is confirmed on NCEP's site as well as when examining the metadata. Real.exe also only ran when p_top_requested was 5000 Pa (50mb/50hPa) or greater (6000, 7000, 8000). Other sites such as AWS, Google, Utah Archive, and THREDDS point to similar conclusions for the model top height of the HRRR too.

However, I also tried running another test case, but instead of using HRRR isobaric data, I used HRRR native model levels with p_top_requested = 1500. It ran successfully. I also tried a simulation running with p_top_requested = 1000 (which should not be possible) and it ran successfully.

Has anyone noticed this before or come across this issue?

My questions are:

1.) What is the actual model top of the HRRR? (15 hPa? 50 hpa?)

2.) Why did real.exe run when p_top_requested = 1000? This should be impossible considering this exceeds what *I think* the model top of the HRRR is. Does this mean there is a glitch in the source code that somehow allows this to happen?

3.) According to this post: viewtopic.php?f=6&t=834 that p_top_requested must be set to a level less than the model top of the input data. So for example, say the HRRR's model top is 15 hPa, according to this post, the p_top_requested setting should equal 16 - which would give me the highest possible model level. Why does real.exe still work even when set to 15 hPa (and even 10 hPa)?

4.) Could anyone provide some recommendations on high altitude modeling optimization settings to use when running WRF? I've considered using more vertical levels and concentrating the eta levels near critical areas in the upper atmosphere.

Thank you!
Posts: 1
Joined: Mon Jun 14, 2021 11:55 am

Return to Dynamics and Numerics

Who is online

Users browsing this forum: No registered users and 2 guests