UNIVERSITY OF ATHENS

1
UNIVERSITY OF ATHENS FACULTY OF PHYSICS DEP. OF
APPLIED PHYSICS LAB. OF METEOROLOGY
Sensitivity tests in the dynamical and
thermal part of the MRF-urban PBL scheme  in
the MM5 model
Aggeliki Dandou, Maria Tombrou
2
1
Meteorological Model MM5 (Version 3-6)
The Penn State/NCAR Mesoscale Model MM5 (Grell et
al., 1994) is a terrain following numerical
weather prediction model, with a multiple-nest
capability, nonhydrostatic dynamics and a
four-dimensional data assimilation capability.
Parameterization schemes considered

• simple ice Hsie et al. (1984), for the
  moisture parameterization
• cloud-radiation scheme Dudhia (1989), for the
  radiation parameterization
• Grell (1993), for the clouds parameterization
• Five-Layer Soil model (Dudhia, 1996), for the
  soil parameterization

3
2
Meteorological Model MM5 (Version 3-6)
PBL parameterization schemes

• MRF (Hong and Pan, 1996) high resolution
  non-local scheme based on Troen and Mahrt (1986)
  representation of counter gradient term and K
  profile in the well mixed PBL

• MRF-urban (Dandou et al., 2005) a modified
  version of MRF whereby urban features were
  introduced both in the thermal part and the
  dynamical part

• MRF-not urban a modified version of MRF,
  whereby the city of Athens is replaced by dry
  cropland and pasture surface, as the surrounding
  area

• Anthropogenic heat as a temporal and spatial
  function of the diurnal variation of the
  anthropogenic emissions
• Heat storage the OHM scheme (Grimmond et al.,
  1991)

• Heat and momentum fluxes (under unstable
  conditions) according to Akylas et al. (2003)
• Diffusion coefficients (under stable
  conditions) according to King et al. (2001)

• Updated field for the roughness length based on
  literature values in combination with satellite
  detailed information on land use (spatial
  resolution 30 m)

4
3
Area of application
Ground stations

• National Observatory of Athens (NOA), an urban
  station (4 km inland from the shore), located in
  a park, on top of a hill (107 m asl), with urban
  characteristics 85 and z00.8 m

Input data

• Meteorological data ECMWF (0.5o x 0.5o) every 6
  hours

• Sea Surface Temperature SST (1o x1o) every 6 hours

• Marousi, a suburban station (13 km inland from
  the shore), inside a grove surrounded by
  buildings of different heights, with urban
  characteristics 52 and z00.5 m

• USGS data (25 categories) (30 x 30) for
  topography and land use

• Two-way nesting

• Peiraias, an urban station at the harbor, with
  urban characteristics 100 and z01 m

5
4
14 September 1994 (MEDiterranean CAmpaign of
PHOtochemical Evolution MEDCAPHOT-TRACE
experiment, Ziomas, 1998)
Available measurements

• Sensible heat flux (sonic anemometer)

NOA, Marousi

• Friction velocity (sonic anemometer)

• Air temperature

NOA, Marousi, Peiraias

• Wind velocity

• Landsat TM satellite image (acquisition date 13
  Joune 1993)

6
5
Results

• Surface fluxes

• Surface fluxes

• Air Temperature

• Diffusion coefficients

• PBL height

7
6

• Diurnal variation of surface fluxes (MRF-urban
  scheme)

(urban, downtown)
(urban, at the harbor)
(semi-urban)
Q- net all wave radiation QH-sensible heat
flux QE-latent heat flux QF-anthropogenic heat
?Qs-heat storage
8
7

• Diurnal variation of sensible heat flux

Decrease MANGE 29
Decrease MANGE 40
Model results versus measurements
Measurements (Batchvarova and Gryning, 1998)
Schemes intercomparison
soil characteristics (bare rocks) and surface
cover (olive tree plantation)
temperature gradients (higher location)
(Mean Absolute Normalized Gross Error)
-predicted values
-observed values
9
8

• Spatial distribution of the sensible heat flux

W/m2
300 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
1400 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
10
9

• Diurnal variation of friction velocity

Decrease MANGE 6
Decrease MANGE 2
Measurements (Batchvarova and Gryning, 1998)
Model results versus measurements
Schemes intercomparison
increase of z0 increase in the diffusion
processes decrease in the wind speed
increase of u normalization in temperature
gradients decrease of u
decrease of u
11
10

