Environmental Engineering Reference
In-Depth Information
satellite data. The logical approach to achieve satellite measurement consistency is
to perform rigorous cross-calibration intercomparisons, which are needed to estab-
lish traceability.
As the technology of satellite observations of the Earth matured, user
requirements and expectations for satellite data have also evolved not only for
climate but also for weather applications. Although qualitative applications of
satellite data, such as generating cloud maps, have less stringent requirement for
calibration accuracy, most quantitative applications require accurately calibrated
satellite data, and this need has rapidly evolved in the last decade. For example, in
the early days of meteorological satellites, numerical weather prediction (NWP)
required retrieved quantities of vertical temperature, moisture, and water vapor
profiles from satellite data. In the past decade, NWP has witnessed the greatest
evolution of direct radiance assimilation from satellite measurements into numeri-
cal weather prediction models. As a result, resolving the observation versus model
differences, aka, the biases, becomes a prerequisite for direct radiance assimilation.
Satellite observations with unresolvable biases will be rejected by the NWP models.
Therefore, the importance of instrument calibration for satellite applications
cannot be overstated. Calibrated radiances are the fundamental building blocks for
all satellite products, including the radiances for data assimilation in NWP, reanaly-
sis, and fundamental climate data records for climate change detection. Calibration is
the centerpiece of data quality assurance in satellite data processing, distribution, and
archive and is part of the core competency of any satellite program.
2.2 Satellite Instrument Calibration Methodologies
Calibration is the process of quantitatively defining the sensor responses to known
and controlled signal inputs. These signals should ideally be traceable to established
reference standards. Traceability requires the establishment of an unbroken chain of
comparisons to stated references each with a stated uncertainty. It should be noted
for satellite sensors on-orbit, the calibration signals may become neither well
known nor controllable. Also, operationally, calibration is the process of converting
the Earth observation raw signals to physical quantities to generate SDR (sensor
data records) or Level 1b data.
Calibration is generally divided into three areas: radiometric, spectral, and
spatial calibration. Radiometric calibration focuses on the accuracy and traceability
of the radiometric quantity of the measurement. Spectral calibration ensures that the
spectral responses of the system are accurate at the operating conditions and
changes over time are well known. Spatial calibration ensures the geometric
distortion is well characterized by a number of metrics including the point spread
function or modulation transfer function. Spatial calibration also ensures the band-
to-band co-registration and geo-location/navigation for each pixel. The radiometric,
spectral, and spatial calibrations go hand in hand (Fig. 2.1 ). A problem in either the
spectral or spatial calibration introduces uncertainties in the radiometric calibration.
Search WWH ::




Custom Search