Application of Cloud Analysis in GRAPES_RAFS

earsplittinggoodbeeInternet and Web Development

Nov 3, 2013 (4 years and 7 days ago)

95 views

Application

of Cloud Analysis in
GRAPES_RAFS


Lijuan ZHU
[1]
, Dehui CHEN
[1]
, Zechun LI
[1]
, Liping LIU
[2]
,
Zhifang XU
[1]
, Ruixia LIU
[3]


[1]
National Meteorological Centre (NMC)

[2]
Chinese Academy of Meteorological Sciences (CAMS)

[3]
National Satellite Meteorological Center (NSMC)

China Meteorological Administration (CMA)

Beijing, 100081

(25 October 2011, for workshop
-
NWP nowcasting in Boulder
-
USA)

Outline


1 Motivations


2 Cloud Analysis in GRAPES


3 Data used by C.A. of GRAPES


4 Preliminary results


5 Summary


1 Motivations

1 Motivations


There are a lot of data sets which are yet difficult
to be directly assimilated, but could be fused for
the model initialization for some reasons of
technique approaches or computation
effectiveness.


These data sets are available, such as the
satellite images or retrieved cloud products,
surface visual + instrumental observations of
cloud, visibility, lightning and so on, specially the
radar reflectivity.

CMA’s Radar Network: CINRAD

The observations of ~158 radars, which have been deployed in whole
China (most along with East coast line) , are available to be used.

1 Motivations


In other hand, a “cold
-
start” GRAPES is poor to
provide the initial information of cloud for the
microphysical scheme, and the associated
moisture field and vertical motions.


It is naturally motivated for us to fuse the
available data sets for generating a more
reasonable initial field with a detailed 3D cloud
specification to produce the meso
-
scale cloud
analysis products, and to improve short
-
time
H.I.W. forecasts.

2 Cloud Analysis in
GRAPES

Cloud Analysis in GRAPES_RAFS

1


Cloud analysis scheme from ADAS of ARPS Model developed by
CAPS,OU
(
Xue et al.,

MAP,
2003 ; Hu, Xue et al.,
MWR
, 2006

)
based on LAPS
(
Albers et al., 1996
)

Fusion of all cloud, precipitation observations

Synop

Satellite
IR ,VIS

Radar
Ref

Background
moisture

Cloud
field

Cloud
amount

Cloud
base

Cloud
thick

Cloud
type





Hydrom.

Background


observations

3
D

cloud

field

,

cloud

amout

Cloud

type


Cloud

water,

cloud

ice

Qc

on

cloud

type

(Cumulus)

Precipitation

type


Precipitation

(qr,

qs

,

qh,


)


Be

nudged


dynamical relaxation factor

And then the cloud analyzed information can be included
by nudging method for the model initialization

Cloud Analysis in GRAPES_RAFS

2


Cloud analysis can be called every 1 hour or every 3 hours.

Changes in the original C.A.


(1) Correction in the code about Synop
application to modify the background cloud
base specification
(
b
arnes interpolation weights
):

original

modified

(2) The introduction of saturation on ice
-
surface scheme


Org: only
water surface saturation


Modified by adding ice surface saturation

with ice surface saturation

Org: water surface saturation only

TRMM

(3) Permitting cloud water, cloud ice as well

NCEP’s RUC:
more suitable to s
tratus
-
cumulus

(smaller upward motion in cloud),
which dominate in most cases in China;

Original scheme:
more focused on deep
convective cumulus (stronger upward
motion in cloud)

(4) Quality control of radar reflectivity

Ground Clutter, Clear air echo, etc.

TRMM

Cloud Water

original

modified

Cloud Ice

TRMM

original

modified

3 Data used by C.A. of
GRAPES

Data used



Background:
3D grid fields of RH, Temperature,
Pressure, surface temperature from 3DVAR analysis



SYNOP:
Cloud base ,Cloud amount



Radar 3D Mosaic Reflectivity

Composite reflectivity
over whole China or
domain specified
;

Satellite

FY
-
2 IR TBB

FY
-
2 VIS CTA

SAT

advantage: to specify the cloud top

FY
-
2 Geostationary satellite, FY2D/2E

every 30min

bu琠 jus琠
hourly da瑡 used by RAFS



Data use

捯nt.


4 Preliminary results

Specification of the experiment


Case

a Tropical Storm landed on
Guangdong coast line


Model:
15km GRAPES using T213 for
3DVAR FG and BC


Background analysis:
3DVAR analysis
downscaling to cloud analysis mesh of 5km

as background of C.A.


Initial Time

Aug. 6, 2009 at 00UTC

b. cloud

modified c. base

used IR TBB

used radar reflect.

used visible image

Impact on cloud cover analysis

IR TBB Obs.

Corrected the cloud base

Before

After

Cloud top compared to MODIS

MODIS

Cloud analysis

Cloud Type

Radar Ref

1 St:
Stratus

2 Sc:Stratocumulus

3 Cu:Cumulus

4 Ns:Nimbostratus

5 Ac:
Altocumulus

6 AS:
Altostratus

7 Cs:Cirrostratus

8 Ci:Cirrus

9 Cc:Cirrocumulus

10

Cb

:
Cumulonimbus

Compared to cloudsat

cloudsat

Cloud analysis

Height(km)

Analyzed hydrometeors

Radar reflectivity(Ob)

Cloud water

Cloud ice

Qr

Qs

Impact on forecast

3h forecast

Radar
obs

With cloud analysis

Without cloud analysis

With cloud analysis

6h forecast

12h forecast

Radar
obs

Radar
obs

Without cloud analysis

Without cloud analysis

With cloud analysis

All china

<10mm

<25mm

<50mm

<100mm

Warm start

0.395

0.203

0.068

0.017

Warm start+cloud
analysis

0.398

0.206

0.066

0.033

TS
-
verification of
6H Precipitation
forecasts


(
for July 5~30, 2009)

5 Summary

Conclusion and discussion


The cloud analysis scheme ADAS has been
adapted to GRAPES_RAFS, and with some
modifications.


The preliminary experiments have showed the
positive impacts. It still needs much further
assessments.


The quality control of the radar reflectivity is still a
big challenge for real time application, not only
due to the reflectivity quality itself, but also due to
effectively receive the data in time.





Conclusion and discussion (cont.)


The cloud analysis is a complicated issue. It is
particularly necessary to adapt it according the
s
tratus
-
cumulus which dominate in most cases in
China.


A lot of works are ongoing for real
-
time
implementation of RAFS with C.A. at NMC/CMA.





Thanks!