SIMULATED SNOWFALL MEASUREMENTS
FROM SATELLITE OBSERVATIONS
AT MICROWAVE FREQUENCIES
ALBERTO MUGNAI(CNR/ISAC, ROMA, ITALY)
2nd INTERNATIONAL PRECIPITATIONWORKING GROUP WORKSHOP
MONTEREY, CALIFORNIA, USA – 25-28 OCTOBER 2004
2
References
Mugnai, A., S. Di Michele, E.A. Smith, F. Baordo, P. Bauer, B. Bizzarri , P. Joe, C. Kidd, F.S. Marzano, A. Tassa, J. Testud, and G.J. Tripoli, 2004: Snowfall measurements by the proposed European GPM mission. Measuring Precipitation from Space: EURAINSAT and the Future (V. Levizzani, P. Bauer, and F.J. Turk, Eds.), Kluwer Academic Publ., in press.
Mugnai, A., et al., 2004: EGPM-Ice Final Report. Final Report of ESA/ESTEC Contract on “Modelling and Retrieval of Light Rainfall and Solid (Ice and Snow) Precipitation for the EGPM Mission”, CNR-ISAC, in preparation.
ESA, 2004: EGPM – European Contribution to Global Precipitation Measurement. The Six Candidate Earth Explorer Missions: Report for Mission Selection SP-1279 (5), ESA/ESTEC, 66 pp.
3EGPM – EUROPEAN CONTRIBUTION TOGLOBAL PRECIPITATION MEASUREMENT (GPM)
Improve accuracy of global precipitation estimates with focus on snowfall and light rain – unique feature within GPM;
Improve global and regional NWP and climate model forecasts;
Improve near-real-time monitoring of hazardous and flash-flood producing storms.
EGPM mission goal is to retrieve precipitation with emphasis on Europe and Canada.
EGPM mission is integral – but distinct component of GPM constellation
ESA’s Earth Observation Programme Board (PB-EO) has recently recommended that “EGPM mission should be furthered within ESA Earth Watch framework”.
4
SIGNIFICANCE OF LIGHT RAINFALL AND SNOWFALL
At high latitudes there is large fraction of light and solid precipitation
Latitude / precipitation rate category diagramderived from COADS Dataset
5
satellite orbit
swath width
flight direction
Radiometer FOV
fore-view
aft-view
45° flight direction
fore-view
aft-view view from the top
Radar FOV
VIEWING GEOMETRY OF EGPM RADIOMETER & RADAR
Radiometer Antenna Size @ 500 km: 1.0 mRadar Antenna Size @ 500 km: 1.2 m
Radar:
Radiometer Calibration
Intercalibration
Extension to high latitudes
6
EGPM MICROWAVE RADIOMETER CHANNELS
36.5
Footprint[km x km]
Polarization
26 x 26 H + V
26 x 26 H or V
13.4 x 13.4 H + V
5.5 x 5.5 H + V
Frequency[GHz]
18.7
23.8
89.0(*)
150.0(*)
5.5 x 5.5 H + V
Window
Footprint[km x km]
Polarization
13.4 x 13.4
H or V(***)
13.4 x 13.4
H or V(***)
13.4 x 13.4
H or V(***)
13.4 x 13.4
H or V(***)
13.4 x 13.4
H or V(***)
13.4 x 13.4
H or V(***)
13.4 x 13.4
H or V(***)
Frequency[GHz]
50.3(**)51.8(**)52.8(**)53.8(**)
118.75 ± 8.5(~**)
118.75 ± 4.2(~**)
118.75 ± 2.3(~**)
118.75 ± 1.2(~**)
13.4 x 13.4
H or V(***)
Temperature-Sounding
(*) channel should provide contiguous and concentric coverage along and across
(**) NAST-M channels (NPOESS Aircraft Sounder Testbed – Microwave)
(***) channels should provide either all H or all V polarization
Bauer, P., and A. Mugnai, 2003: Precipitation profile retrievals using temperature-sounding microwave observations. J. Geophys. Res., 108 (D23), 4730, doi:10.1029/2003JD003572.
Blackwell, W.J., J.W. Barrett, F.W. Chen, R.V. Leslie, P.W. Rosenkranz, M.J. Schwartz, and D.H. Staelin, 2001: NPOESS Aircraft Sounder Testbed-Microwave (NAST-M): Instrument description and initial flight results. IEEE Trans. Geosci. Remote Sensing, 39, 2444-2453.
