Problem in wrfxtrm file

Hello all!
The temperature unit (T2MAX, T2MEAN, T2MIN) seems to totally mess up in my wrfxtrm output. In some timestep, it seems to be in K (the value is over 273.15) and in some timestep, the value is just 20 something so i assume it is calculated in degreeC. The temperature unit is different from one timestep to another. And i have no idea what cause this problem.
By the way, i am turning on the adaptive timestep, will this cause such problem? but the wrfout* file seems fine.

here is my namelist.input for reference.
&time_control
run_days = 0,
run_hours = 0,
run_minutes = 0,
run_seconds = 0,
start_year = 2011, 2011, 2000,
start_month = 05, 05, 06,
start_day = 28, 28, 24,
start_hour = 12, 12, 12,
end_year = 2011, 2011, 2000,
end_month = 07, 07, 01,
end_day = 01, 01, 25,
end_hour = 00, 00, 12,
interval_seconds = 21600
input_from_file = .true.,.true.,.true.,
history_interval = 360, 180, 60,
frames_per_outfile = 400, 400, 1000,
restart = .false.,
restart_interval = 14400,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
io_form_auxinput4 = 2
io_form_auxinput2 = 2
auxinput4_inname ="wrflowinp_d<domain>"
auxinput4_interval = 360
override_restart_timers = .true.
adjust_output_times = .true.
output_diagnostics = 1
auxhist3_outname = "wrfxtrm_d<domain>_<date>"
auxhist3_interval = 1440, 1440,
frames_per_auxhist3 = 1000, 1000,
io_form_auxhist3 = 2,
/

&domains
time_step = 140,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 2,
e_we = 109, 117, 94,
e_sn = 86, 133, 91,
e_vert = 38, 38, 33,
p_top_requested = 5000,
num_metgrid_levels = 38,
num_metgrid_soil_levels = 4,
dx = 20000, 5000, 3333.33,
dy = 20000, 5000, 3333.33,
grid_id = 1, 2, 3,
parent_id = 0, 1, 2,
i_parent_start = 1, 41, 30,
j_parent_start = 1, 16, 30,
parent_grid_ratio = 1, 4, 3,
parent_time_step_ratio = 1, 4, 3,
feedback = 0,
smooth_option = 0
use_adaptive_time_step = .true.
step_to_output_time = .false.
target_cfl = 1.25, 1.25,
target_hcfl = 0.84, 0.84,
max_step_increase_pct = 5,51,
starting_time_step = 120, 30,
max_time_step = 180, 45,
min_time_step = 40, 10,
/

&physics
mp_physics = 3, 3, -1,
cu_physics = 1, 1, 0,
ra_lw_physics = 24, 24, -1,
ra_sw_physics = 24, 24, -1,
bl_pbl_physics = 1, 1, -1,
sf_sfclay_physics = 1, 1, -1,
sf_surface_physics = 2, 2, -1,
radt = 30, 30, 30,
bldt = 0, 0, 0,
cudt = 5, 5, 5,
icloud = 1,
num_land_cat = 21,
sf_urban_physics = 0, 0, 0,
surface_input_source = 3
sst_update = 1
tmn_update = 1
sst_skin = 1
kfeta_trigger = 2
isfflx = 1
/

&fdda
grid_fdda = 2,
gfdda_inname = "wrffdda_d<domain>",
gfdda_end_h = 1000,
gfdda_interval_m = 360,
fgdt = 0,
fgdtzero = 0,
xwavenum = 3,
ywavenum = 3,
if_no_pbl_nudging_uv = 1,
if_no_pbl_nudging_t = 1,
if_no_pbl_nudging_ph = 1,
if_zfac_uv = 0,
k_zfac_uv = 10,
if_zfac_t = 0,
k_zfac_t = 10,
if_zfac_ph = 0,
k_zfac_ph = 10,
guv = 0.0003,
gt = 0.0003,
gph = 0.0003,
io_form_gfdda = 2,
/

&dynamics
hybrid_opt = 2,
w_damping = 0,
diff_opt = 1, 1, 1,
km_opt = 4, 4, 4,
diff_6th_opt = 0, 0, 0,
diff_6th_factor = 0.12, 0.12, 0.12,
base_temp = 290.
damp_opt = 3,
zdamp = 5000., 5000., 5000.,
dampcoef = 0.2, 0.2, 0.2
khdif = 0, 0, 0,
kvdif = 0, 0, 0,
non_hydrostatic = .true., .true., .true.,
moist_adv_opt = 1, 1, 1,
scalar_adv_opt = 1, 1, 1,
gwd_opt = 0,
/

&bdy_control
spec_bdy_width = 5,
specified = .true.
nested = .false.,.true.
/

&grib2
/

&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/

Which version of WRF did you use?
Please turn off adaptive time step and let me know whether the problem still exists. Thanks.

Ming Chen said:

Which version of WRF did you use?
Please turn off adaptive time step and let me know whether the problem still exists. Thanks.

Click to expand...

we are using WRF 4.2.
I turned off the adaptive time step but the problem exists.

Hi,
I just run a test case with the option output_diagnostics = 1. wrfxtrm files look just fine. Note that I turn off sst_update and spectral nudging.
This test indicates that the option of output_diagnostics works as expected. If you get those weird results, it is possibly because something is wrong in the physics.
I would suggest you turn off spectral nudging, because this option is designed for large -domain, long-term simulation, which is not your case.
Other than this, all other options in your namelist.input look fine. Still, I would suggest you turn off some options, e.g., sst_update, sst_skin, etc.fewer options will be helpful to trace what is wrong.

Hi, I have the same problem in the wrfxtrm files.

I noticed that the clWRF statistics files (wrfxtrm) have some erroneous values (eg., surface temperatures of thousands of K, among others). This only happens for some time steps (days, for these files) but when it happens it affects all variables contained in the wrfxtrm files.

I have confirmed in the corresponding wrfout files that, for the same variables and times, these erroneous values are not present.

Any insight on this?

Where did you download the clWRF codes? Can you also tell me whether there is any contact person who is rresponsible for clWRF? Thanks.

The clWRF codes are included in the standard WRF code

No I don't think so. As far as I know, the group that develops clWRF never shared information with us.

It is, you activate it with compilation flags. Been using it for years

output_diagnostics = 1 turns on the option to output 36 surface diagnostic arrays (max/min/mean/std). This is an option in standard WRF, and it is not required to compile WRF with the specific flag for clWRF.