Daily Temperature Oscillation Enhancement of
Multitemporal LST Imagery
George Ch. Miliaresis
Abstract
This paper addresses a major limitation of remote sensing
in biophysical modeling-capturing the diurnal temperature
range (
DTR
) with global datasets at moderate resolution
scale.
DTR
relates to the variation in temperature that oc-
curs between daytime and nighttime daily temperatures. A
new context for
MODIS
land surface temperature modeling is
proposed on the basis of the 01:30, 10:30, 13:30, and 22:30
local time Aqua and Terra acquisitions. First, cubic spline
interpolation produces an image time series with a uniform
six hours sampling per day. Second, the inverse Fourier
transform considers the harmonics with period less than or
equal to a day, to reconstruct a new image time series, and
enhance
DTR
. Finally, cluster analysis of the reconstructed
data set identifies eight clusters that are spatially arranged
into a Southern and a Northern group. The temporal varia-
tion for each cluster reveals a season dependent
DTR
that
is a key issue in supporting land cover studies in Greece.
Introduction
Currently, biophysical data sets are computed from the satel-
lite-based remotely sensed images with high temporal resolu-
tion at a moderate resolution scale allowing the day and night
monitoring of Earth’s surface. Among the biophysical parame-
ters, temperature is the most significant one in environmental
analysis studies (Withers
et al.,
2009); an example being the
land surface temperature (
LST
) acquisitions from MODerate-
resolution Imaging Spectroradiometer (
MODIS
) on board the
two
EOS
satellites, Terra and Aqua (Wan, 2007). Data analysis
of long time series of
MODIS LST
images provide information
about the spatial and temporal changes in temperature. For
example, Miliaresis (2009) mapped thermal invariant regions
in both space and time from
MODIS LST
imagery, while the
elevation, latitude, and longitude decorrelation stretch of
multitemporal
LST
imagery revealed thermal anomalies across
vast regions (Miliaresis, 2012, 2013, and 2014).
The moderate resolution biophysical data sets can be used
to forward many research questions in agricultural sciences.
For example, Sakamoto
et al.
, (2010), used day and night
digital images for monitoring the seasonal changes in crop
growth, while Vintrou
et al.
(2012) proved that agricultural
landscape can be characterized with a set of coarse resolu-
tion satellite-derived metrics. Such efforts seek to better
understand the links between agricultural activities and the
variety of impacts that climate change may have on food
supplies (Thenkabail,
et al.
, 2012). Due to climatic change,
night temperatures are expected to increase at a faster rate
than day temperatures due to less radiant heat loss because
of increased cloudiness (Alward
et al
., 1999). The duration
of a crop growth cycle is conditioned by the daily tempera-
tures absorbed by the plant. Therefore, an increase in daily
temperature will speed up plant development by reducing
the duration between sowing and harvesting (Hertel
et al
.,
2010). Thus, crop productivity may fall with the shortening
of a cycle. High night temperature decreases production by
decreasing the photosynthetic function (Turnbull
et al
., 2002).
Diurnal temperature range (
DTR
) is a
term
that relates to the variation in temperature that occurs be-
tween daytime (maximum) and nighttime (minimum) daily
temperatures (Hughes
et al.,
2007). Daily temperature oscilla-
tion is greatest in the planetary boundary layer, particularly,
and it is of importance in agriculture (Mohammed
et al.,
2009).
DTR
is controlled by many factors including: latitude,
land cover, elevation, atmospheric circulation, clear skies,
and the intensity of solar radiation (Lu
et al.,
2006). For
example,
DTR
generally increases on high mountain plateaus
and with distance from the sea (Weber
et al.,
1997), while
low lying humid areas typically present the least temperature
oscillations (Tang
et al.,
2006). The day and night temperature
oscillations are also controlled by thermal inertia (a measure
of the subsurface’s ability to store heat during the day and
reradiate it during the night), which facilitates lithologic map-
ping on planet Mars, by the Mars Global Surveyor Thermal
Emission Spectrometer (Mellon
et al
., 2000; Kirk
et al
., 2005).
Although spatial modeling of
DTR
from meteorological
stations can provide reliable estimates (Hill, 2013), meteoro-
logical stations are often too sparse to make reliable estimates
by interpolation. Towards this end, the capturing the diurnal
temperature range (
DTR
) with daily global datasets is per-
formed with geostationary weather satellites. For example,
Meteosat have been used to drive energy balance models of
sensible and latent heat flux (Rosema
et al
., 2013). Meteosat
data set consists of hourly thermal infrared images. Although
it provides higher temporal resolution than
MODIS
, the low
spatial resolution limits the land cover mapping applications.
Towards this end,
MODIS
near-diurnal
LST
oscillations
(day minus night
LST
) were computed per pixel for the 01:30
(night) and 13:30 (day) local crossing time passes of the Aqua
satellite for the year 2008 in Greece (Miliaresis and Tsatsaris,
2011) and the temporal and the spatial characteristics were
mapped. These estimates do not represent the diurnal temper-
ature range since 01:30 local time
LST
is not minimum while,
13:30 local time
LST
is not maximum. Two more thermal
images for the 10:30 (day) and the 22:30 (night) local crossing
time passes of Terra satellite are acquired daily.
The aim of the current research effort is to apply a technique
to all four daily
MODIS
LST
acquisitions from both the Aqua and
the Terra satellites, in order to produce a new multitemporal
LST
Open University of Cyprus, Faculty of Pure & Applied Sciences,
P.O. Box 12794, Latsia 2252, Cyprus (
)
Photogrammetric Engineering & Remote Sensing
Vol. 80, No. 5, May 2014, pp. 423–428.
0099-1112/14/8005–423
© 2014 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.80.5.423
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
May 2014
423
