ESA Cloud-CCI Phase 1 Results Climate Research Perspective

Claudia Stubenrauch Laboratoire de Météorologie Dynamique, France

and Cloud-CCI Team

Outline

Ø Challenges to retrieve cloud properties

Ø What do we know about clouds from satellite retrievals and where does ESA Cloud-CCI stand ?

Ø How to use satellite cloud data for climate model evaluation?

Ø Challenges to build climatologies on cloud properties

Ø How to get a more complete cloud picture?

Copyright: 1998 Wadsworth Publishing Company; C. Donald Ahrens, Essentials of Meteorology

Cumulus (low fair weather clouds) Cumulonimbus (vertically extended)

Clouds are extended objects of many very small liquid / ice particles

Cirrus (high ice clouds) Cloud structures over Amazonia

bulk quantities at spatial & temporal scales

to resolve weather & climate variability

satellite radiometers

Space-based Global Observing System Geostationary satellites : frequent observation, no global coverage Polar orbiting satellites : observations at same local time, twice per day

ESA Cloud CCI observation strategy: 1) IR-NIR-VIS imagers (AVHRR) on polar orbiting satellites NOAA / Metop filled in with MODIS on Terra, Aqua & (A)ATSR on ENVISAT calibration of multi-satellites unique retrieval strategy based on Optimal Estimation

2) Combined retrieval using AATSR & MERIS on ENVISAT (2001-2012) OE & O2 A-band

3) IR sounder (IASI) on Metop (phase 2)

uses IR-VIS imagers on both, to resolve diurnal cycle

GEWEX cloud dataset

Ø information on uppermost cloud layers Ø  ‘radiative’ cloud height Ø perception of cloud scenes depends on instrument => cloud property accuracy scene dependent : most difficult scenes: thin Ci overlying low clouds, low contrast with surface (thin Ci, low cld, polar regions )

Cloud properties from space lidar – radar : vertical structure of clouds

IR-NIR-VIS Radiometers, IR Sounders, multi-angle VIS-SWIR Radiometers exploiting different parts of EM spectrum

How does this affect climatic averages & distributions ?

lidar, CO2 sounding, IR spectrum

IR-VIS imagers

solar spectrum

thin Ci over low clouds : Interpretation of Cloud height

≤ 20% of all cloudy scenes (CALIPSO)

ISCCP Cloud CCI

properties: (GCOS ECV’s) •  cloud amount CA (0.01-0.05) + rel. cloud type amount •  pressure/ height CP/CZ (15-50 hPa) •  temperature CT (1-5 K) •  IR emissivity CEM •  eff cloud amount CAE (= cloud amount weighted by emissivity) •  VIS optical depth COD •  Water path CLWP/CIWP (25%) •  eff part. radius CRE (5-10%)

1° x 1° monthly statistics per obs time: ● averages, ●monthly variability, ● histograms distinguish : tot, High, Mid, Low Water, Ice CP< 440 hPa, CP>680hPa CT>260K, CT<260 / 230K

Cloud Assessment co-chairs: C. Stubenrauch, S. Kinne

http://climserv.ipsl.polytechnique.fr/gewexca

2005-2012 (Stubenrauch et al. BAMS 2013, WCRP rep. 21/2012) Ø first coordinated intercomparison of L3 cloud products

of 12 global ‘state of the art’ datasets Ø database facilitates assessments, climate studies & model evaluation

CAHR + CAMR + CALR = 1

How many of detected clouds are high, midlevel & low clouds?

CAHR depends on sensitivity to thin Ci (30% spread) 42% are high clouds (COD>0.1) -> 20% with COD>2 (MISR, POLDER) thin Ci over low cloud misidentified as midlevel clouds by ISCCP, ATSR, POLDER 42% are single-layer low clouds, 60% are low clouds (MISR, CALIPSO, surface observer)

GEWEX Cloud Assessment key results and how does ESA Cloud CCI compare?

Cloud Amount : 0.68 ± 0.03 for clouds with COD>0.1

+ 0.05 subvisible Ci, -> 0.56 (clds with COD > 2)

synoptic (day-to-day) variability : 0.25-0.30 inter-annual variability : 0.025

0.10-0.15 larger over ocean than over land

global ocean-land

lidar, CO2 sounding IR spectrum

IR-VIS imagers

solar spectrum

CAHR + CAMR + CALR = 1

InterTropical Convergence Zone: high convection + cirrus anvils

stratocumulus

winter storm tracks

©1994 Encyclopaedia Britannica Inc.

Even if absolute values depend on Ci sensitivity, geographical cloud distributions agree!

MODIS-CE

uncertainty on regional variability:

9  

ESA Cloud CCI compared to GEWEX reference (xESACCI – <x>GEWEX)/ σxGEWEX

CA underestimation over ocean in 60N-60S (3-5σ from ref)

CAHR underestimation over land (2-4σ from ref) underestimation in SH midlatitudes (multi-layer?)

