Page 1
Preliminary tests on the Cycle of Preliminary tests on the Cycle of
Hourly Assimilation and Forecast Hourly Assimilation and Forecast
system in Guangzhousystem in Guangzhou
WAN Qi-lin, CHEN Zi-tong, DING Wei-yu, HUANG Yan-yanWAN Qi-lin, CHEN Zi-tong, DING Wei-yu, HUANG Yan-yan
Institute of the tropical and marine meteorology, CMA, GuangzhouInstitute of the tropical and marine meteorology, CMA, Guangzhou
Page 2
Outline
Motivation
system introduce
Some tests
conclusions
Page 3
Motivation
How to make use of high frequent observation data for improving numerical analysis and forecast .
How to give a good support to nowcasting
Page 4
system introduce
For short, the system is named as CFAF
CHAF - The Cycle of Hourly Assimilation and Forecast
based on Grapes - Global/ Regional Assimilation PrEdiction
System
Page 5
Domain
Grid resolution:0.12°
CHAF domain
Page 6
System flow chart
Pretreatment
ObservationDATA
Cycle
Boundary(from coarse model)
OperationPrediction
Increment interpolation
Increment Digital Filter
GRAPES_ MODELGRAPES-Model
background
GRAPES_ 3DVARGRAPES-Var
SimpleCloud analysis
nudging
Page 7
The Case choosing to understand the basic character of CHAF
The case is about a typhoon , the typhoon is landing and tend to weaken.
At 12UTC, 13 August,2005, a typhoon have been landing at Guangdong Province.
In the simulation test, we chose 12UTC as initial time ,and have performed 24h forecast.
Page 8
The scheme of experiment--to check if the cycle would give birth to noise, and if the cycle would benefit to improve analysis filed.
Test 1 : The model run 24 hour, without cycle;
Test 2 : The model run 24 hour, with the cycle of hourly assimilation and forcast.
The DFI window is six hour ( -3 , +3 ), and, the observation was not import 。
Test 3 : The model run 24 hour, with the cycle of hourly assimilation and forcast.
The DFI window is six hour ( -3 , +3 ), and, the Sounding and surface observation was ingest 。
Page 9
Test 1 –Control test
Page 10
Test 2
test3 test1test2
Test2
Test1
Page 11
Test 3
Surface observation Surface observation
Sounding
Page 12
Initial field_SLP12z 13AUGCycle- test 3
observationControl Test
Page 13
SLP12H Forecast00z 14AUG
Test 1
Test 3observation
Page 14
500hPa Height12H forecast
Test 1
Test 3observation
Page 15
12H FORECAST Rainfall :6H Accum00z14AUG
Test 1
Test 3observation
Page 16
conclusion
According to the result of the tests , we think that: the cycle of CHAF system did not arouse addi
tional error evidently, the CHAF system must be capable of assimil
ation some high frequent observation.
Page 17
The tests to assimilate high frequent observation.
Cloud-drift wind Airline observation ( AMDAR ) Calculated Cloud (Cloud analysis)
Page 18
Hourly asimilation of cloud-drift wind
The Vector of cloud-drift wind on 200hpa( 2005-07-11 00 : 00~12 : 00UTC )
Page 19
The scheme of comparative experiment
Test 1
Initial value00 UTC
Analysis field
at 12UTC
forecast
Initial value 00 UTC
forecast
Hourlr assimilate cloud-drift wind
Analysis field
at 12UTC
Test 2
Page 20
Height and Wind on 500hpa (analysis)
a. With cloud-drift wind
b. Without CDW
c . Velocity difference between a and b
d . NCEP analysis
a b
c d
Page 21
24H precipitation ( 2005-7-11 12 : 00UTC~2005-7-12 12 : 00UTC )
a . With cloud-drift wind b . without c . Difference d . Real
a b
c d
Page 22
The observation distribution from Air report between 00 - 23UTC on 18 Jun. 2 ( left: time-height cross-section ; right: planform – look down )
Test about Air report assimilation
The three-dimensional distribution of Airline observation between 04 - 06UTC on 18 Jun.
Page 23
The comparative test of Air report assimilation
12Z18Jun00Z18Jun
Test2:1-h assimilation of Air report,and sounding,surf,ship
12Z18Jun
24h Forecast
24h Forecast
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23探空 46地面 346 1053 409 867 239 334 384船舶 4 14 4 7 4 15 4卫星 160飞机 128 127 75 131 155 149 162 112 54 118 105 139 200 155 38 25 15 5 5 15 6 7 35
00Z18Jun
Test1:1-h assimilation of sounding,surf,ship
Page 24
The 24H precipitation
With airline observ-ati
on
Without
Difference Obs.
Page 25
The 3km cloud at 2005082012UTC . The below pictures are thecomparing the calculated rain content and the reflection of radar
Page 26
The 5km cloud at 2005082012UTC . The below pictures are thecomparing the calculated rain content and the reflection of radar
Page 27
The nudging of analysis cloud has effect on 0-1h precipitation forecast
With nudging cloudWithout
Difference Obs.
Page 28
The nudging of calculated cloud has effect on 1-2h precipitation forecast
Difference
With nudging cloudWithout
Obs.
Page 29
The case of Real time Running
Red - CYCLE;Green - operation ; Black - observation
cycle obs
3day 2day 1day
Page 30
Summary
In the CHAF system, the cycle did not arouse additional error evidently.
The analysis and forecast can be improved by use of high frequent observation data.
The CHAF system would have a good future in application to nowcasting.
But More research must be done to improve the cloud
analysis,3DVAR, cycle system flow, etc. Farther, a long time running must be performed t
o inspect the CHAF system.
Page 31
The end,thank you !