• Spatial distribution of the friction velocity

m/s
300 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
1400 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
12
5
Results

• Surface fluxes

• Surface fluxes

• Air Temperature

• Air Temperature

• Diffusion coefficients

• PBL height

13
11

• Diurnal variation of air temperature (at 10 m
  agl)

Decrease MANGE 38
Decrease MANGE 41
Decrease MANGE 53
Model results versus measurements
Schemes intercomparison
increase due to the dynamical part decrease due
to the thermal part
total decrease
14
12

• Spatial distribution of the air temperature
  differences (at 2m agl)

• Spatial distribution of the air temperature
  (at 2m agl)

?? (oC)
oC
MRF-dyn - MRF
MRF-ther - MRF
MRF-urban - MRF
MRF-urban
300 LST
300 LST
300 LST
300 LST
MRF-dyn - MRF
MRF-ther - MRF
MRF-urban - MRF
MRF-urban
1400 LST
1400 LST
1400 LST
1400 LST
15
5
Results

• Surface fluxes

• Surface fluxes

• Air Temperature

• Air Temperature

• Diffusion coefficients

• Diffusion coefficients

• PBL height

16
13

• Diffusion coefficient profiles

increase due to the dynamical part decrease due
to the thermal part
total decrease
17
14

• Spatial distribution of diffusion coefficients
  at the surface layer

m2/s
300 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
1400 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
18
5
Results

• Surface fluxes

• Surface fluxes

• Air Temperature

• Air Temperature

• Diffusion coefficients

• Diffusion coefficients

• PBL height

• PBL height

19
15

• Diurnal variation of the PBL height

increase due to the dynamical part decrease due
to the thermal part
total decrease
20
16

• Diurnal variation of the PBL height

20-9-2002
20-9-2002
15-9-1994
Tombrou et al. (2006)
21
17

• Spatial distribution of the PBL height

m
300 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
1400 LST
MRF
MRF-dyn
MRF-ther
MRF-urban
22
18

• Summarizing the results

n77 (number of grids) t12 (number of hours for
the day and night)
Day (0700-1800 LST) MRF-dynall - MRF () MRF-ther - MRF () MRF-urban - MRF ()
Sensible heat -1.92 -28.15 -28.99
Friction velocity -3.98 -10.84 -10.79
Air temperature 2.10 -15.21 -9.12
Diffusion coefficients 36.58 -34.58 -20.97
Wind speed 6.87 -4.85 -6.47
PBL height 10.45 -10.47 -5.11
Night (0000-0600 LST, 1900-2300 LST) MRF-dynall - MRF () MRF-ther - MRF () MRF-urban - MRF ()
Sensible heat -3.38 -47.70 -51.79
Friction velocity 1.19 7.69 7.65
Air temperature 1.08 3.80 4.03
Diffusion coefficients 10.21 5.30 83.9
Wind speed 0.53 64.90 60.06
PBL height 5.95 86.40 88.30
23
19

• Spatial distribution of the wind velocity (at 10
  m)

• Spatial distribution of the wind speed
  differences (at 10m)

MRF-urban MRF-not urban
MRF-urban
300 LST
300 LST
m/s
m/s
MRF-urban MRF-not urban
MRF-urban
1400 LST
1400 LST
24
20

• Vertical cross sections of the wind velocity
  along the sea-breeze axis

MRF-urban
MRF-not urban
300 LST
300 LST
m/s
1400 LST
1400 LST
0.4 m/s
4 m/s
25
21
Conclusions

• Both modifications play an important role and
  improve the models results

• The increase in temperature and diffusion
  coefficients calculated by the dynamical part
  is compensated by the decrease in the thermal
  part, resulting in a total decrease

• A decrease in turbulence and fluxes is calculated
  by both modifications

• During the day

• A slowing in the sea-breeze front and a
  frictional retard concerning its penetration over
  the Athens city is calculated due to the
  increased roughness length

• The total increase in temperature, diffusion
  coefficients, turbulence and fluxes is due to
  both modifications

• During the night

• The maximum wind speeds calculated in the lower
  atmosphere is due to the urban heat island