INTRODUCTION
The National Aeronautics and Space Administration has established the Earth System Science Pathfinder (ESSP) Program within the Office of Mission to Planet Earth (MTPE) to provide a means by which small, scientific space missions can be proposed and developed by individual investigators and their teams in response to research priorities not adequately addressed by current space missions such as those in EOS. The ESSP philosophy is to identify low-cost (less than $90M), quick-turnaround missions of limited duration that will provide data required to answer focused questions of importance to earth system science. As such, the ESSP program has been described as another tool among those in the existing NASA observational tool kit for improved understanding and comprehension of the components and interactions of the Earth system. The ESSP program is on-going and anticipates yearly mission launches.In response to this opportunity, the Vegetation Canopy Lidar (VCL) mission was proposed by the University of Maryland, College Park, NASA Goddard Space Flight Center, and other university and industrial collaborators, to collect canopy and topography landcover data critical for terrestrial ecology and climate studies. In spring of 1997, VCL was selected as the first ESSP mission at a total cost of $59.8 M to NASA, along with the second mission, the Gravity Recovery and Climate Experiment (GRACE).
The datasets derived from the laser altimetry instrument will make a
unique and catalytic contribution to pressing environmental
issues -- climatic change and variability, biotic erosion and
sustainable landuse, and should dramtically improve our estimation of
global biomass and carbon stocks, fractional forest cover,
forest extent and condition, and provide canopy data critical for biodiversity
studies. In addition, through its characterization of land surface
canopy and topographic structure, VCL provides new knowledge of
biophyscial parameters critical for climate, hydrologic, and other
global modeling activities, activities that are essential for our ability
to plan and predict for global change in the next century.
Lastly,
VCL provides a dense, global network of accurate topographic heights,
including sub-canopy topography,
data which are invaluable for future topographic missions,
as well as for natural hazard and climate studies.
VCL is an active remote sensing system built around the Multi-Beam
Laser Altimeter (MBLA), a five-beam instrument with 25 m contiguous
along track resolution. The five beams are in a circular configuration
8 km across and each beam traces a separate ground track spaced 2 km apart,
eventually producing 2 km coverage between 65° N and S, with
orbit crossovers producing a denser grid away from the equator.
For understanding and managing these environmental issues,
there is a prerequisite
for describing landcover status and dynamics to initialize,
parameterize and validate modeling efforts.
This description is required in many forms.
To represent landcover in models of the climate system,
it must be first be parameterized with respect
to albedo, aerodynamic roughness,
and surface resistance to evaporation, among others;
these parameters being functions of vegetation
structure,
e.g. height, canopy closure, and leaf-area index (LAI),
and composition
(annual/ perennial, etc.).
Similarly,
for
biogeochemical modeling,
biotic erosion,
and land degradation,
the vegetation component of landcover must be
represented in structural terms (biomass)
and by composition
(annual/perennial, woody/non-woody, and their age structures).
These arguments are not new.
Indeed,
they have justified the
significant efforts by the global remote sensing community to better
characterize the global land surface.
The principal data source thus far for these
efforts has been that from the AVHRR instrument of the NOAA spacecraft series,
e.g. the NASA AVHRR Pathfinder Program.
The ISLSCP Initiative (Sellers et al. 1995) is one
example of the intensive use of
existing data to parameterize landcover
for modeling purposes.
While these AVHRR-derived datasets represent a significant
contribution,
there remains considerable uncertainty associated with the estimated
parameters.
For example,
important parameters such as height, LAI, woody biomass,
aerodynamic roughness, and surface resistance are poorly
estimated by existing methods using AVHRR data alone.
Nor will any of the EOS-suite of instruments provide the
requisite observations of these variables.
However,
given an independent and systematic sampling of vegetation
based on laser returns,
global data sets can be developed that will provide much improved
estimates of critical variables such as canopy height, aerodnamic roughness
and biomass.
Furthermore,
the recovery of explicit vegetation structure should provide new
insight into traditional measures of surface reflectance and structure,
obtained from passive remote sensing instruments,
such as NDVI, LAI, BRDF,
thereby extending the usefulness of the existing and planned
remote sensing archive.
VCL also provides a near-global dense reference
network of height transects including sub-canopy topography.
