MODIS, AIRS, and Midlevel Cloud Phase

1
MODIS, AIRS, and Midlevel Cloud Phase

• Shaima Nasiri
• CIMSS/SSEC, UW-Madison
• Brian Kahn
• Jet Propulsion Laboratory

MURI Hyperspectral Workshop7-9 June,
2005 UW-Madison
2
Outline

• Introduce midlevel/potentially mixed phase clouds
• MODIS IR phase retrievals for midlevel clouds
• Frequency of midlevel clouds
• Radiative transfer calculations for MODIS and
  AIRS for a variety of cloud heights

3
Midlevel Clouds

• Definition Clouds occurring at a level (height)
  where water cloud, ice clouds, or mixed-phase
  clouds could occur (potentially mixed-phase
  clouds).
• Typically between 253 and 263 K
• Cloud height depends on latitude and season
• 6 km in midlatitudes
• 1 km in polar regions (midlevel clouds a
  misnomer here)

4
ARM M-PACE 12 October, 2004
2200 UTC Aqua MODIS False Color Phase Image
Proteus flight track
Barrow
N. Alaska
False color cloud phase image R 0.65 µm
Ref., G 2.1 µm Ref., B 11 µm BT
5
ARM M-PACE 12 October, 2004
2200 UTC Aqua MODIS False Color Phase Image
Barrow
Slightly Broken Cloud
Uniform Cloud
N. Alaska
Ice
False color cloud phase image R 0.65 µm
Ref., G 2.1 µm Ref., B 11 µm BT
6
ARM M-PACE 12 October, 2004
2200 UTC Aqua MODIS 11 µm Brightness Temperature
Slightly Broken Cloud
Uniform Cloud
Ice
MODIS Band 31 11 µm BT (K)
7
ARM M-PACE 12 October, 2004
2200 UTC Aqua MODIS Infrared Cloud Phase
Retrieval
Slightly Broken Cloud
Uniform Cloud
Ice
MODIS IR Cloud Phase at 1 km
8
ARM M-PACE 12 October, 2004
2200 UTC Aqua MODIS Infrared Cloud Phase
Retrieval
Scene is unknown because BT11 µm is between
253 and 263 K, and BTD8.5-11 µm is between
-1.0 and 0.25 K
MODIS IR Cloud Phase at 1 km
9
ARM M-PACE 12 October, 2004What are we looking
at?
Barrow

• Lidar shows water cloud with ice precipitating
• MODIS 11 µm brightness temperature is around 258
  and 259 K

N. Alaska
10
ARM M-PACE 12 October, 2004
AHSRL Circular Depolarization Ratio ()

• Lidar shows water cloud with ice precipitating
• MODIS IR phase classifies scene as Unknown

11
ARM M-PACE 12 October, 2004
AHSRL Circular Depolarization Ratio ()
Question remains What is the infrared cloud
phase?

• Lidar shows water cloud with ice precipitating
• MODIS IR phase classifies scene as Unknown

12
12 October

• Averaged S-HIS spectra with averaged MODIS
  observations overlaid

MODISBand 32
MODISBand 31
MODISBand 29
13
Modeling

• Beginning set of modeling studies to characterize
  HSR IR cloud sensitivities for the M-PACE case

MODIS 8.5 µm
MODIS 12 µm
MODIS 11 µm
Simulations using LBLDIS
Still need to simulate mixed-phase and
overlapping cloud cases
14
How much of a problem are these potentially mixed
phase clouds?

• Unable to answer the question directly, but we
  can look at frequency of MODIS IR phase
  retrievals of ice, water, unknown.
• Consider 8 days (and nights) of global Aqua MODIS
  data from 1-8 April 2003.

15
Water CloudsCloudy pixels classified as liquid
water using IR MODIS channels
Between 60 S and 60 N, Water Cloud Frequency 52
16
Ice CloudsCloudy pixels classified as ice using
IR MODIS channels
Between 60 S and 60 N, Ice Cloud Frequency 28
17
Unknown CloudsCloudy pixels classified as
unknown phase using IR MODIS channels
Between 60 S and 60 N, Unknown Cloud Frequency
14
18
Mixed Phase CloudsCloudy pixels classified as
mixed phase using IR MODIS channels(mixed phase
should be considered another flavor of unknown)
Between 60 S and 60 N, Mixed Cloud Frequency 6
Scale is from0 to 0.5
19
Zonal frequencies of cloud phase(frequency out
of all cloudy data)
Mean cloud phase frequencies between 60S and
60N Water 52 Ice 28 Unknown 14 Mixed
(Unknown) 6
Sum of curves 1
Aqua MODIS 1-8 April, 2003
20
Zonal frequencies of cloud phase(frequency out
of all data)
Mean cloud phase frequencies between 60S and
60N Water 35 Ice 18 Unknown 9 Mixed
(Unknown) 4
Sum of curves Frequency of cloudiness
Aqua MODIS 1-8 April, 2003
21
Mid-talk Summary

• 20 of MODIS IR cloud retrievals are classified
  as unknown
• Because of the IR phase algorithm, these unknown
  classifications are likely due to midlevel or
  potentially mixed-phase clouds

22
The Question

• Can hyperspectral IR data improve our
  characterization of midlevel/potentially mixed
  phase clouds?

23
Brightness Temperature Differences for Water and
Ice Clouds
Midlatitude winter atmospheric profile
Optical thickness at 11 µm
MODIS sims from DISORT
AIRS sims from CHARTS
24
Brightness Temperature Differences for Water and
Ice Clouds
9 km, T 226 K7 km, T 238 K
3 km, T 262 K2 km, T 265 K1 km, T 269 K
Optical thickness at 11 µm
MODIS sims from DISORT
AIRS sims from CHARTS
25
Brightness Temperature Differences for Water and
Ice Clouds
? 1 K
Optical thickness at 11 µm
MODIS sims from DISORT
AIRS sims from CHARTS
26
Brightness Temperature Differences for Midlevel
Water and Ice Clouds
Optical thickness at 11 µm
MODIS sims from DISORT
AIRS sims from CHARTS
27
Brightness Temperature Differences for Midlevel
Water and Ice Clouds
? 0.5 K
Optical thickness at 11 µm
MODIS sims from DISORT
AIRS sims from CHARTS
28
Brightness Temperature Differences for Midlevel
Water and Ice Clouds
AIRS Sims show little sensitivity to water cloud
height
Optical thickness at 11 µm
MODIS sims from DISORT
AIRS sims from CHARTS
29
Summary

• 20 of MODIS IR cloud retrievals are classified
  as unknown
• Because of the IR phase algorithm, these unknown
  classifications are likely due to midlevel or
  potentially mixed-phase clouds
• Simulations shows that hyperspectral IR data can
  potentially differentiate between ice and water
  in midlevels (????????????

30
Future Plans

• Continue simulations for
• More atmospheric profiles
• Overlapping clouds (cirrus over water)
• Mixed-phase clouds
• Consider different channel combinations
• Apply what we learn from simulations to S-HIS and
  AIRS data from M-PACE
• Consider combined hyperspectral and MODIS phase
  algorithm