Trends & Variability of Liquid Water Clouds from Eighteen Years of Microwave Satellite Data: Initial Results 6 July 2006 Chris O’Dell & Ralf Bennartz University.

Trends & Variability of Liquid Water Clouds from Eighteen Years of Microwave Satellite Data:

Initial Results

6 July 2006

Chris O’Dell & Ralf Bennartz

University of Wisconsin-Madison

Talk Outline• Motivation

• Description of sensors & retrieval product

• Mean climatology & comparison with ERA40, ISCCP

• Diurnal cycle

• Long-term trends?

Motivation for a cloud liquid water path (LWP) climatology

• Anthropogenic trends in cloud properties are possible, due both to global warming and aerosol effects.

• A robust LWP climatology can serve as a benchmark for global climate models.

• The 18-year passive microwave record contains a robust and independent measure of liquid clouds.

Previous successes from passive microwaves:

• Increases in lower tropospheric temperature, decreases in stratospheric temperatures from Microwave Sounding Unit (Mears et al. 2003, Christy et al. 2003, Vinnikov & Grody 2003)

• Increases in global (especially northern hemisphere) water vapor path (Trenberth et al., 2005)

Microwave sensors measuring LWP

Sensor - Platform Dates Used Ascending Overpass Time

SSM/I – F08 1/1988 – 12/1991 06:14

SSM/I – F10 12/1990 – 11/1997 07:38 – 10:26

SSM/I – F11 12/1991 – 5/2000 05:00 – 07:38

SSM/I – F13 5/1995 – 12/2005 05:39 – 06:33

SSM/I – F14 5/1997 – 12/2002 08:49 – 08:16

TMI - TRMM 12/1997 – 12/2005 N/A (equatorial orbit)

SSM/I – F15 12/1999 – 12/2005 09:33 – 08:42

AMSRE - Aqua 6/2002 – 12/2005 13:30

All instruments are conical scanners, with footprints ~ 40 km

Calibration/Retrieval

• All satellites have been intercalibrated – the radiances are consistent from one satellite to the next (RSS, unpublished!)

• All satellites use the same, modern retrieval algorithm to simultaneously retrieve LWP, water vapor path, and surface wind speed.

• Probably better than older algorithms which often used 2 channels to retrieve a given quantity, algorithms tended to be regression-based, and tended to retrieve different quantities independently.

LWP Agreement between sensors is good

No global trend with simple average!

Processing Scheme

At Remote Sensing Systems• Retrieved LWP binned daily onto a 0.25º grid (1440x720) for both

morning & evening overpasses• Even pixels with heavy rain retrieve LWP (but not water vapor or

surface winds)

At Wisconsin• Quantities further binned to 2.5º grid, monthly average for each

sensor & local overpass time.• Monthly diurnal cycle fits made for each pixel (average of all years).• Diurnally-corrected monthly means calculated for each pixel.• Seasonal & annual LWP trends calculated for each pixel.

MAM Mean LWP

Does microwave LWP agree with ERA40?

Does microwave LWP agree with ERA40?

Problem in Sc regions!

Does microwave LWP agree with ISCCP*?

Problems at higher latitudes

Ice?

* ISCCP D3 water path (WP)

Diurnal Cycle Fitting

• Goal is to make a diurnally-corrected LWP climatology

• Previous work with TRMM only retrieved diurnal cycle for tropics. Possible midlatitude diurnal cycle?

• For each 2.5º pixel & month, fit local time versus LWP to this function:

( corresponds to 24 hours)

• Use resultant fits to correct each monthly binned observation.

)(2cos)(cos),( 2211 ttAttALtzLWP

LWP Diurnal Cycle Strength

Wood et al., Geophysical Research Letters, 29 (23), 2002

F13

F11

F14F15

SSM/I

TRMM-TMI

Liquid Water Path [kg/m2]

Local Time [hours]

Normalized Diurnal Amplitude

Liquid Water Path [kg/m2]

Local Time [hours]

Liquid Water Path [kg/m2]

Local Time [hours]

Liquid Water Path [kg/m2]

Local Time [hours]

Liquid Water Path [kg/m2]

Local Time [hours]

Liquid Water Path [kg/m2]

Local Time [hours]

Robust Regional Trends in LWP?

Local trends – tropical western pacific

Local trends – northern midlatitudes

Local trends – southern mid-high latitudes

Local trends – Arctic ocean

Conclusions Existing passive microwave observatinos appear to

provide a stable, long-term record for climate studies of liquid clouds.

ERA40’s cloud parameterization seems to poorly characterize LWP seasonal and interannual variability in the subtropical high stratocumulus regions.

The diurnal cycle of LWP has been well-characterized in most ocean locations, and is generally in agreement with previous studies.

Initial studies of LWP trends are promising, with hints of regional trends (especially in the northern high latitudes), but no significant long-term global trend.

“To Do List”

Principle component analysis - may reveal interesting patterns of variability or problems with the data set.

Further investigation of the derived diurnal cycles – how constant are they from year-to-year? How well do they compare with CA & precip diurnal cycles?

More sophisticated statistical analyses of LWP trends…(hint to audience for guidance)

Make the complete LWP climatology available on the web.