Arctic Cloud Properties and Radiative Forcing From Observations and Their Role in Sea Ice Decline Predicted by the NCAR CCSM3 Model During the 21st Century - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 46

“umbrella” effect: melting of ice promotes greater cloud for-
mation, which reduces the amount of SW radiation absorbed
by the ocean. This diminishes the ice-albedo feedback and
mitigates the increase in surface temperature. Second is the
“greenhouse” effect: increased cloudiness and atmospheric
temperatures accompanying the sea ice reduction in the Arctic
enhance LW cloud warming of the surface. This extra energy
causes earlier and/or greater snowmelt, which, in turn, trig-
gers the ice-albedo feedback by reducing the surface albedo.
In the present chapter we use satellite and ground-based
observations to estimate and compare the effects of clouds
on the shortwave and longwave radiation. Further, we use
the NCAR CCSM3 coupled global climate model to exam-
ine the change of sea ice and clouds during the 20th and
21st centuries. The chapter is structured as follows: Section
2 describes observations and the model. Section 3 reviews
the studies based on observations subdivided into two parts:
section 3.1 discusses the effects of sea ice and clouds on
the SW fluxes at the top of atmosphere using satellite data,
and section 3.2 reviews Arctic cloud radiative properties
using ground-based observations. Section 4 presents the
CCSM3 model simulations of sea ice, clouds, and radiation
during the 20th and 21st centuries. The summary and the
discussion of the study are given in section 5.

associated with seasonal variations in sea ice concentrations.
Significant improvements in the radiative flux retrievals in
the polar regions are achieved in the new product from the
Clouds and the Earth's Radiant Energy System (CERES)
program, which is currently becoming available after a se-
ries of validations [ Kato et al. , 2006]. However, the TOA
albedo from the CERES data differs significantly from the
ERBE [ Bender et al. , 2006] and requires more comparison
with observations prior to be used in sensitivity studies. The
choice in the present study is given to the widely used ERBE
data set, taking into account its shortcomings.
Sea ice concentration (SIC) data used for the analysis of
sea ice effects on the TOA shortwave fluxes are from the UK
Met Office Hadley Centre's sea ice and sea surface tempera-
ture data set (HadISST1) available from 1870 to the present
on a 1° latitude-longitude grid [ Rayner et al. , 2003]. Begin-
ning in 1978, the data are derived from Special Sensor Mi-
crowave/Imager and the Scanning Multichannel Microwave
Radiometer [ Gloersen et al. , 1992]. The microwave radi-
ance data have a monthly averaged SIC error of about 7%,
increasing up to 11% during the melt season [ Gloersen et al. ,
1992]. The biases are greatly reduced in the HadISST1 ho-
mogenization process using other satellite and in situ sea ice
concentration and sea ice extent data [ Rayner et al. , 2003].
Cloud and surface radiative flux ground-based data are ob-
tained during the Surface Heat Budget of the Arctic Ocean
(SHEBA) program conducted on an ice floe that drifted more
than 1400 km in the Beaufort and Chukchi seas between 74°N
and 81°N latitudes and between 165°W and 140°W longitudes
from late October 1997 to mid-October 1998 [ Uttal et al. ,
2002]. The major meteorological and surface energy budget
measurements were conducted on multiyear pack ice with
summertime melt ponds and occasional nearby leads [ Per-
ovich et al. , 2002; Tschudi et al. , 2001]. Upwelling and down-
welling fluxes were measured using broadband radiometers
[ Persson et al. , 2002]. The uncertainty in the downwelling
and upwelling LW flux was estimated at ±2.5 W m -2 and at
±4 W m -2 for the net LW radiation. The estimated uncertainty
in the downward and upward SW flux is ±3% with a bias from
-5 to +1 W m -2 for the downward and 0 to -6 W m -2 for the
upward SW flux [ Persson et al. , 2002]. Cloud presence and
base height were derived from a combination of lidar and ra-
dar measurements [ Intrieri et al. , 2002b]. Column-integrated
liquid water path (LWP) was derived from brightness tem-
peratures measured by microwave radiometer with 25 g m -2
accuracy [ Intrieri et al. , 2002b; Westwater et al. , 2001].

2. DATA AND METHODOLOGY

2.1. Observational Data

A suite of accurate up-to-date data of Arctic cloud proper-
ties, radiative fluxes, and sea ice are used in this study. Short-
wave radiative fluxes and albedo spectrally integrated over
the 0.2-5.0 mm band at the top of the atmosphere (TOA) de-
rived from the Earth Radiation Budget Experiment (ERBE)
data are used to analyze sea ice effects on the TOA shortwave
radiation for all-sky conditions. The program combines the
ERBS, NOAA 9, and NOAA 10 satellite measurements for
the period from November 1984 to February 1990 [ Bark-
strom et al. , 1989; Barkstrom and Smith , 1986]. We use the
narrow field of view product with a spatial resolution of
2.5° ´ 2.5°. Global error for ERBE monthly SW fluxes is
5.5 W m -2 [ Wielicki et al. , 1995]. Larger errors in the polar
regions (up to 20 W m -2 for all-sky fluxes and 50 W m -2 for
clear-sky fluxes) are caused by inaccuracies in surface scene
and cloud identification over sea ice (a review of the ERBE
data errors is given by Gorodetskaya et al. [2006]). Large un-
certainties in clear-sky identification over ice surfaces make
separate analyses of the data for clear-sky conditions and
cloudy-sky conditions unreliable [ Li and Leighton , 1991].
For all-sky data, the errors contribute to the scatter, but they
are substantially smaller than the changes in the SW fluxes

2.2. NCAR CCSM3 Model

The output from the 20th and 21st century simulations of
the NCAR CCSM3 global coupled model is examined to