PE&RS November 2015 - page 861

An Effective Antarctic Ice Surface Temperature
Retrieval Method for MODIS
Tingting Liu, Zemin Wang, Xin Huang, Liqin Cao, Muye NIU, and Zhongxiang Tian
Abstract
Given the close relationship between surface melt in polar
areas and ice surface temperature (IST), it is important to
develop an effective IST retrieval method. However, the studies
concerning this topic are relatively limited. In this context, this
paper proposes an effective approach to retrieve IST in the
Antarctic area by presenting a modified split-window algo-
rithm (SWA) and introducing a polynomial fitting for atmo-
spheric transmittance simulation. The effectiveness was quan-
titatively validated by a comparative study with a Moderate
Resolution Imaging Spectroradiometer (MODIS) IST product
(MOD29) and automatic weather station (AWS) data. The com-
parisons indicated that the proposed method shows a robust
performance in Antarctic IST retrieval for MODIS data: the
bias was −0.61 K and the root-mean-square error (RMSE) was
1.32 K for the Zhongshan Station data set; the bias was −1.62 K
and the RMSE was 2.34 K for the Ross Ice Shelf data set.
Introduction
The study of the polar areas has been the subject of great in-
terest in recent decades, as the polar areas are sensitive to the
global environmental change (GEC) and interact closely with
many environmental factors in the GEC systems (Wynne and
Lillesand, 1993; Peter et al., 1999; Harvey and Green, 2013).
Among the various changes in the polar areas, the issue of
surface melt deserves special attention and discussion since
it contributes to the accelerated disintegration of the ice sheet
(Zwally et al. , 2002). Surface melt has a direct relationship
with surface temperature. Consequently, extracting accurate
ice surface temperatures (ISTs) can provide further insight into
the process of the surface melt.
Although a lot of automatic weather stations (AWSs) have
been established in the polar areas (Steffen et al. , 1996; Laz-
zara et al., 2012), they are far from sufficient for IST moni-
toring in the broad polar areas. Compared with AWS data,
remotely sensed imagery can provide fruitful and timely data
for Earth observation across both time and space. In particu-
lar, the application of thermal infrared (TIR) data can offer a
more practical approach to IST information extraction. Recent
studies have seen an increasing application of TIR data to land
surface temperature (LST) retrieval (Tang et al., 2008; Xiao
et al., 2008; Peña, 2009; Nichol, 2009; Rajasekar and Weng,
2009; Jimenez-Munoz et al., 2014).
Research into TIR-based LST estimation has experienced
significant progress in recent decades. Multispectral TIR sen-
sors (e.g., the Moderate Resolution Imaging Spectroradiometer
(MODIS)) have gradually replaced other sensors (e.g., the Scan-
ning Multichannel Microwave Radiometer (SMMR) on Nimbus-7
(Comiso, 1994), the Along-Track Scanning Radiometer (ATSR)
on ERS -1 (Stroeve et al., 1996), and the Advanced Very High
Resolution Radiometer (AVHRR) on NOAA (Leshkevich et al.,
1993, Bolgrien et al., 1995; Key et al., 1997; Veihelmann et al.,
2001)) as the main data source for LST estimation. Furthermore,
the split-window algorithm (SWA ) has been found to be an ef-
ficient approach for LST estimation. Recently, a few studies have
investigated TIR-based IST estimation. Hall et al. (2004) pro-
posed a procedure to produce a standard MODIS-based polar sea
ice product suite (MOD 29). The relationship between melt sea-
son IST and the mass balance of Greenland (2000 to 2005) was
analyzed in the research of Hall et al. (2006). The difference
between IST products and AWS data over Greenland was further
analyzed by Hall et al. (2008), to address the uncertainties and
limitations associated with the existing IST products. Other
studies have focused on ice surface emissivity (Hori et al., 2006;
Hwang et al., 2008), IST and ice extent analysis (Liu et al., 2009;
Sobota, 2011; Shu et al., 2012; Hall et al., 2013), IST-based ap-
plication (Ciappa et al., 2012), and so on. However, an effective
IST estimation process for the Antarctic area is still required.
Most of the existing SWAs (Coll et al ., 1994; Franca and Crack-
nell, 1994; Thenkabail et al., 2007) require some parameters and
coefficients that are difficult to estimate in the real world, since
the in situ data used to calculate the parameters and coefficients
are often quite difficult to obtain in the Antarctic area.
In this paper, to address these shortcomings, an effec-
tive SWA-based approach is proposed for IST retrieval in the
Antarctic from MODIS data. The notable advantages of the
proposed method include the following two aspects:
1. Development of an effective IST estimation method for the
Antarctic area. First, the modified SWA developed by Qin
et al. (2001) is introduced into the IST retrieval, which
requires only two parameters (surface emissivity and
atmospheric transmittance), without the complicated esti-
mation of other coefficients and parameters. Furthermore,
in the procedure of atmospheric transmittance estimation,
the relationship between water vapor and atmospheric
transmittance is simulated by a series of polynomial func-
tions, replacing the traditional linear fitting.
2. Validation of the proposed method based on AWS data.
Although Qin’s method (2001) has been successfully
used in many areas (Mao et al., 2005), its effectiveness
in the polar areas has not yet been verified. For this
purpose, in this study, the method is modified and
Tingting Liu, Zemin Wang, and Muye NIU are with the
Chinese Antarctic Center of Surveying and Mapping, Wuhan
University, 129 Luoyu Road, Wuhan, P.R. China
(zmwang@whu.edu.cn).
Xin Huang is with the State Key Laboratory of Information
Engineering in Surveying, Mapping and Remote Sensing,
Wuhan University, 129 Luoyu Road, Wuhan, P.R. China.
Liqin Cao is with the School of Printing and packaging,
Wuhan University, 129 Luoyu Road, Wuhan, P.R. China.
Zhongxiang Tian is with the National Marine Environmental
Forecasting Center, 8 Dahuisi Road, Beijing, P.R. China.
Photogrammetric Engineering & Remote Sensing
Vol. 81, No. 11, November 2015, pp. 861–872.
0099-1112/15/861–872
© 2015 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.81.11.861
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING November 2015 861