WRFDA使用CV7背景误差gen_be出现问题

Nozigaka46

在WRFDA同化的时候，选择了cv_option=7的选项，然后出现一下问题，同时也导致同化后的VARBC.out时NAN空值

(base) [lwang1@wxh_server RANDMCV]$ tail rsl.error.0000
     Input water vapour profile exceeds lower coef limit (profile number =        1)
2023/09/04  16:17:00  Limit   =     1.2912
2023/09/04  16:17:00  p (hPa) =   334.8650
2023/09/04  16:17:00  Value   =     0.5691
2023/09/04  16:17:00  rttov_checkinput.F90
     Input water vapour profile exceeds lower coef limit (profile number =        1)
2023/09/04  16:17:00  Limit   =     1.2912
2023/09/04  16:17:00  p (hPa) =   334.8650
2023/09/04  16:17:00  Value   =     0.5691


请问如何这是什么原因导致的或者说该如何解决？


             这是namelist.input的设置：
&wrfvar1
print_detail_grad=.false.,
/
&wrfvar2
analysis_accu=3600,
/
&wrfvar3
ob_format=1,
/
&wrfvar4
use_mhsobs=true,
/
&wrfvar5
put_rand_seed=false,
/
&wrfvar6
/
&wrfvar7
je_factor = 2.0,
cv_options=7,
/
&wrfvar8
/
&wrfvar9
/
&wrfvar10
/
&wrfvar11
seed_array1=1,
seed_array2=1,
/
&wrfvar12
/
&wrfvar13
/
&wrfvar14
rtminit_nsensor=1,
rtminit_platform=1,
rtminit_satid=19,
rtminit_sensor=15,

thinning_mesh=80.0,
thinning=true,
qc_rad=true,
write_iv_rad_ascii=false,
write_oa_rad_ascii=true,
rtm_option=1,
use_varbc=true,
use_crtm_kmatrix=true,
use_rttov_kmatrix=false,
rttov_emis_atlas_ir=0,
rttov_emis_atlas_mw=0,
crtm_coef_path = "/public2/users/lwang1/fzl/WRFDAdomain/wrfdatest/crtm_coeffs"
/
&wrfvar15
/
&wrfvar16
ensdim_alpha = 0,               
alphacv_method = 2,
alpha_corr_type=3,
alpha_corr_scale = 500.0,
alpha_std_dev=1.000,
alpha_vertloc_opt = 2,
ep_format = 1,
/
&wrfvar17
analysis_type="QC-OBS",
n_randomcv=5,
/
&wrfvar18
analysis_date="2020-04-01_02:00:00.0000",
/
&wrfvar19
/
&wrfvar20
/
&wrfvar21
time_window_min="2020-04-01_01:00:00.0000",
/
&wrfvar22
time_window_max="2020-04-01_04:00:00.0000",
/
&time_control
start_year=2020,
start_month=04,
start_day=01,
start_hour=02,
end_year=2020,
end_month=04,
end_day=01,
end_hour=02,
/
&fdda
/
&domains
e_we=347,
e_sn=324,
e_vert=40,
dx=3000,
dy=3000,
/
&dfi_control
/
&tc
/
&physics
mp_physics=6,
ra_lw_physics=1,
ra_sw_physics=1,
radt=10,
sf_sfclay_physics=1,
sf_surface_physics=2,
bl_pbl_physics=1,
cu_physics=0,
cudt=5,
num_soil_layers=4,
mp_zero_out=2,
co2tf=0,
/
&scm
/
&dynamics
/
&bdy_control
/
&grib2
/
&fire
/
&namelist_quilt
/
&perturbation
/

任性小王子

您好，楼主，您使用CV7格式，新NCAR的背景误差模型是由GEN_BE制作的吗？楼主用的什么观测数据，怎么制作背景误差模型，可以分享一下经验吗？