4

I am wondering if anybody has an idea about the difference between regional climate models and numerical weather prediction models that are run at the same domain.

I have a basic idea that the numerical weather prediction models can be run at much higher resolution and due to the time stepping and CFL schemes, it is not always possible to run them for a long time. So I have a general query and I want to infer about the time stepping and CFL criteria used in an RCM and a NWP model. Are they different or they both are similar, it's just that an RCM is run at coarser resolution?

Edit: I want to know if it is possible to get a climate model from a generic NWP model. I have seen the CWRF documentation which illustrates its highlights and mentions some different physical parameterizations corresponding to different processes. I realized that the convergence of model solution is more important even if the time stepping allows to get a long time run. I still have confusions about the new schemes used in CWRF, do they neglect subgrid scale processes like the other climate models? Why are these necessary to be implemented? Here, I also saw the mentioned description about CMM5 but I have the same doubts here. I want to know about the physical parameterizations and numerical schemes used in unified models and how are they different from that of a generic NWP model.

Agni
  • 645
  • 3
  • 9
  • Suspect there may also be differences in what physics is and isn't included, what assumptions are made, what is explicitly modelled vs paramaterised, etc. – Semidiurnal Simon Sep 28 '18 at 13:01
  • I guess you are talking about subgrid scale processes such as convection. This should consume more computational cost. I don't know what changes can be made to those NWP models so that they can be run at climate scale. – Agni Sep 29 '18 at 16:05
  • Why the obsession with the N in NWP? You know that everything calculation run on a computer is a "numerical something", right? – AtmosphericPrisonEscape Sep 30 '18 at 00:10
  • I agree. It was generally called N, so I used it – Agni Sep 30 '18 at 01:30
  • Thanks, it is more informative as I had not seen it before. But apart from model physics, subgrid scale processes and assimilation, I wanted to ask about converting a weather model into a climate one with no change in underlying physics. The complications will arise with time stepping I know, but I want to know if people are working to achieve this goal and what factors do they consider. – Agni Sep 30 '18 at 05:27
  • I still have no clue how the extended and seasonal scale predictions are made. Are weather models used for the seasonal scale in this case? – Agni Sep 30 '18 at 05:32
  • 1
    @nshwal - http://cwrf.umd.edu/ –  Oct 03 '18 at 01:34
  • @nshwal - If you still have that question reframe it terms of how CWRF does it. But that I think is best handled at the CWRF forums. –  Oct 04 '18 at 10:12
  • 1
    @gansub my question seemed ambiguous so I have edited it as per my need. Thanks for the link, I also found a mention about CMM5. Both the CWRF and CMM5 use different physics options but I am unable to get them all in a nutshell. Therefore, I am hoping if anyone can tell me more about unified models that are derived from NWP systems. – Agni Oct 07 '18 at 15:52

0 Answers0