CALR underestimation in stratocumulus regions

AVHRR

10  

Height stratification according to pressure

Tropics, 1:30 – 3:00 PM, 2007

bimodal T/p distributions in tropics : not well observed by ESACCI retrieval of T, p or z: atmospheric profiles for T->p, p->T,z->T : retrieved, from reanalysis or forecast

high-level clouds: CP < 440 hPa

low-level clouds: CP > 680 hPa

Applications: evaluation of climate models

cloud radiative effects

12  

Comparison to climate models Satellite observations view clouds from above:

Ø passive remote sensing only gives information on uppermost clouds

Ø observations at specific local time

Ø instrument & retrieval method sensitivity, retrieval filtering, partial cloudiness may lead to biases

Climate models prescribe cloudiness per pressure layer (H2O saturation)

Ø clouds built from adjacent layers & max / random overlap per lat x long grid

ü filter local time, cloud detection sensitivity (in optical thickness)

ü cloud property grid averages from cloud overlap scheme

Satellite Simulators or simpler methods take care of these issues However, they can not repair insufficient instrument / retrieval sensitivity

13  

Comparison to climate models: latitudinal averages

land ocean

January/winter

CA

general good agreement of latitudinal variation in CA using several datasets as reference -> uncertainty in data

EC-Earth & ERA-Interim

LWP

IWP

LWP & IWP averages are much more difficult to compare!

Bulk microphysical properties

essential to be taken into account when comparing to models!

Single scattering properties in radiative transfer depend on thermodynamical phase / particle shape

Cloud Water Path: Liquid: 40 – 120 gm-2 Ice: 25 – 300 gm-2

ice

averages & distributions strongly depend on retrieval filtering & partly cloudy fields (MODIS-ST, ATSR retrieval filtering COD > 1, AIRS COD < 4)

ESA CCI COD-CP histograms compared to GEWEX reference

all ESACCI datasets are missing thin cirrus

AIRS-LMD

when evaluating cirrus processes in climate models, be sure to use a dataset sensitive to this kind of clouds !

REF

ATSR

AVHRR

AATSR-MERIS

CP

COD

L3 COD-CP histograms to determine cloud radiative effects

2) weight fluxes by COD-CP histograms (monthly 1° x 1° map resolution)

assessing cloud climatologies in terms of TOA fluxes (ESA Cloud CCI phase 2)

440

680

hPa 3.6 23 or radiative flux kernels of (Zelinka et al. 2012)

1)  determine radiative fluxes of cloud types over the globe, at different seasons

0.21 0.09 0.04 0.13 0.11 0.03 0.19 0.18 0.03

ISCCP 0.29 0.11 0.0 0.12 0.06 0.0 0.17 0.24 0.0

AIRS-LMD

differences in COD-CP distributions lead to differences in radiative effects (transformation of IR emissivity to COD -> COD < 10 => underestimation of SW effect)

Challenges in longterm monitoring

Ø climate change studies: be aware of temporal changes in coverage! Ø Interannual variability increases with decreasing Earth coverage!

Monitoring of Earth coverage / day at specific obs

Global CA / CT anomalies in time

global CA within ±0.025, CT within ±2K (~ interannual mean variability)

Investigation of possible artifacts in ISCCP cloud amounts (W. B. Rossow, Ann. 2 of WCRP report) Changes in radiance calibration, geographic & day-night coverage, satellite viewing geometry reduce magnitude of CA variation only by 1/3 merging different instruments / satellites challenging

-> look at histograms / regions

IR-VIS Synergy -> multi-layer clouds

IR Sounder - Imager Synergy: multi-layer situations in daylight

single-layer semi-transparent Cirrus (COD<3)

semi-transparent Cirrus above lowlevel clouds

from CALIPSO-ST :

IR sounding provides high-level & VIS provides low-level clouds

Conclusions & thoughts for discussion ESA Cloud CCI: §  common retrieval strategy (Optimal Estimation) on AVHRR,MODIS, AATSR, AATSR-MERIS §  L2 -> L3 processing will be done as in GEWEX cloud assessment

Ø comparison with GEWEX Cloud Assessment database: 1) AVHRR correlates in general well in seasonal / regional variation 2) promising, but still improvement needed in detection of thin cirrus & stratocumulus

Ø Retrieval development iterative process: analyses -> problems in some variables (averages or histograms) -> feedback to teams -> correction by teams & sending in new data

Ø comparison with climate models : models profit to compare to several datasets, being aware of different sensitivities

Ø Phase 2 & beyond: assessing cloud radiative effects, longterm, IR-VIS synergy -> 3-dim cloud distribution, interaction aerosols –clouds ? (needs transport studies)

Ø independent longterm dataset / cooperation with GEWEX,… Ø synergy observation & modelling groups -> analysis strategies

seasonal cycle : ESACCI compared to GEWEX reference

seasonal cycle agrees in general well with other datasets

in SHtropics less sensitivity to high-level clouds

good correlation for AVHRR in 60N-60S

less good for AATSR

including MERIS improves correlation

correlation in time over one year AATSR AVHRR AATSR-MERIS

CA

AATSR-MERIS

24  

Thermodynamic phase & retrieval of optical / microphysical properties

Retrieval of optical / bulk microphysical properties needs thermodynamic phase distinction: •  polarization (POLDER, CALIPSO) •  multi-spectral (PATMOS-x, MODIS, ATSR) •  temperature (ISCCP, AIRS, TOVS)

RVIS -> COD RVIS & RSWIR -> COD & CRE (smaller particles reflect more) assumptions in radiative transfer: particle habit, size distribution, phase

WP = 2/3 x COD x ρ x CRE (vertically hom.)

IR: small ice crystals in semi-transparent Ci lead to slope of CEM’s between 8 & 12 µm

Choi et al. 2010

CALIPSO

Zheng, PhD 2010

POLDER