PRT5s were used to study the spatial, spectral and temporal variabilities of brightness temperature along 10
km transects representative of land use and vegetation types at the Barrax site.They measured brightness
temperatures along the Barrax-La Gineta road on different days and at different times of the day, in
approximate coincidence with satellite overpasses. In June, 11 and 12, they performed seven transects of
brightness temperature in approximate coincidence with NOAA-IO, NOAA-11 and LANDSAT-5.
One of the transects was performed almost in coincidence with LANDSAT-5 overpass in June 12.
This experiment aimed to quantify the scaling process by direct comparison of data at the highest TIR
spatial resolution available from space, that is, 120 m. This transect was identified on the TM scene and
after correcting for atmospheric effects, the result of the comparison can be seen in Fig. 1.
OSTANO Cm)
IOOO 2000 5000 *000 5000 6000 7000 6000
■ E2—“7 s-»», 0 — ^ —
Fig. 1. Comparison of ground and satellite (TM) transects.
Emissivity measurements and temperature transects were also made in the Barrax area by the
University of Valencia to validate methods developed in the Group for measuring emissivities in the field
and for estimation of atmospheric effects. They were also aiming at the definition of effective emissivity
and temperature at NOAA scale. Emissivity measurements were mainly performed on barley and maize for
dry and irrigated crops. With these measurements, emissivity maps at LANDSAT and NOAA scales were
also obtained. For temperature, and coinciding with LANDSAT and NOAA overpasses, transects were
made for homogeneous surfaces (maize, barley and bare soil), and across the heterogeneous area along
Barrax-La Gineta road. Surface brightness temperature was corrected for emissivity and atmospheric
reflection effects.
3 - EVALUATION OF VISIBLE AND NIR SATELLITE DATA
The intercomparison of measurements of different remote sensing systems with each other and with
ground measutements has identified three major problem areas for the "portability" of remote sensing
methods and these have been addressed by EFEDA (Billing et al., 1993b):
- The calibration of the sensors is not very well documented and not quickly enough available for
operational use. This has required some calibration efforts during the field campaign itself.
- Atmospheric correction methods without aerosol information is still a severe problem.
- The strong and from ecosystem to ecosystem highly variable dependence of angular reflectance on
the illumination and the observing geometry requires more detailed investigations for a variety of land-
surface types. Furthermore, for many ecosystems, albedo, needed for radiation budget computations, is
rather ill-defined because of its strong dependence on the solar zenith angle and cloudiness.
3.1 - Remote Sensing Images and Corrections
3.1.1. Calibration and compatibility of spectral channels. Two different extrapolation formulas exist for
AVHRR, those of Kaufman et al. (1992) and of Che and Price (1992). Since calibration involves the same
corrections discussed here, it is difficult to determine how accurate these calibrations are (Billing et al.,
1993b). For June 1991, the two calibration coefficients depart by 8% for ch. 2 and zero for ch. 1. As for
LANDSAT-TM, Brockmann (1992) is of the opinion that data distributed by ESA (which are used in this
study) are not corrected for sensor degradation and suggests that TM data are in error by -7% to -10%.
