Sensitivity%20Study%20of%20Precipitation%20and%20T2m%20to%20Soil%20Moisture%20Using%20NCEP%20WRF%20Ensemble - PowerPoint PPT Presentation

Title: Sensitivity%20Study%20of%20Precipitation%20and%20T2m%20to%20Soil%20Moisture%20Using%20NCEP%20WRF%20Ensemble


1
Sensitivity Study of Precipitation and T2m to
Soil Moisture Using NCEP WRF Ensemble

• Jun Du and George Gayno
• EMC/NCEP/NOAA

2
In Memory of Prof. Fred Sanders of MIT
(The photo taken at NCEP in July 1994)
3
outline

• Sensitivity study
• Basic physical mechanism understanding
  
• Soil moisture perturbation design (on going)
• Ensemble system improvement (to be done)
• Summary

4
MotivationUnder dispersive issue in warm
season precipitation and near-surface variables
such as T2m
5
Ensemble spread of 12h-apcp from NCEP 21-member
SREF (Jun 2006)
6
( from NMC/CMA, Y. Li)
7
(from NMC/CMA, Y. Li)
8
3 WRF_NMM (NCEP, 40km) and 3 WRF_ARW (NCAR, 45km)
9
Warm season case
10
0-10cm Soil moisture diff (nam-gfs)
NOAH
Diff 0.1
Nam is wetter than gfs in general (except For SE
US and Greenland)
NOAH
11
(WRF_NMM model)
Diff between two NMM 3h-apcp forecasts
With NAM soil moisture
With GFS soil moisture
12
(WRF_ARW model)
Diff of two ARW 3h-apcp forecasts
With NAM soil moisture
With GFS soil moisture
13
Within WRF_NMM model Impact on T2m is
significant!
With nam soil moisture (NMM)
T2m diff (namSM gfsSM, NMM)
With gfs soil moisture (NMM)
14
Within WRF_ARW model Impact on T2m is
significant.
With nam soil moisture (ARW)
T2m diff (namSM gfsSM, ARW)
With gfs soil moisture (ARW)
15
(3h-apcp)
Impact comparison between IC diff and soil
moisture diff within WRF_NMM
note (a) lt0.1 contour not plotted, (b) its
only 3h accumulation
Difference caused by different ICs
Difference caused by different soil moisture
Impact from diff soil moisture is comparable to
that from diff ICs within NMM!
16
(3h-apcp)
Impact comparison between IC diff and soil
moisture diff within WRF_ARW
Difference caused by different ICs
Difference caused by different soil moisture
Impact from diff soil moisture is comparable to
that from diff ICs within ARW!
17
3h-apcp
Impact comparison between model difference and
soil moisture difference
Difference caused by different models
Difference caused by soil moisture diff (nmm)
Impact from soil moisture diff is slightly less
than but still comparable in some areas to that
even from model diff (NMM vs. ARW)!
18
3h-apcp
Impact comparison between model difference and
soil moisture difference
Difference caused by different models
Difference caused by soil moisture diff (arw)
Impact from soil moisture diff is slightly less
than but still comparable in some areas to that
even from model diff (NMM vs. ARW)!
19
T2m -- impact comparison (NMM) Soil moisture
impact on T2m is bigger Than that of IC and of
model difference!
different IC (NMM namSM)
Diff soil moisture (nam gfs NMM)
different models (nmm vs. arw namSM)
20
T2m -- impact comparison (ARW) Soil moisture
impact on T2m is comparable To that of IC but
smaller than model diff
Different IC (ARW, namSM)
Diff soil moisture (nam gfs ARW)
Diff models (nmm vs. arw gfsSM)
21
Sensitivity/response of precip to soil moisture
perturbation varies with model
In WRF_NMM (less sensitive)
In WRF_ARW (more sensitive)
22
Sensitivity/response of T2m to soil moisture
varies with model
In WRF_NMM (more sensitive)
In WRF_ARW (less sensitive)
23
Physical mechanism soil moisture ?
evapotranspiration ?partitioning of net available
surface thermal energy between latent heat and
sensible heat flux (temp and moisture) ?
convection
(a) dry case increased sensible heat will have
an earlier initiation of convection (e.g.
afternoon thunderstorm) (b) wet case in
contrast to dry case, reduced sensible heat
will delay initiation convection (e.g.
late night thunderstorm) (c) perturb soil
moisture ? local thermal gradient ? local
circulation ? precipitation
24
(No Transcript)
25
0-10cm Soil moisture diff (max-min)
Wet 0.5
Diff 0.3-0.4
Dry 0.1
26
With WRF_NMM model
Local early morning
Local afternoon
T2m diff (wet dry, NMM)
T2m diff (wet dry, NMM)
27
With WRF_ARW model
Local afternoon
Local early morning
T2m diff (wet dry, ARW)
T2m diff (wet dry, ARW)
28
Afternoon (dry)
Afternoon (wet)
Late night (wet)
Late night (dry)
