Monitoring of cloud properties has been demonstrated for cloud altitude and appears promising in
the case of the phase (liquid or frozen) of cloud particles. Research to identify and remove the effects of
thin cirrus cloud contamination in Earth surface observations is in progress.
SPECIFIC COMMENT: MODELS AND INVERSION.
Models describing the ID, 2D and 3D light scattering process in plant canopies have been available for
some years. Several recent models have coupled the radiation transport processes in the atmosphere and
plant canopies. Recent models, which now often include the effects of the hotspot, have been based upon
the radiation transport equation, geometric-optics methods, radiosity theory and semi-empirical approaches.
And in fact, the state of the art continues to advance with the plethora of new models. Even so, there are
now so many models that the world should never lack for want of one.
Efforts to intercompare models have lagged perhaps and deserve the increased attention of the
co mmuni ty. Continued acquisition and sharing of high quality data suitable for model intercomparison and
validation tests should continue.
Progress has been made in the development of tractable models for inverting canopy reflectance
properties to estimate physical parameters of a canopy — such as leaf area index and canopy water content.
Solution of the inversion problem, insidiously difficult because canopy reflectance properties are not
uniquely related to canopy physical properties, does appear feasible for estimating biologically important
variables for man y canopies. Practical, computer-efficient inversion models have been demonstrated. The
use of both spectral and directional variables in an inversion model appears especially promising.
Global climate models require estimates of plant canopy albedo, which have been obtained with the
aid of canopy models for the spectral BRDF. In an alternative approach, the canopy albedo has been
estimated from 'strings' of multi-directional, multi-temporal canopy radiance measurements, taking advantage
of a priori knowledge of how canopy BRDF varies with angle. The approach potentially could be applied
to satellite data.
The issue of score heterogeneity, important because the ground pixel size of satellite data will
soon range between 10 m and 6 km, was not addressed at this congress. The scale treated here is the parcel.
There has been only limited attention paid to inversion of satellite data having pixels larger than even a few
hundred meters. While the mixed pixel problem has received significant research scrutiny during the past
20 years, we believe the area merits additional research on inversion of data having large heterogeneous
pixels such as will be provided by such planned future sensors as MISER, MODIS and POLDER.
SPECIFIC COMMENT: MEASUREMENTS.
Significant progress has been made both toward collection of calibrated, multi-spectral, multi-directional
data sets supporting remote sensing research — and toward development of tools facilitating collection of
such data.
Recent campaigns such as FIFE and HAPEX/Sahel and planned future campaigns such as Boreas
have placed special emphasis upon collection of sets of high quality data describing the light scattering
properties of the land surface and atmosphere. We believe collection of such data should continue because
they have played a key role supporting remote sensing research. We applaud and support efforts to share
these data with the world-wide remote sensing community.
Development of the tools necessary for estimating the light scattering properties of the atmosphere
and land surface continues with the deployment of such sensors as POLDER, CASI, CEASER, MODIS
Airborne Simulator and improved versions of ASAS and AVIRIS, all aircraft instruments. An improved
version of the PARABOLA sensor for collecting ground based measurements will soon be available
commercially. Improved sun photometers and several other instruments for estimating atmospheric
properties have recently become available from the Cimel Company in France and the University of Arizona.
These instruments, which together provide usable, high quality tools for estimating the surface
BRDF/BPDF, are proving critically important for supporting the increased data collection needs of the
remote sensing community.
