Snowfall Retrieval from Space

Guosheng Liu

Florida State University

• Mid-latitudes vs. Polar regions

• Radars vs. Radiometers

The difference between mid-latitudes and polar regions- surface & airborne radar obs

Barrow– Alaska (71°N) 2002 -2005 winters

MMCR

Wakasa Bay – Japan (~35°N) 2001&2003 January

APR-2 & Obama Doppler

Ze=56S1.2 (Matrosov 2006)

0 dBZ

20 dBZ

↓

Snowfall examples from CloudSat

snow rainsnow

02/08/2007

Ze-R relationships for Solid Precipitation

Precipitation (mm/h)

0.1 1

Ze (m

m6 m

-3)

10-1

100

101

102

103

104

Sector-13.4GHz (DDA)Sector-35.6GHz (DDA)Sector-94.0GHz (DDA)Dendrite-13.4GHz (DDA)Dendrite-35.6GHz (DDA)Dendrite-94.0GHz (DDA)Wakasa snow (Aonashi,2003)Snow_dry (Puhakka,1975)Snow (Sekhon & Srivastava,1970)Snow_dry (Imai,1960)Plate, column (Ohtake & Henmi,1970)Single crystals(Carlson&Marshall,1972)Snow_dry(Fujiyoshi et al.,1990)-9.4GHzGraupel_13.4GHz, den=0.25 (Mie)Graupel_35.6GHz, den=0.25 (Mie)Graupel_94.0GHz, den=0.25 (Mie)

5

13.4 GHz

35.6 GHz

94 GHz

Z-R relations for snowfallDepends on:

Size Distribution + particle shape

(Matrosov 2007)

High frequency measurement of snowfallδ T

B (K

)

-80-60-40-20

02040

89150220340

δ TB (K

)

-80

-60

-40

-20

0

20

183+-1183+-3183+-7

AMSUB (A1-B1)T B (K

)

200

210

220

T B (K

)210

220

230

Latitude

36.5 37.0 37.5 38.0 38.5

T B (K

)

230

240

250

183+-7

183+-3183+-1

150

89

Snowfall Retrieval From High –Frequency Microwave Satellite Observations

• Use surface and airborne radar snowfall observations as the basis to build the a-priori database

• Radiative transfer model utilizes nonspherical snowflakes scattering calculations

A Case Study (Jan 14 2001)

AMSU-B Retrieval

GMS IR Radar (AMeDAS)

3 Cases (1ox1o)

Retrieved Snowfall Rate (mm h-1)

0 1 2 3 4 5 6

AMeD

AS S

now

fall

Rat

e (m

m h

-1)

0

1

2

3

4

5

6

+ 0 hr+ 1 hr+ 2 hr+ 3 hr

Compare AMSU-B Retrieval with AMeDAS radar (3 cases, correlation coefficient: 0.79) Noh, Liu, et al. (2006)

Winter Snowfall Distribution

Winter Rainfall Distribution

SSM/T-2

snowfall

COADS snow

fraction

SSM/I

rainfall

COADS rain

fraction

Below 60Below 60°°NN

Liu & Curry (1996)

SSM/I 85GHz ΔTB 92/12/25-93/01/03, N. Pole – 70 N

-100

-50

0

50

100

-20

-10

0

10

20

0

2

4

6

8

10

0

100

200

300

0

2

4

6

8

-20 -10 0 10 20 30-110

-70

-30

10

50

-20 -10 0 10 20 30

85 GHz Brightness Temperature Anomaly (K)

(a) (b)

(c) (d)

(e) (f) CC: 0.32

CC: 0.71CC: 0.45

CC: 0.79CC: 0.76

CC: 0.59

LW DOWN SURF TEMP

CLOUD AMT

LWP

RADAR, 300m

DAILY SNOW

Liu & Curry (2003)

Detecting Snowfall From Space- what we knew and what we do not know

• From space radar (CloudSat, future new radars)– Pros: physically direct, only need Z to R conversion– Cons: CloudSat - spatial coverage, only a 1.4 km wide strip per orbit– Future radar ? (GPM radars can go up to ~67 degree latitudes), how about polar

regions ?– Need minimum detectable dBZ value -10 dBZ or smaller

• From High-Frequency Microwave Radiometers (AMSU-B, SSMIS, GMI)– Pros: More spatial and temporal coverage (4~6 satellites)– Cons: 1. weaker signature, 2. physically less direct, 3. surface contamination– Below ~60N degree, particularly over ocean, there have been attempts in retrieving

snow from HF-Microwave radiometer obs, – Above ~60 degree latitudes, ???

• Polar Regions:– Satellite radars see snow;– Radiometers? Need to explore

• Studies shows increase of TB associated with cloud/snow at 85, 37 GHz• How about even higher frequencies?