7
EGPM NADIR-POINTING PRECIPITATION RADAR (NPR)
NPR instantaneous footprints
Parameter Mission Requirement
Sensitivity to precipitation 0.1 mm/hour
Fundamental integration length < 5 km
Footprint diameter < 5 km
Vertical resolution < 250 m
Vertical range - 3 to 15 km
Requirements
8
CLOUD RESOLVING MODEL (CRM) SIMULATIONS
University of Wisconsin – Non-hydrostatic Modeling System (UW-NMS) / Tripoli (1992 , 2004)
Three-dimensional, time-dependent, explicit, non-hydrostatic CRM; Bulk microphysics scheme for six different hydrometeors:
Cloud droplets Rain drops Graupel particles Pristine ice crystals Snowflakes Ice aggregates
All particles are assumed to be spherical; Size distributions:
Cloud droplets & pristine ice crystals : monodispersed; Other hydrometeors : inverse exponential constant-slope / constant-intercept n(D) = A exp (-BD)
Density: Ice crystals, graupel & snowflakes : constant Ice aggregates : size-dependent
UW-NMS Simulations1) January 2000 snowstorm that produced significant snowfall over east coast of
United States.2) January 2003 frontal system that produced light snow over Baltic area (Gotland).3) October 2002 frontal system that produced light-to-medium rainfall over southern
United Kingdom and English Channel.
9US EAST COST SNOW STORM (JANUARY 24-26, 2000):UW-NMS NESTED GRID CONFIGURATION
Inner Grid:468 x 468 km2
2.34 km res.
10UW-NMS US EAST COST SNOW STORM SIMULATION:COLUMNAR LWC/IWC’S (06.00, JAN. 25, 2000)
11
US EAST COST SNOW STORM : RAIN / SNOW RATE
12MODEL SIMULATION OF EGPM RADIOMETER OBSERVATIONS:LOW WINDOW FREQUENCIES
US East Coast Snow StormB
rig
thn
ess
Tem
per
atu
re (
TB
)
Over ocean: all sensitive to rainfall Over land: 36.5 GHz responds to scattering by large snow amounts
Co
lum
nar
Wat
er C
on
ten
t (C
WC
)
13MODEL SIMULATION OF EGPM RADIOMETER OBSERVATIONS:HIGH WINDOW FREQUENCIES
US East Coast Snow Storm
Bri
gth
nes
s T
emp
erat
ure
Over ocean: both highly sensitive to scattering by ice particles aloftOver land: both respond very strongly to scattering from snow
Co
lum
nar
Wat
er C
on
ten
t
14
TB’s
Four channels sense clouds and precipitation to different depths.
Co
lum
n W
ater
Co
nte
nt
MODEL SIMULATION OF EGPM RADIOMETER OBSERVATIONS:O2 SOUNDING CHANNELS
US East Coast Snow Storm
Four corresponding channel pairs provide information on ice amount from top to different depths inside cloud and reduce sensitivity to surface.
15CROSS SECTION A : LIQUID/ICE WATER CONTENTS
L / O
16
CROSS SECTION A : TB’S
17
CROSS SECTION C: TB’SC
olu
mn
ar W
ater
Co
nte
nt
Dis
tan
ce
US East Coast Snow Storm
18
Co
lum
n W
ater
Co
nte
nt
MODEL SIMULATION OF EGPM RADAR OBSERVATIONS
US East Coast Snow Storm
0 dBZ5 dBZ
16 dBZ
Five (5) dBZ sensitivity of EGPM radar will allow detection of light precipitation at surface and of full precipitation profile.
19
SNOW RETRIEVAL
Average retrieval accuracy of the near-ground (up to 500 m) snow equivalent LWC
(“true” average model value = 0.7841 gm-3)
20SNOWFALL MEASUREMENTSUMMARY
Snowfall over land or snow cover has distinct signatures at high-frequency window channels (89 and 150 GHz), as well as at selected temperature-sounding channel pairs of two oxygen bands near 50-54 and 118 GHz. In particular, TB-depressions and relative behavior of 89-150 GHz and 54-118 GHz frequencies carry information on whether snow is precipitating out or is still aloft, on vertical depth of snow layer, on snowfall rate, as well as on possible concomitant presence of rain.
Thus, snowfall over land can be detected and measured by space-borne radiometer employing such frequency combinations – emphasizing value of sounding channels because they are significantly less sensitive to surface emissivity properties than window channels.
Snowfall over ocean can be observed through its scattering/emission signatures at all frequencies.
Often, retrieval problem is more complex because oceanic snowfall melts into liquid precipitation above sea surfaces with SST > 0 °C.
EGPM radar will allow detection of light precipitation and precipitating ice aggregates and will help interpretation/retrieval of radiometric observations.
21
GOTLAND : SNOW RETRIEVAL (OVER LAND)
22
UK FRONTAL SYSTEM : SNOW + RAIN RETRIEVAL
SEA / LAND
23LIGHT-TO-MODERATE RAINFALL MEASUREMENTCONCLUSIONS
Light-to-moderate rainfall over ocean can be directly sensed by lower window channels (18.7 and 36.5 GHz), as well as by water vapor channel at 23.8 GHz.
