The University of Bristol worked on the line to demonstrate the ability to link ground based spectral
measurements to causal biophysical properties (Harrison et al., 1993a). They mainly studied semi-natural
vegetation and were able to discriminate up to five common species by means of spectral reflectance
measurements. After sampling at one location the instrument projected Field-of-View (FOV) was
determined and marked on the upper surface of the vegetation sample. This area was recorded by video
camera to allow the composition of the projected FOV to be determined. Destructive sampling through the
column formed by projecting this area through the vegetation sample to the ground was carried out.
Biophysical parameters collected within this column include LAI, leaf chlorophyll data and leaf moisture
content. LAI measurements were made by harvesting the leaves within the FOV and calculating their
surface area from classified video images. Leaf surface area divided by the area of the instrument FOV
gives LAI. Surface Leaf Area (SLA) was determined by classifying all the directly irradiated leaves within
the FOV. Total chlorophyll absorption values were obtained spectrophotometrically according to
Mackinney (1941). There appears to be a strong relationship between SLA and NDVI which seems to be
species specific. For the species examined, NDVI seems to be more highly correlated to SLA than LAI.
By examining the spectra using factor decomposition techniques, the results show that there are four
statistically significant factors causing the variance amongst the population of spectral curves. Brightness is
factor 1, major source of variation and factor 2 appears to be reacting to changes in the ratio of visible and
near infrared reflectance, which is related to vegetation amount. As a consequence of this study, correlations
were carried out between the weighting of factor 2 and the SLA for all samples. The results are very
promising, suggesting that there is a non-species specific relationship between factor 2 and SLA. The next
major step is to extrapolate these ground based properties to the remotely sensed images.
The Remote Sensing Group of the University of Valencia carried out a field radiometry campaign at
the three sites, Barrax, Rada de Haro and Tomelloso, in order to spectrally characterize the main land uses
to discuss the factors which determine reflectance values and to study their influence in the intercomparison
of reflectances determined at different places and at different times (Gilabert et al., 1993b). This work was
carried out within the framework of a monitoring campaign where reflectance measurements were obtained
from the main crops, following phenological cycles, and measuring every 16 days, nearly synchronously
with LANDSAT-5 overpasses. Simultaneously, different plant parameters were measured by the University
of Castilla-La Mancha. The main goal of this work was the study of correlations between NDVI and
agronomical parameters (LAI and biomass). Some parcels were associated to irrigated crops in the Barrax
area and show the typical spectral behaviour of green vegetation. Other parcels were associated to dry crops
in Tomelloso and their spectral response does not exhibit the typical absorption by chlorophyll, so
indicating the presence of yellow vegetation and a greater contribution of the soil background response.
Some other measurements were made in the Rada de Haro area, mainly corresponding to natural vegetation
surfaces.
A very homogeneous bare soil area was selected at the Barrax site to take ground measurements for
reference and for estimation of correction factors for atmospheric effects. These measurements were taken
just at the time of satellite and aircraft overpasses. That bare soil spot was measured on June 12 and 28 (at
LANDS AT overpassing time), and on June 29 (at ER-2 overpass, for reference of TMS and AVIRIS data).
Correlations between NDVI and agronomical parameters were analyzed for each parcel separately,
taking into account all the values obtained during the phenological cycle. In general, results from the 1991
campaign showed the suitability of the NDVI to characterize the different crops as a function of their
biophysical characteristics. In 1992, a follow-up field radiometry campaign was carried out as complement
of the 1991 measurements, this time on a com canopy (Gilabert et al., 1993a). The main findings of this
experiment are also presented at this Symposium by Gilabert and Meliâ, (1994).
The Free University of Berlin carried out spectral albedo measurements at the end of April, during
June and at the end of August, with concentration in June, 1991 in the Rada de Haro region (Billing et al.,
1993a). For some of the most characteristic vegetation types, the change of albedo during the vegetation
period was determined. Directional reflectance measurements were also taken to interprète observed spectra
features in the discrete albedo measurements and to relate these measurements to satellite and aircraft data.
They could also study the daily course of broadband albedo of a wheat field in April and bear soil and vine
in August, in the Tomelloso area. In June they studied matorral vegetation and a vine field in Rada de Haro.
In addition to the "meteorological" albedo, the albedo of the PAR was separately measured for a vine field
at the Rada de Haro site in June. The PAR sensor was mounted on the radiation balance station to measure
incoming PAR from where they could derive the PAR absorbed by the vegetation canopy (APAR).
2.2 - Longwave Infrared Measurements
The Groupe de Recherche en Télédétection et Radiométrie (GRTR) from University of Strasbourg
carried out radiometric field measurements in the thermal infrared (Nerry, 1993). By means of two
BARNES-PRT5 radiometers and using the box method they could measure broadband emissivity for
different kinds of bare soil, com, barley, vineyard, sunflower, etc. Spectral emissivity signatures were
obtained at the laboratory and they carefully studied the three typical bare soils of the main sites. The two
