Opened 3 years ago

Closed 3 years ago

Last modified 3 years ago

#284 closed Support (fixed)

Strange big bouncing of particles when hitting ground

Reported by: fcheng Owned by: pesei
Priority: major Milestone:
Component: FP other Version: FLEXPART 10.4
Keywords: Cc:

Description

Dear FLEXPART community,

I observe that the particles released by FLEXPART have seemingly overexaggerated bouncing behavior when they are hitting the ground, sometimes they can be reflected by a few km or even up to 10km in just one hour. This really confuses me since it just does not make sense.

I think the source code that controls the bouncing of particles is related to advance.f90, and maybe also the default setting of SPECIES 24. But I just could not really identify where the problem comes from.

For your information, here is my RELEASES file

&RELEASES_CTRL
 NSPEC      =           1, ! Total number of species
 SPECNUM_REL=          24, ! Species numbers in directory SPECIES
 /
&RELEASE                   ! For each release
 IDATE1  =       20080601, ! Release start date, YYYYMMDD: YYYY=year, MM=month, DD=day
 ITIME1  =          00000, ! Release start time in UTC HHMISS: HH hours, MI=minutes, SS=seconds
 IDATE2  =       20080601, ! Release end date, same as IDATE1
 ITIME2  =          00000, ! Release end time, same as ITIME1
 LON1    =       -179.000, ! Left longitude of release box -180 < LON1 <180
 LON2    =        179.000, ! Right longitude of release box, same as LON1
 LAT1    =        -89.000, ! Lower latitude of release box, -90 < LAT1 < 90
 LAT2    =         89.000, ! Upper latitude of release box same format as LAT1
 Z1      =          0.000, ! Lower height of release box meters/hPa above reference level
 Z2      =       5000.000, ! Upper height of release box meters/hPa above reference level
 ZKIND   =              1, ! Reference level 1=above ground, 2=above sea level, 3 for pressure in hPa
 MASS    =       1.0000E0, ! Total mass emitted, only relevant for fwd simulations
 PARTS   =        6000000, ! Total number of particles to be released
 COMMENT =    "RELEASE 1 test, particles released in an hour.", ! Comment, written in the outputfile
 /

And my COMMAND

&COMMAND
 LDIRECT=               1, ! Simulation direction in time   ; 1 (forward) or -1 (backward)
 IBDATE=         20080601, ! Start date of the simulation   ; YYYYMMDD: YYYY=year, MM=month, DD=day
 IBTIME=            00000, ! Start time of the simulation   ; HHMISS: HH=hours, MI=min, SS=sec; UTC
 IEDATE=         20080601, ! End date of the simulation     ; same format as IBDATE
 IETIME=           210000, ! End  time of the simulation    ; same format as IBTIME
 LOUTSTEP=           3600, ! Interval of model output; average concentrations calculated every LOUTSTEP (s)
 LOUTAVER=           3600, ! Interval of output averaging (s)
 LOUTSAMPLE=          900, ! Interval of output sampling  (s), higher stat. accuracy with shorter intervals
 ITSPLIT=        99999999, ! Interval of particle splitting (s)
 LSYNCTIME=           900, ! All processes are synchronized to this time interval (s)
 CTL=          -5.0000000, ! CTL>1, ABL time step = (Lagrangian timescale (TL))/CTL, uses LSYNCTIME if CTL<0
 IFINE=                 4, ! Reduction for time step in vertical transport, used only if CTL>1
 IOUT=                  1, ! Output type: [1]mass 2]pptv 3]1&2 4]plume 5]1&4, +8 for NetCDF output
 IPOUT=                 1, ! Particle position output: 0]no 1]every output 2]only at end 3]time averaged
 LSUBGRID=              0, ! Increase of ABL heights due to sub-grid scale orographic variations;[0]off 1]on
 LCONVECTION=           1, ! Switch for convection parameterization;0]off [1]on
 LAGESPECTRA=           0, ! Switch for calculation of age spectra (needs AGECLASSES);[0]off 1]on
 IPIN=                  0, ! Warm start from particle dump (needs previous partposit_end file); [0]no 1]yes
 IOUTPUTFOREACHRELEASE= 0, ! Separate output fields for each location in the RELEASE file; [0]no 1]yes
 IFLUX=                 0, ! Output of mass fluxes through output grid box boundaries
 MDOMAINFILL=           1, ! Switch for domain-filling, if limited-area particles generated at boundary
 IND_SOURCE=            1, ! Unit to be used at the source   ;  [1]mass 2]mass mixing ratio
 IND_RECEPTOR=          1, ! Unit to be used at the receptor; [1]mass 2]mass mixing ratio 3]wet depo. 4]dry depo.
 MQUASILAG=             1, ! Quasi-Lagrangian mode to track individual numbered particles
 NESTED_OUTPUT=         0, ! Output also for a nested domain
 LINIT_COND=            0, ! Output sensitivity to initial conditions (bkw mode only) [0]off 1]conc 2]mmr
 SURF_ONLY=             0, ! Output only for the lowest model layer, used w/ LINIT_COND=1 or 2
 CBLFLAG=               0, ! Skewed, not Gaussian turbulence in the convective ABL, need large CTL and IFINE
 OHFIELDS_PATH= "../../flexin/", ! Default path for OH file
 /

And, here is the hourly altitude (meters a.g.l.) of a particle that had a big bounce (and there are many particles with such a behavior)

397.4110413
40.72389984
537.1908569
289.5689392
937.1464233
323.7579346
98.35718536
138.9232025
74.352005
155.2883759
11351.6709
11257.2832
11149.76074
11025.70117
10902.66504
10848.78516
10827.65918
10822.45605
10803.74316
10752.75
10698.94824

Is such a big bouncing behavior normal? Any hints and help are much appreciated. Thanks a lot!

Warm regards,
Fandy

Change History (4)

comment:1 Changed 3 years ago by pesei

  • Owner set to pesei
  • Status changed from new to accepted

I see that you have convection activated. Therefore I assume that what you see is a particle lifted in convection from 155 m to 11351 m. I do not see "bouncing" in your data.

comment:2 Changed 3 years ago by fcheng

Thank you for the reply. From my experience analyzing the particle's position, I found that such an abrupt jump often occurred near the ground level, that was why I thought the particle might hit the ground and somehow trigger the bouncing motion.

I think for a particle that travels vertically by 10km in just one hour seems a bit strange. I will check the results again by turning the convection scheme off. Thanks a lot for the hint.

Cheers,
Fandy

comment:3 Changed 3 years ago by pesei

  • Resolution set to fixed
  • Status changed from accepted to closed

It is not strange, it is how convection works. And penetrating convection starts near the surface. However, if you think that too many particles are affected, and if you see this happen even if you turn convection off, then please reopen the ticket.

comment:4 Changed 3 years ago by fcheng

Well understood. Thanks a lot for the comment. I really appreciate it.

Best,
Fandy

Note: See TracTickets for help on using tickets.
hosted by ZAMG