evaluation of WRFV4-SUEWS results

[sunt05]

@hamidrezaomidvar you can use this script for analysis: https://github.com/Urban-Meteorology-Reading/WRF-SUEWS/blob/master/post-processor/WRF-SUEWS-ana.py

[hamidrezaomidvar]

Thanks. there were a couple of bugs in the automated Json which I am fixing now, and will push it later. Please use the updated version for tets

[hamidrezaomidvar]

Ting, do you have namelist.wps of the London domain that I used for the run? I think the python file needs that one!

[hamidrezaomidvar]

Here are some results:

[sunt05]

Ting, do you have namelist.wps of the London domain that I used for the run? I think the python file needs that one!

It's here: https://github.com/Urban-Meteorology-Reading/WRF-SUEWS/blob/master/wrf_input/London-3domains/namelist.wps-london-3

[sunt05]

Here are some results:

Good! Re: getting results out!

Let's focus on the inner-most domain, d03, first. Have you had a look at the radiation data? I know there's some issue with the incoming solar radiation simulated by WRF, leading to poorly simulated heat fluxes. Let's resolve this issue first.

[hamidrezaomidvar]

Ok, I will look at the radiation data and post the results here

[suegrimmond]

Excellent

Best wishes Sue

Prof Sue Grimmond Meteorology, University of Reading

---

From: Hamidreza Omidvar notifications@github.com Sent: Thursday, March 14, 2019 4:58:11 PM To: Urban-Meteorology-Reading/WRF-SUEWS Cc: Subscribed Subject: Re: [Urban-Meteorology-Reading/WRF-SUEWS] evaluation of WRFV4-SUEWS results (#27)

Ok, I will look at the radiation data and post the results here

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/Urban-Meteorology-Reading/WRF-SUEWS/issues/27#issuecomment-472962869, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ATSs3tkXqzUOEWTvapaoXh4olMWQaMqeks5vWn-jgaJpZM4bzscy.

[hamidrezaomidvar]

Ting, when you want to plot maps from wrf outputs, what do you usually use? it seems when I want to plot them, I get Segmentation fault (core dumped) which is a memory problem I assume

[sunt05]

xarray: multi-dimensional pandas

salem: I feel this should be the WRF companion in Python.

[sunt05]

Ting, when you want to plot maps from wrf outputs, what do you usually use? it seems when I want to plot them, I get Segmentation fault (core dumped) which is a memory problem I assume

In particular, I find this is very useful for our scenario: https://salem.readthedocs.io/en/v0.2.3/plotting.html#plotting-with-salem

[hamidrezaomidvar]

Here are some of the plots of Radiation components. https://www.dropbox.com/sh/5jomvq84awm3eps/AAALXSEqIihEVLNpCXbvDXnFa?dl=0 I am plotting some maps and will update it here

[hamidrezaomidvar]

[hamidrezaomidvar]

[sunt05]

Great! Would be good to compare the simulations against the observations in London so we know how well/poorly the coupled model is doing.

[hamidrezaomidvar]

Ok, I will start working on it!

[hamidrezaomidvar]

Quick question. I was looking at the London data you gave me, and tried to calculate the net all wave radiation (Kdn-Kup+Ldn-Kup), and compared with Qstar in the data (which is net all wave radiation). Note that I am talking about the data, not the WRF model. Here is what I get. why is there this gap between them? am I missing something?

[hamidrezaomidvar]

never mind, I had a mistake in calculations

[suegrimmond]

Lup are you using that?

Best wishes Sue

Prof Sue Grimmond Meteorology, University of Reading

---

From: Hamidreza Omidvar notifications@github.com Sent: Monday, March 18, 2019 5:57:44 PM To: Urban-Meteorology-Reading/WRF-SUEWS Cc: Sue Grimmond; Comment Subject: Re: [Urban-Meteorology-Reading/WRF-SUEWS] evaluation of WRFV4-SUEWS results (#27)

never mind, I had a mistake in calculations

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/Urban-Meteorology-Reading/WRF-SUEWS/issues/27#issuecomment-474032446, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ATSs3i3uk7BqSgbU-5nAPQr2x8kV0w2Yks5vX9OYgaJpZM4bzscy.

[hamidrezaomidvar]

Yes that was the problem. I mistakenly was using Kup instead

On Mar 18, 2019, at 6:28 PM, Sue Grimmond notifications@github.com wrote:

Lup are you using that?

Best wishes Sue

Prof Sue Grimmond Meteorology, University of Reading

---

From: Hamidreza Omidvar notifications@github.com Sent: Monday, March 18, 2019 5:57:44 PM To: Urban-Meteorology-Reading/WRF-SUEWS Cc: Sue Grimmond; Comment Subject: Re: [Urban-Meteorology-Reading/WRF-SUEWS] evaluation of WRFV4-SUEWS results (#27)

never mind, I had a mistake in calculations

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/Urban-Meteorology-Reading/WRF-SUEWS/issues/27#issuecomment-474032446, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ATSs3i3uk7BqSgbU-5nAPQr2x8kV0w2Yks5vX9OYgaJpZM4bzscy. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[hamidrezaomidvar]

Here are two plots comparing downward SW and LW for the data and the model:

[suegrimmond]

What cloud physics is being used? Do we understand why the first Longwave is always 0?

