Backward trajectories from Nottingham

[fherreazcue]

Backward trajectories from Nottingham at

• 1200 UTC 1 July (60–0 h),
• 1800 UTC 1 July (66–0 h),
• 1800 UTC 3 July (114–54 h, 114–0 h), and
• 0000 UTC 4 July (120–54 h, 120–0 h).

Ending locations at 950, 925, 900, 850, 800, 750, 700, 650, 600, 500, 400, 300, and 250 hPa.

Simulations: Control, ALB-40, ALB-Std, ALB-90

Plots and CSV file with following columns:

• time in model simulation
• pressure
• latitude
• longitude
• height above mean sea level
• height above model terrain
• actual air temperature
• potential temperature
• equivalent potential temperature
• dewpoint temperature
• mixing ratio
• relative humidity

[fherreazcue]

The diagnostic names that can in principle be extracted from rip

• time in model simulation
• pressure: prs: Pressure, mb. (3D)
• ? latitude xlat: Latitude (if available as a separate RIP data file). (2D)
• ? longitude xlon: Longitude (if available as a separate RIP data file). (2D)
• ? height above mean sea level ght: Geopotential height above mean sea level, m. (3D)
• ? height above model terrain ghtagl: Geopotential height above ground level, m. (3D) ter: Terrain height, m. (2D)
• actual air temperature tmk: Temperature, K. (3D)
• potential temperature the: Potential temperature, K. (3D)
• equivalent potential temperature eth: Equivalent potential temperature, K. (3D)
• dewpoint temperature tdk: Dew point temperature, K. (3D)
• ? mixing ratio qci: Cloud ice mixing ratio, g kg-1. (3D). -- Note: If non-mixed-phase microphysics was used, this field contains the model output cloud water only where T < 0 °C, and is zero where T > 0 °C. qcl: Total cloud mixing ratio, g kg-1. (3D). -- Note: This is just qcw + qci. qcw: Cloud water mixing ratio, g kg-1. (3D). -- Note: If non-mixed-phase microphysics was used, this field contains the model output cloud water where T > 0 °C, and is zero where T < 0 °C. qgr: Graupel mixing ratio (if available as a separate RIP data file), g kg-1. (3D) qpr: Total precipitation hydrometeor mixing ratio, g kg-1. (3D). -- Note: This is just qra + qsn (+ qgr if available). qra: Rain water mixing ratio, g kg-1. (3D). -- Note: If non-mixed-phase microphysics was used, this field contains the model output rain water only where T > 0 °C., and is zero where T < 0 °C. qsn: Snow mixing ratio, g kg-1. (3D). -- Note: If non-mixed-phase microphysics was used, this field contains the model output rain water only where T < 0 °C, and is zero where T > 0 °C. qvp: Water vapor mixing ratio, g kg-1. (3D)
• ? relative humidity rhi: Relative humidity with respect to ice, percent. (3D) rhu: Relative humidity with respect to liquid water, percent. (3D)

[douglowe]

1. Latitude and Longitude can be the xlat / xlon values
2. Height above mean sea level / terrain needs calculating from the ght, I'll post something below (once I've looked it up)
3. mixing ratio I would usually assume should be water vapor mixing ratio (qvp) - but double check this with Dave
4. relative humidity is most commonly expressed with regards to liquid water (rhu), I think that's a safe bet to use

[douglowe]

This is the code from the NCL routine to get the height data:

  if( any( variable .eq. (/"geopt","geopotential","z","height", "height_agl"/) ) ) then
       ; Height [=full geopotentail height / 9.81]
       if(isfilevar(nc_file,"PH")) then
         var = _get_wrf_var(file_handle,"PH",time)
         PHB = _get_wrf_var(file_handle,"PHB",time)
         var = var + PHB
         z = wrf_user_unstagger(var,var@stagger)
         z@description = "Geopotential"
       elseif(isfilevar(nc_file,"GHT")) then
         ;; may be a met_em file - see if we can get GHT - data in met_em file is Height in M
         z = _get_wrf_var(file_handle,"GHT",time)
         z = z * 9.81
         z@description = "Geopotential"
         z@units       = "m2 s-2"
       else
         print("wrf_user_getvar: Error: cannot find 'PH' or 'GHT' on the file")
         return
       end if

       if( any( variable .eq. (/"z","height", "height_agl"/) ) ) then
         z = z / 9.81
         z@description = "Height"
         if (variable .eq. "height_agl") then
           if (isfilevar(nc_file,"HGT") ) then
             ter = _get_wrf_var(file_handle, "HGT", time)

             rank := dimsizes(dimsizes(ter))

             if (rank .eq. 2) then
                ter_conform = conform_dims(dimsizes(z), ter, (/1,2/))
             elseif (rank .eq. 3) then
                ter_conform = conform_dims(dimsizes(z), ter, (/0,2,3/))
             elseif (rank .eq. 4) then
                ter_conform = conform_dims(dimsizes(z), ter, (/0,1,3,4/))
             elseif (rank .eq. 5) then
                ter_conform = conform_dims(dimsizes(z), ter, (/0,1,2,4,5/))
             elseif (rank .eq. 6) then
                ter_conform = conform_dims(dimsizes(z), ter, (/0,1,2,3,5,6/))
             end if

             z = z - ter_conform
             z@description = "Height_AGL"
           else
             print("wrf_user_getvar: Error: cannot find 'HGT' in the file")
             return
           end if 
         end if 
         z@units       = "m"
       end if
       return(z)
  end if

[douglowe]

Looks like the process to get height above sea level is: 1) Add PH + PHB 2) Unstagger the grid of the resultant data 3) Divide by gravitational acceleration rate (9.81 m/s^2)

[douglowe]

Ahh - looks like RIP already returns GHT though - so perhaps just use that.

(I had assumed these were wrfout variables listed here, but perhaps RIP does this postprocessing work too)

[fherreazcue]

From Dave: """ Can you output both ght and ghtagl? By mixing ratio, I implied water vapor mixing ratio, so variable qvp. Rather than tdk and tmk, could you do tdp and tmc? I forgot about all those variables. As long as we're doing that, could you output these, too? lcl, lfc, omg, pvm, pvo, sateth, stb, stbe, stbz, tdd, wsp, wdr, www Looks like you'll be able to get CIN and CAPE easily (cin3 and cap3, mcap and mcin). """

[fherreazcue]

All diagnostics added in https://github.com/UoMResearchIT/SpanishPlumeAnalysis/commit/beef8cfefe374e55bbb6f63ac51469313e6ed38a, ready to be run on csf. However, some issues with memory were encountered on local tests (see this comment), so it might be that we need to run the trajectory calculation twice to get all the desired diagnostics (running the trajectory with only the second half of diagnostics does work, so it must be that it runs out of memory). Before changing the code though, I need to see if the same problem is encountered when ran on the csf.

[fherreazcue]

The issue persisted on the csf, so it was necessary to split the run in two (see https://github.com/UoMResearchIT/SpanishPlumeAnalysis/commit/7d0572a345cab453a5da91e6b9d33c80563f039c and https://github.com/UoMResearchIT/SpanishPlumeAnalysis/commit/2d6adab34a3be94b18d3cbd5b8187607288e88d8). Trajectories have been generated, with outputs for csvs split in files with g1 or g2 in the name, and the results are in dropbox.

[fherreazcue]

CSV files along trajectories of 90h ran successfully, results are in Dropbox.