(Aligo, Gallus and Segal, 2006 WAF)
29
Wet soil moisture ? more rainfall In general
especially in mid-high Latitudes than dry soil
situation
Wet soil moisture
Diff (wet dry) with ARW
Dry soil moisture
30
Wet soil moisture ? more rainfall In general
especially in mid-high Latitudes than dry soil
situation
Wet soil moisture
Diff (wet dry) with NMM
Dry soil moisture
31
Local circulation/precipitation modification
32
Response of 3h-apcp forecast to IC difference
depends on soil moisture content More sensitive
in wet condition than in dry condition (WRF_NMM)
Wet soil moisture condition
Dry soil moisture condition
33
T2m is more sensitive to IC in dry soil than in
wet soil condition
(WRF_NMM model)
Wet soil moisture condition
Dry soil moisture condition
34
Response of 3h-apcp forecast to IC difference
depends on soil moisture content More sensitive
in wet condition than in dry condition (WRF_ARW)
Wet soil moisture condition
Dry soil moisture condition
35
T2m is similar maybe slightly more sensitive to
IC in dry soil than in wet soil
(WRF_ARW model)
Wet soil moisture condition
Dry soil moisture condition
36
Response of 3h-apcp forecast to model difference
depends on soil moisture content More sensitive
in wet condition than in dry condition
Wet soil moisture condition
Dry soil moisture condition
37
Similar sensitivity of T2m to model difference is
observed under wet vs. dry tests
Wet soil moisture condition
Dry soil moisture condition
38
Sensitivity of T2m to model diff depends on which
soil moisture initial states to use
With GFS soil moisture (dryer)
With NAM soil moisture (wetter)
39
Cold season case
40
(Difference in 3h-apcp forecasts with WRF_NMM
model)
Due to soil moisture difference
Due to atmos IC difference (benchmark)
41
(Difference of 3h-apcp forecasts with WRF_ARW
model)
Due to soil moisture difference
Due to atmos IC difference (benchmark)
42
Difference of T2m forecasts with WRF_NMM model
Due to soil moisture difference
Due to atmos IC difference (benchmark)
43
Difference of T2m forecasts with WRF_ARW model
Due to soil moisture difference
Due to atmos IC difference (benchmark)
44
Soil moisture perturbation design
45
Response of T2m sensitivity to soil moisture
perturbations spatial structure (NMM)
Randomly perturbed (0.05)
Uniformly perturbed (0.05)
46
Response of T2m sensitivity to soil moisture
perturbations spatial structure (ARW)
Uniformly perturbed (0.05)
Randomly perturbed (0.05)
47
Response of 3h-apcp sensitivity to soil moisture
perturbations spatial structure (NMM)
Uniformly perturbed (0.05)
Randomly perturbed (0.05)
48
Response of 3h-apcp sensitivity to soil moisture
perturbations spatial structure (ARW)
Uniformly perturbed (0.05)
Randomly perturbed (0.05)
49
Response of T2m sensitivity to soil moisture
perturbation size (NMM)
Uniformly perturbed with 0.05 pert size
Uniformly perturbed with 0.03 pert size
50
Response of T2m sensitivity to soil moisture
perturbation size (ARW)
Uniformly perturbed with size of 0.05
Uniformly perturbed with size of 0.03
51
Response of 3h-apcp forecast to soil moisture
perturbation size (NMM)
Uniformly perturbed with 0.05 pert size
Uniformly perturbed with 0.03 pert size
52
Response of 3h-apcp sensitivity to soil moisture
perturbation size (ARW)
Uniformly perturbed with size of 0.05
Uniformly perturbed with size of 0.03
53
f24h
f12h
Difference of T2m forecasts due to soil moisture
difference with WRF_NMM
f36h
f48h
54
f12h
f24h
Difference of T2m forecasts due to atmos IC
difference with WRF_NMM model
f48h
f36h
55
f12h
f24h
Difference of T2m forecasts due to soil moisture
difference with WRF_ARW
f36h
f48h
56
f12h
f24h
Difference of T2m forecasts due to atmos IC
difference with WRF_ARW model
f48h
f36h
57
f48
f60
Sensitivity of 3h-apcp to soil moisture with NMM
model
f84
f72
58
f48
f60
Sensitivity of 3h-apcp to atmos IC with NMM model
f72
59
f48
f60
Sensitivity of 3h-apcp to soil moisture with ARW
model
f72
f84
60
f48
f60
Sensitivity of 3h-apcp to atmos IC with ARW model
f72
f84
61
Next Step to test a Breeding-like cycled
approach to have land surface initial states
perturbation be consistent with atmospheric
variables perturbation.Hopefully an Improved
ensemble system after adding soil moisture
perturbations
62
Summary

