3 Role of SAR in Surface Energy Flux Measurements Over Sea Ice
45
nisms for first year ice scattering during the advanced melt period are also largely
unknown.
In this section we introduced the scientific rationale for this work within the context of global climate change and operational ship navigation in ice-infested waters.
We then described the average seasonal conditions of the physical, electrical, energy
balance and microwave-scattering characteristics of the marine cryosphere within
the microwave-scattering seasons proposed by Livingstone et al. (1987). In the next
section we investigate the statistical linkages between microwave scattering and various components of the energy balance. Dielectrics and physical properties are then
used to explain the physical mechanisms responsible for the statistical relationships.
These linkages then constitute a precursor to developing image products (Sect. 3.4)
based on the time series evolution of synthetic aperture radar (SAR) scattering coefficients.
3.3
Building Statistical Relationships
The research results presented in this chapter have been collected during the Seasonal
Sea Ice Monitoring and Modelling Site (SIMMS) experiment. Members of this project
team collected data annually over various ice types located in the Canadian Arctic
Archipelago between 1990 and 1995. Details of the experiment are available elsewhere
(LeDrew and Barber 1994).
3.3.1
Univariate Analysis
A time series of ERS-l data were calibrated and used to produce a summary view of the
seasonal evolution of (TO for a first-year sea ice surface (first-year ice site) during
SIMMS'92. In this analysis we focus on two time series: the "early" seasonal period (days
105-130, approximately the month of May) and a portion of the "transitional" and
"accelerated" periods (days 140-170, approximately the month of June).
The conductive components of the energy balance appear to exhibit a reasonable
association with the seasonal evolution of (TO in time series 1. The daily totals of Qco,
show no covariance with (TO until after about day 120 (Fig. 4). The pattern of the QCi (Fig.
5) more closely approximates the pattern of (To, suggesting that the snow volume may
be having little or no effect relative to the ice surface on (TO during this period. In time
series 2 Qco follows a similar pattern to cro while QCi appears to be nearly constant. It
appears from this relationship that changes in the snow cover both affecting and affected by Qco may be detected by the SAR, whereas QCi appears to be independent of the
observed changes in (TO in time series 2.
In general, as the conductive flux decreases there is a decrease in (TO. The similarity
in pattern between QCi and (TO in time series 1 suggests a slight covariance between the
ice surface flux and (TO. In time series 2 the covariance of Qco with (TO and the lack of
association with QCi suggests that (TO may be arising (at least partially) from the snow
volume. It is important to note that Qco began varying in phase with (TO after day 120
and QCi stopped varying with (TO after about day 150. During the "early" seasonal period the snow surface is well insulated from the thermodynamics of the underlying vol-
45
nisms for first year ice scattering during the advanced melt period are also largely
unknown.
In this section we introduced the scientific rationale for this work within the context of global climate change and operational ship navigation in ice-infested waters.
We then described the average seasonal conditions of the physical, electrical, energy
balance and microwave-scattering characteristics of the marine cryosphere within
the microwave-scattering seasons proposed by Livingstone et al. (1987). In the next
section we investigate the statistical linkages between microwave scattering and various components of the energy balance. Dielectrics and physical properties are then
used to explain the physical mechanisms responsible for the statistical relationships.
These linkages then constitute a precursor to developing image products (Sect. 3.4)
based on the time series evolution of synthetic aperture radar (SAR) scattering coefficients.
3.3
Building Statistical Relationships
The research results presented in this chapter have been collected during the Seasonal
Sea Ice Monitoring and Modelling Site (SIMMS) experiment. Members of this project
team collected data annually over various ice types located in the Canadian Arctic
Archipelago between 1990 and 1995. Details of the experiment are available elsewhere
(LeDrew and Barber 1994).
3.3.1
Univariate Analysis
A time series of ERS-l data were calibrated and used to produce a summary view of the
seasonal evolution of (TO for a first-year sea ice surface (first-year ice site) during
SIMMS'92. In this analysis we focus on two time series: the "early" seasonal period (days
105-130, approximately the month of May) and a portion of the "transitional" and
"accelerated" periods (days 140-170, approximately the month of June).
The conductive components of the energy balance appear to exhibit a reasonable
association with the seasonal evolution of (TO in time series 1. The daily totals of Qco,
show no covariance with (TO until after about day 120 (Fig. 4). The pattern of the QCi (Fig.
5) more closely approximates the pattern of (To, suggesting that the snow volume may
be having little or no effect relative to the ice surface on (TO during this period. In time
series 2 Qco follows a similar pattern to cro while QCi appears to be nearly constant. It
appears from this relationship that changes in the snow cover both affecting and affected by Qco may be detected by the SAR, whereas QCi appears to be independent of the
observed changes in (TO in time series 2.
In general, as the conductive flux decreases there is a decrease in (TO. The similarity
in pattern between QCi and (TO in time series 1 suggests a slight covariance between the
ice surface flux and (TO. In time series 2 the covariance of Qco with (TO and the lack of
association with QCi suggests that (TO may be arising (at least partially) from the snow
volume. It is important to note that Qco began varying in phase with (TO after day 120
and QCi stopped varying with (TO after about day 150. During the "early" seasonal period the snow surface is well insulated from the thermodynamics of the underlying vol-