RegCM中测试NNRP1强迫场报错

彭小清

在RegCM4.1中，同一个研究区域，用ERA40的数据通过了测试，但是把这个脚本用在NNRP1的数据中（当然其中的数据类型改成了NNRP1），转到第16天就报错了。错误是：求指导
ATM variables written at   2010011600  0.000000000000000E+000
SRF variables written at   2010011600  0.000000000000000E+000
RAD variables written at   2010011600  0.000000000000000E+000
SEARCH BC data for 2010011606
READY  BC from     2010011600 to 2010011606
soil balance e nstep =     10806 point =   215 imbalance =    0.00 W/m2
     at day =   15.0340, ktau =      21650 :  1st, 2nd time deriv of ps =  0.39901E-03 0.27821E-05,  no. of points w/convection =       0
soil balance e nstep =     10835 point =   244 imbalance =    0.00 W/m2
     at day =   15.0687, ktau =      21700 :  1st, 2nd time deriv of ps =  0.41244E-03 0.31605E-05,  no. of points w/convection =       0
soil balance e nstep =     10859 point =   244 imbalance =    0.00 W/m2
soil balance e nstep =     10868 point =   244 imbalance =    0.00 W/m2
WHUUUUBBBASAAAGASDDWD!!!!!!!!!!!!!!!!
No more atmosphere here....
CFL violation detected, so model STOP
#####################################
#            DECREASE DT !!!!       #
#####################################
  -------------- FATAL CALLED ---------------
  Fatal in file: mod_tendency.F90 at line:   2291
  CFL VIOLATION
  -------------------------------------------
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.
NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------

letsdo8

得顺便贴上你的区域（中心点、大小、投影方式等），那些也有影响。

letsdo8

这样才容易得到帮助。

彭小清

这是namelist

! Chinese domain
&dimparam
iy     = 30,
jx     = 32,
kz     = 14,
nsg    = 1,
/
&geoparam
iproj = 'LAMCON',
ds = 20.0,
ptop = 5.0,
clat = 38.0,
clon = 103.0,
plat = 38.0,
plon = 103.0,
truelatl = 36.0,
truelath = 40.0,
i_band = 0,
/
&terrainparam
domname = 'SYH',
ntypec = 10,
ntypec_s = 10,
smthbdy = .false.,
lakedpth = .false.,
fudge_lnd   = .false.,
fudge_lnd_s = .false.,
fudge_tex   = .false.,
fudge_tex_s = .false.,
h2opct = 75.,
dirter = '/input15',
inpter = '/GLOBEDAT',
/
&ioparam
ibyte = 4,
/
&debugparam
debug_level = 1,
dbgfrq = 3,
/
&boundaryparam
nspgx  = 12,
nspgd  = 12,
/
&modesparam
nsplit = 2,
/
&globdatparam
ibdyfrq = 6,
ssttyp = 'OI_WK',
dattyp = 'NNRP1',
ehso4 = .false.,
globidate1 = 2010020100,
globidate2 = 2010123118,
dirglob = '/input15',
inpglob = '/GLOBEDAT',
/
&globwindow
lat0 = 0.0
lat1 = 0.0
lon0 = 0.0
lon1 = 0.0
/
&aerosolparam
aertyp = 'AER00D0'
ntr = 10,
nbin = 4,
/
&restartparam
ifrest  = .false. ,
idate0  = 2010020100,
idate1  = 2010020100,
idate2  = 2010123118,
/
&timeparam
radfrq  =    30.,
   abemh   =    18.,
abatm   =   120.,
dt      =   30.,
/
&outparam
ifsave  = .true. ,
  savfrq  =    7200.,
iftape  = .true. ,
tapfrq  =     6.,
ifrad   = .true. ,
   radisp  =     6.,
ifbat   = .true. ,
ifsub   = .true. ,
   batfrq  =     3.,
ifchem  = .false.,
   chemfrq =     6.,
dirout = '/output15',
/
&physicsparam
iboudy  =          5,
ibltyp  =          1,
icup    =          2,
   igcc    =          2,
ipptls  =          1,
iocnflx =          2,
ipgf    =          0,
iemiss  =          0,
lakemod =          0,
ichem   =          0,
scenario = 'A1B',
idcsst = 0,
iseaice = 0,
idesseas = 1,
iconvlwp = 1,
/
&subexparam
qck1land=   .250E-03,
qck1oce =   .250E-03,
cevap   =   .100E-02,
caccr   =      3.000,
cftotmax =      0.75,
/
&grellparam
/
&emanparam
elcrit  = 0.0011D0,
coeffr  = 1.0D0,
/
  &chemparam
idirect   =    1,
inchtrname  = '',
inchtrsol   =   .00,
inchtrdpv   =  .00000, .00000,
industbsiz  =   .00,  .00,
/
  &clmparam
dirclm = '/input15',
imask  = 1,
clmfrq = 12.,
/

letsdo8

It seems to me that RegCM 4.1 is too much unstable.

I would suggest try RegCM 4.3.5.6, or just RegCM 4.0.

Your domain is too small, not practical, I think.

Use such a small domain and 12 grid of LBC buffer zone is meaningless.

letsdo8

And check the NNRP1 data you dowloaded.

彭小清

letsdo8 发表于 2014-1-3 11:10

                               
登录/注册后可看大图

And check the NNRP1 data you dowloaded.

我用的是RegCM4.1  这个NNRP1的数据是从它的官网下载的啊   你说的domain太小了啥意思（你是说研究区域的范围吗还是。。。）

彭小清

彭小清 发表于 2014-1-3 21:51

                               
登录/注册后可看大图

我用的是RegCM4.1  这个NNRP1的数据是从它的官网下载的啊   你说的domain太小了啥意思（你是说研究区域的 ...

大侠 我又重新下载了一遍数据，目前还在运行之中   还没有报错  估计是数据的问题  我下载的是1990年，前满报错用的是2009,2010年的数据。

letsdo8

Probably, the subroutine cdc6hour, which read NNRP1 netcdf files, should be modified.

彭小清

letsdo8 发表于 2014-1-4 20:52

                               
登录/注册后可看大图

Probably, the subroutine cdc6hour, which read NNRP1 netcdf files, should be modified.

可能吧   大侠有空帮忙看一下这个帖子的错误  http://bbs.06climate.com/forum.php?mod=viewthread&tid=19951谢啦