DETECTION OF PERIODIC CLIMATE ANOMALIES OVER GREENLAND
WITH MICROWAVE RADIOMETERS
Wolfgang Wagner‘, Per Gloersen?
! Student, Institute for Photogrammetry and Remote Sensing, Vienna University of Technology
? Senior Scientist, Oceans and Ice Branch, Laboratory for Hydrospheric Processes, NASA Goddard Space Flight Centre
Commission VII, Working Group 8
KEY WORDS: Radiometry, Change Detection, Spectral Analysis, Greenland, Ice Sheet
ABSTRACT
We investigate the potential of microwave radiometers for the detection of periodic climate anomalies over Greenland. A multiple-
window harmonic analysis technique is applied to a nine year record of Scanning Multichannel Microwave Radiometer (SMMR)
data. We concentrated on the 37GHz vertically polarized channel to obtain a good correlation between the signal and ice crystal
size over the skin depth. Year to year differences in microwave emission from the percolation facies of the Greenland Ice Sheet are
explained by the absence or varying degree of surface melt during summer. Quadrennial oscillations were found to occur more
frequently in the brightness temperature record than oscillations with periods of one or two years.
1. INTRODUCTION
The Greenland ice sheet is the largest ice sheet in the Northern
Hemisphere. The present ice sheet has a surface area of
1,750,000 km? (Steffen et al., 1993), that is twenty one times
the size of Austria and roughly the size of Mexico. Little is yet
known about the Greenland ice sheet despite its important role
for climate studies and for climate itself (Thomas, 1993).
Research has been hampered by its vast size and the harshness
of the polar environment. Remote sensing satellites are able to
provide measurements over large areas and will thus contribute
to our understanding of the Greenland ice sheet. The
usefulness of microwave radiometers, like the Scanning
Multichannel Microwave Radiometer (SMMR), for the study
of the polar ice sheets has already been demonstrated. The
most prominent application of microwave radiometers is
detection of surface melting on the ice sheets. Since moisture
in the near-surface firn causes a marked increase in microwave
brightness temperatures (Ty), very high values of T, indicate
surface melting. Thereby the extent and duration of surface
melting can be derived. Presently long term-series of such
information are being established (Zwally and Fiegels, 1994).
In this study we investigated the potential of microwave
radiometers for the detection of periodic climate anomalies
over Greenland. By studying the spatial and temporal
variabilities of geophysical parameters, global patterns can be
depicted which makes it possible to link weather and climate
anomalies in one region of the globe to another (Lau and
Busalacchi, 1993; Gloersen, 1995).
2. METHOD
To find periodic signals in the microwave emission from the
Greenland ice sheet we applied a novel harmonic analysis
technique to a nine year record of SMMR brightness
temperatures. The harmonic analysis technique was developed
for investigating seismic, atmospheric carbon dioxide, and
other time series records (Park et al., 1987; Kuo et al. 1990).
Several names can be found for it in the literature. We will
refer to this technique as the ,multiple-window harmonic
analysis". The multiple-window harmonic analysis aims for
finding the harmonic components of a time series. The method
is novel in that it uses several windows to fit a sinusoid model
to the data and that it utilizes a F-test statistic to test the fit of
the sinusoid model.
We analyzed data from the Scanning Multichannel Microwave
Radiometer (SMMR) which was operated on board of the
Nimbus 7 satellite from October 26, 1978 to August 20, 1987.
The SMMR is a 10-channel instrument receiving both
horizontally (H) and vertically (V) polarized radiation at 5
frequencies. Only the 37 GHz V and H channels were
considered because of the greater sensitivity of the 37 GHz
data to near-surface changes compared to the longer-
wavelength data (Schuman et al., 1995). The spatial resolution
of the data at 37 GHz is approximately 30 km. The SMMR
instrument was generally operated on an every-other-day basis.
Our data set thus comprised 1611 images for both the 37 GHz
vertically polarized channel and the 37 GHz horizontally
polarized channel. The SMMR data have been mapped onto a
rectangular grid with a grid cell size of 25 km x 25 km over a
stereographic projection. For a detailed description of the
SMMR see Gloersen et al. (1992).
Brightness temperature time series were established for each
grid cell. The time series were detrended assuming that the
trend is linear. Following the procedures described in the
Appendix ,,Multiple-window harmonic analysis“ we calculated
the complex amplitude, u, of the sinusoid model and the F-test
parameter, F, from the detrended data. The sum over time is
taken on six-day intervals instead of the original two-day
intervals to match the revisit time of the SMMR. The complex
amplitude gives information about the magnitude and phase of
the sinusoid model immersed in the time series, and the F-test
744
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
p?
CO
ca
an
ye
tel
sk
tei
bl:
VO
rej
(R
ref
de
Th
de
to
ur
tra
lay
19
cri
(M
Do
sul
po!
lay
der
sca
cre
tha
lar,
adc
Th
hoz
Init
pol
are
fin:
the
gro
ice
pol
V-s
the
Inte
she
To.
doe