Light-to-moderate rain over land is indirectly observed through scattering of radiation by ice particles aloft at high-frequency window channels (89 and 150 GHz) and at temperature-sounding channels of oxygen line near 118 GHz (to lesser extent, at 36 GHz and at temperature-sounding channels of oxygen band near 50-54 GHz).
Generally, variability of surface emissivity obfuscates rain emission and thus retrieval problem over land is more complex than over ocean. However, additional information provided by sounding channels reduces retrieval uncertainty (reduces both bias and rms errors).
EGPM radar will allow detection of light-to-moderate precipitation and will help interpretation/retrieval of radiometric observations.
24
Backup Slides
25
DISTRIBUTION OF RAINFALL
Cumulative distribution of rainfall accumulations (solid) and occurrence (dashed)against rainfall intensity derived from European radar (Europe) and from TRMM
(Tropics).
Rainfall Accumulation & Occurrence
26
DISTRIBUTION OF SNOWFALL
Snow to Total Precipitation Ratio
Snowfall Accumulation
27
CONCEPTUAL SCHEME OF THE EGPM RADIOMETER SIMULATOR
28US EAST COST SNOW STORM:SURFACE TEMPERATURE & COLUMNAR WATER VAPOR
29US EAST COST SNOW STORM: UW-NMS SIMULATEDCOLUMNAR LWC/IWC’S (06.00, JAN. 25)
30
MICROWAVE SNOW EMISSIVITY
Weng, F., and B. Yan: A Microwave Snow Emissivity ModelEmpirically derived from satellite retrievals and ground-based measurements
Large variation with frequency depending on snow type;Large variation with snow type at high frequencies.
31CROSS SECTION A :ΔJ’S (RAIN)
32CROSS SECTION A :ΔJ’S (SNOW)
33CROSS SECTION C : SNOW EMISSIVITY & WINDOW FREQUENCIES
L / O
34CROSS SECT. C : SNOW EMISSIVITY & SOUNDING FREQUENCIES
35GOTLAND LIGHT SNOW UW-NMS SIMULATIONCOLUMNAR LWC/IWC’S
A
B
36GOTLAND LIGHT SNOW UW-NMS SIMULATION SURFACE TEMPERATURE & COLUMNAR WATER VAPOUR
37
GOTLAND LIGHT SNOW : LOW WINDOW FREQUENCIES
38
GOTLAND LIGHT SNOW : HIGH WINDOW FREQUENCIES
39
GOTLAND LIGHT SNOW : O2 SOUNDING CHANNELS
40
GOTLAND CROSS SECTIONS : LIQUID/ICE WATER CONTENTS
A B
41
GOTLAND CROSS SECTIONS : TB’S (OVER LAND)A B
42UW-NMS UK FRONTAL SYSTEM SIMULATIONCOLUMNAR LWC/IWC’S
43
UK FRONTAL SYSTEM : SURFACE RAINFALL RATE
44UK FRONTAL SYSTEM SIMULATIONSURFACE TEMPERATURE & COLUMNAR WATER VAPOUR
45MODEL SIMULATION OF EGPM RADIOMETER OBSERVATIONS:LOW WINDOW FREQUENCIES
Co
lum
na
r W
ate
r C
on
ten
t (C
WC
)B
rig
thn
es
s T
em
pe
ratu
re (
TB
)
Over ocean: all frequencies are sensitive to rainfall Over land: 36.5 GHz responds to scattering by large ice amounts
UK Frontal System
46MODEL SIMULATION OF EGPM RADIOMETER OBSERVATIONS:HIGH WINDOW FREQUENCIES
Both over ocean and over land:Both frequencies respond very strongly to scattering from ice particles aloft
Co
lum
na
r W
ate
r C
on
ten
t (C
WC
)
Bri
gth
ne
ss
Te
mp
era
ture
(T
B)
UK Frontal System
47
Co
lum
na
r W
ate
r C
on
ten
t (C
WC
)
MODEL SIMULATION OF EGPM RADIOMETER OBSERVATIONS:O2 SOUNDING CHANNELS
The four channels sense clouds and precipitation to different depths.
TB’sThe four corresponding channel pairs provide information on ice amount from
top to different depths inside the cloud and reduce sensitivity to surface.
UK Frontal System
48
CROSS SECTION : LIQUID/ICE WATER CONTENTS
49UK FRONTAL SYSTEMCOLUMNAR RAIN LWC VS. COLUMNAR SNOW IWC
50
UK TB’SSEA / LAND