Overview of Dust Storm Weather and Cases Simulations in Beijing

1
Overview of Dust Storm Weather and Cases
Simulations in Beijing

• ZHANG Xiaoling
• xlzhang_at_ium.cn
• Institute of Urban Meteorology, CMA,Beijing

2009-10-29
2
OUTLINE

• Overview of dust storm weather in China
• Dust weather in Beijing and its effect on air
  quality
• The development of a new regional dust numerical
  model in Beijing
• Cases simulations and validation with surface
  weather records and observation
• Summary

3
1?Overview of Dust Storm in North of China

• There have been 68 dust storm events since
  2000,average 13.6 events per year. In 2004,there
  were the 15 events.
• Here is the dust events from 2000 to 2004
  over the North China.

4
The Fig. shows the dust storm events to have a
tendency to small from 1995 to 2002.
5
In 2002, there were most dust storm events over
the past five years
In 2002, there are 17 dust events in the whole
year. In spring, 12 dust events,
6
Severe and extensive dust storms occurred during
18-22 March 2002

• In this episod,
• 72 stations recorded dust storm 24 stations
  recorded sever dust storm.
• Affecting more than 1.4 million km2.

7
Satellite image(17BST,20 Mar.) and surface dust
observation from 20BST 19 to 14BST 20 Mar.
Satellite image(March 21 733 ,2002)
8
Dust storm transported across China to
Korea?Japan.
Beijing
Clear day Morning

• Beijing

Korea 21/03
9
Climate background in the early 2002

• Total precipitation(mm) and its anomaly
  percentage()
• from Jan. to Mar. 2002

10
Climate background in the early 2002

• Temperature anomaly from Jan. to Mar. 2002

11
Dust weather in Beijing
Climate change of dust day number from 1950 to
2002
Monthly distribution of dust weather from 1954 to
2001 in Beijing
12
Variation of dust weather in Beijing

13
Dust weather in Beijing in 2002

• Seasonal distribution of dust weather

14
Dust weather effect on air qualitySuccessive
dust weather occurred from the mid of March to
mid of April.
07-08/04/2002
20/03/2002
15/03/2002
15/04/2002
Daily averaged PM10(mg/m3) Concentration in two
stations of Beijing (Located in northwest and
southeast part of Beijing)
15
The development of a new regional dust numerical
model in Beijing

• Since 2000,we have start to pay more attention to
  dust weather study,including synoptic data
  analysis,monitoring and forecast and earlier
  warning.
• The numerical model is playing great role in
  weather forecast and as well as dust forecast.
• Three key factors are meteorology, dust and
  sand source , description of mechanisms of dust
  formation and transport and deposition.

16
Mesoscale meteorological model

• MM5V3 Developed by PSU/NCAR, is a nonhydrostatic
  dynamic model with terrain-following sigma
  vertical coordinate.
• Used for a broad spectrum of theoretical studies,
  real-time data simulation and forecasting, and
  four-dimensional data assimilation.
• Being our operational model for many years and
  provide good forecast for local wind and weather
  situation which are important for dust weather.

17
Dust Model description Basic Equations of dust
outbreak and transport

• QDKdust bin concentration,Ddust ,Kbin
• Pps-pt, ps and pt represent surface pressure
  and model top pressure.
• In terms of the basic equation of dust outbreak
  and transport
• ? dust tendency,
• ?horizontal diffusion,
• ?vertical diffusion,
• ?mass term, ?dust deposition, ?dust
  emission,
• ?divergence term.

18
Dust Model description Dust model coupled with
MM5

• The dust processes are calculated at each model
  time step

• The model includes all the physical processes of
  dust storm occurrence, lifting, transport, dry
  and wet deposition

dust numerical model construct chart
19
Dust model description Dust size distribution

• The airborne particles are divided into 16 size
  bins from 0.25µm to 60.75µm.

20
Dust Model description
Dust flux equation
F dust flux (unit g.cm-2.sec-1) K direct
proportion coefficient ß non-linearity
coefficient of friction velocity S dust erodible
coefficient ,which derived from land use , soil
type, vegetation aland humidity
coefficient ufriction velocity(unitcm.s-1)
21
Cases simulations and validation with surface
weather records

• Four dust events were selected

22
Case simulation

• Simulation domain and dust erodible coefficient
  distribution

• Distribution of dust erodible coefficient
  (0.0-1.0) in the model region

23
Model Design

• Horizontal resolution 30km
• Vertical sigma level 23
• Time step 90 second,
• Centered 400N/1100E.
• Horizontal grid point numbers 121131
• The sponge lateral boundary condition .
• 48 hours simulation, output data every 3h.
• The initial fields and boundary conditions were
  provided by T106-GCM models objective analysis
  data

24
Simulation analysis surface dust concentration
compared with surface weather records
12h results
25
24h results
26
36h results
27
48h results
28
Model results and comparison

• Comparison of the simulated and measured
  concentrations of PM10 and TSP in Beijing During
  19-21 Mar. dust event

PM10 was observed at Yanqing weather station TSP
observed data provided by Dr.Zhang Renjian from
IAP,CAS
29
Simulated hourly dust concentration from 1908UTC
to 2108UTC March,2002
30

• The simulated dust vertical profiles in Zhurihe
  and Beijing at 11h and 15h simulation

Solid line is for Zhurihe Dashed line is
for Beijing
31
Ddust vertical profiles in Zhurihe and Beijing at
21h and 25h simulation
Solid line is for Zhurihe Dashed line is
for Beijing
32
Ddust vertical profiles in Zhurihe and Beijing at
33h and 39h simulation
Solid line is for Zhurihe Dashed line is
for Beijing
33
Vertical cross section of modeled dust mass
concentration(mg/m3) along near 40N
Sigma
Sigma
12h
24h
34
36h
Vertical cross section of modeled dust mass
concentration(mg/m3) along near 40N
Sigma
Sigma
36h
42h
35
Surface friction velocity every 12 hours from
12UTC 19 to 00UTC 21 Mar.
36
Summary

• Dust weather events occurred frequently in recent
  years,especially in 2002.
• A regional dust coupled model (RDCM5) has been
  established by Beijing Institute of Urban
  Meteorology through developing dust model and
  embedded in MM5.
• The model includes all the atmospheric physical
  processes of sand-dust production, lifting,
  transport, dry and wet deposition.
• First, the model simulates or forecasts well for
  the meteorological elements.

37
Summary (continued)

• Second, from the simulation results of this
  selected 4 dust events and other cases previously
  done, it showed that the regional dust model can
  make a good simulation for the spatial pattern
  and temporal evolution of dust storm.
• Third, the dust model could be applied to the
  real-time forecast.
• Further modification and improvements needed.

38

• Thank you