• Impact on precipitation and T2m forecast caused
  by soil moisture uncertainty cannot be neglected.
  It could be comparable to that caused by
  atmospheric IC uncertainty. Therefore, soil
  moisture perturbation needs to be considered in
  mesoscale ensembling. The degree of such
  sensitivity varies with model used. Precipitation
  is less sensitive to soil moisture in cold season
  than that in warm season while T2m remains highly
  sensitive to soil moisture in both warm and cold
  seasons.
• In general, wet soil moisture produces more
  rainfall than dry soil moisture especially in
  high/mid-latitudes. Others also showed that dry
  soil leads early afternoon thunderstorms while
  wet soil late night thunderstorms. Precipitation
  becomes more sensitive to IC and model physics
  when soil moisture is wet than dry. (more
  evapotranspiration ? more moist air?)

63
Summary (cont)

• Soil moisture directly impacts T2m by changing
  surface thermal energy balance. Perturbing soil
  moisture has larger impact on T2m than on precip.
  Sensitivity of T2m to soil moisture seems to
  favor certain geographical locations in a
  particular case. Its also noticed that T2m is
  more sensitive to IC and model physics in dry
  soil than in wet soil. (more sensible heat?)
• Sensitivity of T2m and precipitation depends on
  soil moisture perturbations spatial structure
  and magnitude. It is not an easy task to design a
  realistic, simple and effective soil moisture
  perturbation scheme which also needs to be
  consistent with atmospheric models perturbation.
  A simple scheme is planned to be implemented in
  the NCEP regional ensemble system (SREF) in 2007.

64
SPCs historical, first-ever Day-2 High Risk
outlook for April 7, 2006s over 60 tornadoes
event (SREF played a key role)
SPCs Significant Tornado Probability forecast
Derived from SREF (David Bright)
65
Subject SREF and the sig tornado event last
FridayFrom David Bright David.Bright_at_noaa.govDa
te Mon, 10 Apr 2006 094412 -0500To Jun Du
Jun.Du_at_noaa.govHi Jun,I wanted to show you the
image SREF attached. This is an experimental
parameter one of the forecasters is testing and
it's definitely not ready for prime time (i.e.,
I'm not convinced its formulation is entirely
defensible at this point, and its false alarm
ratio has yet to be determined!). But, this F039
hour SREF forecast from last Thursday (09 UTC
SREF 6 April 2006) shows the primary significant
tornado threat centered over nrn AL, nrn MS, and
parts of TN (with SREF mean 850 mb height
overlaid).The SREF played a large part in the
first ever Day 2 "HIGH RISK" outlook ever issued
by the SPC (http//www.spc.noaa.gov/products/outlo
ok/archive/2006/day2otlk_20060406_1730.html). Of
course, other deterministic models, climatology,
and forecaster experience were involved in the
decision process, but the SREF played a key role.
Again, this product only in the evaluation
stage, but I thought you'd be interested
nonetheless.DavidLouis Uccellini
wroteJoe/Russ Between this SREF result, the
first time Day 2 High Risk and the 1.0 POD for
over 60 tornadoes, I think it is time to
celebrate in a big way. I recommend that you guys
need to put together a one pager and work with
Dennis and Carmeyia to highlight this historic
accomplishment(s). Louis