[hamidrezaomidvar]

Here is the physics part of the namelist.input

 physics_suite                       = 'CONUS'
 radt                                = 30,    30,    30,
 bldt                                = 0,     0,     0,
 cudt                                = 5,     5,     5,
 icloud                              = 1,
 num_soil_layers                     = 4,
 num_land_cat                        = 21,
 sf_urban_physics                    = 0,     0,     0,
 sf_surface_physics                  = 9,     9,     9,
 cu_physics                          = 6,     0,     0,

It seems that icloud=1, but from the documentation, this just works when:

cloud effect to the optical depth in radiation (only works for ra_sw_physics = 1 and ra_lw_physics = 1) 0 = without cloud effect 1 = with cloud effect

But in the physics_suit='CONUS', we have:

where 'CONUS' is equivalent to

 mp_physics         = 8,
 cu_physics         = 6,
 ra_lw_physics      = 4,
 ra_sw_physics      = 4,
 bl_pbl_physics     = 2,
 sf_sfclay_physics  = 2,
 sf_surface_physics = 2,

@sunt05 am I right about this?

[suegrimmond]

I suggest you find a clear day in the observations - and that we runs that period intiailly -so we can check how things are looking before we deal with the clouds. We do have a list of a clear days - so we can use that

[hamidrezaomidvar]

I see! I will do the run for the clear days. @sunt05 I might have some question about the process of using WPAS to create WRF input for these clear days. I will let you know

[sunt05]

@hamidrezaomidvar feel free to call me on Skype if it would be easier.

[hamidrezaomidvar]

@sunt05 after creating the geo_em.nc files for the domain, how do you create the input files (time series of atmospheric data) for the London data?

[sunt05]

have you got met_em*.nc?

[sunt05]

Please go through this section: http://www2.mmm.ucar.edu/wrf/users/docs/user_guide_v4/v4.0/users_guide_chap3.html#_How_to_Run

[hamidrezaomidvar]

No. Should I create them using the gfs data of here: https://ftp.ncep.noaa.gov/data/nccf/com/gfs/prod/ ?

[sunt05]

No. Should I create them using the gfs data of here: https://ftp.ncep.noaa.gov/data/nccf/com/gfs/prod/ ?

No, use this one: https://rda.ucar.edu/datasets/ds083.2/

[sunt05]

Once you get met_em*.nc, do this: http://www2.mmm.ucar.edu/wrf/users/docs/user_guide_v4/v4.0/users_guide_chap4.html#Real

[sunt05]

And, check out this link for various forcing data sources: http://www2.mmm.ucar.edu/wrf/users/download/free_data.html

[hamidrezaomidvar]

Awesome! thanks. I registered to get the data. Once approved, will work on it and let you know if I get to any problems!

[sunt05]

and for data transfer between these academic servers, check out GLOBUS: both RDA and Jasmin support this.

[hamidrezaomidvar]

Quick question: what are the advantages/disadvantages of using 2-way nesting+feedback=0 (basically one-way) over using ndown.exe for running a 1-way nest?

[sunt05]

short answer: ndown.exe-based method is tedious. You need to run this program to get refined results; while one-way nesting (your 2-way nesting+feedback=0) is done on the fly when running WRF.

[hamidrezaomidvar]

Here is the clear sky run and its comparison with the data:

[sunt05]

Nice! Can we see results of more days? Also other components (Ldown, Luo, Q*, Kup, etc. )?

[hamidrezaomidvar]

Ok, I will plot other components. Also I will run for other clear days to see how they look like!

[hamidrezaomidvar]

Here are more results from the same day:

[sunt05]

Great! Then we know something wrong with the albedo set for the grid. How about LW_net?

[hamidrezaomidvar]

Here is the Ldn-Lup

[sunt05]

Great! Now let's see the results of longer series.

And Let's open the abnormal first value Ldown as a separate issue to deal with later; but do plots excluding the first value of LW-related variables.

[hamidrezaomidvar]

Ok. I am running for three more clear days. one 04/14, 04/21, and 06/11 (all 2015). I will post the results once they finish

[sunt05]

Ok. I am running for three more clear days. one 04/14, 04/21, and 06/11 (all 2015). I will post the results once they finish

Updated: Sorry @hamidrezaomidvar , I meant 24 Apr Ok, let's run 14 days for the April period: 10 Apr – 24 Apr as we need some spin-up time.

[hamidrezaomidvar]

Got it! do you mean Apr 1-14?

[hamidrezaomidvar]

Here are the results after the long series run from April 10 to 24: There were 3 clear days in April 2015. 14,15,21

Here are 14 and 15 results:

[hamidrezaomidvar]

and here are 21st results:

[sunt05]

Great! Can we plot all days on the same figure, i.e., including both cloudy and clear days but with different markers?

[hamidrezaomidvar]

Here are all the days: