Problems for a 50-Year (2015-2065) WRF Dynamical Downscaling Simulation

[feifeiyu1225]

Hi everyone, I'm planning to conduct a 50-year (2015-2065) WRF dynamical downscaling simulation, and I have a few questions:

1. How should I divide the simulation period? Should I run it in smaller chunks, and if so, how long should each segment be?
2. What should be the spin-up time for such a long-term simulation?
3. My target region has latitude and longitude intervals of around 20°. Should I replace WRF's default data with 90m resolution SRTM data?

Thanks for your insights!

 

[Ming Chen]

Please see my answers below:

Hi everyone, I'm planning to conduct a 50-year (2015-2065) WRF dynamical downscaling simulation, and I have a few questions:

1. How should I divide the simulation period? Should I run it in smaller chunks, and if so, how long should each segment be?

There is no strict rule how long a segment should be. You can run year-by-year or month-by-month, using the restart capability of WRF.

1. What should be the spin-up time for such a long-term simulation?

For long-term simulation, lateral boundary forcing is more important. Therefore, the spin-up time of 9-12 hours will be enough.

1. My target region has latitude and longitude intervals of around 20°. Should I replace WRF's default data with 90m resolution SRTM data?

No, it is not necessary.

Thanks for your insights!

 

[Alfred-Jose]

Hi all,
Can you clear my doubt on running such long simulation runs, don't we need the soil moisture or other properties to warmed up before starting the run?
How do we decide the Spin-up period actually?

And don't we need to do Spectral Nudging in the Outer Domain, during the run?

Thank you

 

[Ming Chen]

@ Alfred-Jose
You may run offline LSM to create more realistic soil moisture and/or other properties, then use them as the initial condition for your long-term climate simulation. There is no strict rule on how to determine the spin-up time. Some people run LSM for 1-yr or 1-yr plus one extra season to spin-up the soil condition. The other simpler option is to use soil data from reanalysis products.

If your outer domain is large enough (>1500–2000 km across) and your resolution is coarse ( e.g., >10km), you will need to turn on spectral nudging.