The importance of accurate topographic data for global change
and earth system science
is increasingly being recognized,
as evidenced by efforts to assimilate global topographic data,
declassify defense digital elevation data,
and the creation of new technologies for accurate
mapping (such as GPS and interferometric SAR).
This data set should prove invaluable for calibration and registration
of data from future topographic missions,
such as the Shuttle Radar Topography Mapper (SRTM),
as well as providing detailed information on topographic roughness
and scaling.
Heretofore,
the lack of broad scale data on vegetation morphology
has limited the implementation of canopy-related
processes into earth system models (Asrar and Dozier 1994),
and is a priority for earth system science.
Our crude knowledge about fundamental landcover characteristics,
especially their vertical and horizontal structure and gradients,
is remarkable given the level of sophistication and complexity
of many global modeling activities that rely on this data.
DeFries et al. (1995) have outlined the weaknesses
of current landcover schemes in their discreet mapping
of the land surface characteristics required to calculate
land surface parameters such as absorbed radiation,
albedo,
canopy conductance,
roughness,
photsynthesis,
transpiration,
net primary production,
carbon and nutrient dynamics.
Although further development and refinement of the
passive remote sensing techniques generally used for such
characterization may provide some improvement,
truly dramatic gains may be realized only with the advent
of active remote sensing systems,
such as laser altimetry,
which directly measure surface structure
Figure 1 . Return waveform for a laser pulse.
An incident gaussian laser pulse's interaction with surface structural
components leads to the distorted (relative to a gaussian)
return waveform or echo. Measuring the return travel time of
pulse gives distance from the sensor.
By knowing where the last return is from the ground,
shown as the strong pulse,
this distance can be translated into height above the ground.
The magnitude at any height (time) of the waveform is directly related
to the number of intercepting surfaces and their reflectance.
Thus where the amplitude of the waveform is larger implies
more canopy materials, for example.
Figure 2 shows typical laser returns over a vegetation canopy.
Canopy height is calculated by subtracting the
elevations of the first and last returns.
Vegetation height is
is a function of species composition, climate
and site quality,
and can be
used for land cover classification alone or in conjunction with
vegetation indices (e.g., NDVI).
Along track measurements, i.e., footprint-to-footprint, of VCL-derived
height variation provides additional information such
as fractal (Palmer 1988) or autocorrelative
(Cohen et al. 1990) properties of the canopy that further may be used to
differentiate among natural and anthropogenically-disturbed land
cover patterns (Krummel et al. 1987).
When coupled with species composition and site quality
information (e.g., edaphic and climatic variables),
height
serves as an estimate of stand age or successional state which
can be correlated to carbon flux rates (Ustin et al. 1993).
In addition to providing a unique metric, i.e., the vertical
dimension, to classify vegetative cover at global scales, height
is highly correlated with aboveground biomass
(Oliver and Larson 1990, Avery and Burkhart 1994, Nilsson 1996).
Biomass in forests represents the major reservoir of carbon
in terrestrial ecosystems that can be quickly mobilized by
disturbance or land use change (e.g., Houghton et al. 1987,
Dixon et al. 1994).
When combined with greeness measures
from other sensors,
such as TM, MODIS or AVHRR,
VCL observations may be used to determine whether the greeness
signal is the result solely of low-lying vegetation (via
the height distribution).
Many areas of the world have
ground covers with greeness indices
comparable to those of forests,
making landcover discrimination based on greeness
measures alone difficult.
Measures of the vertical organization of canopy components
are also critical for modeling factors that relate to biophysical
and micrometerological processes at the atmospheric-vegetation
boundary layer such as radiative transfer, evapotranspiration,
and trace gas flux (Gro 1993, Fournier et al. 1995).
In contrast VCL would provide direct measurements of canopy top height,
canopy base height,
sub-canopy ground roughness length,
and vertical density of interceptors (woody and non-woody).
Maps of these variables would then be available globally,
and gridded to a resolution as fine as 2 km @times@ 2 km.
These can then be used to derive the bulk aerodynamic
parameters at accuracies and spatial extents never before
possible,
with a resultant improvement in our ability to model
momentum,
water vapor and sensible heat fluxes.
In addition,
a global data base of canopy structure,
especially when combined with other remote sensing data,
such as greeness indices from MODIS,
should enhance our ability to model
the interaction of energy with the surface
via better estimates of LAI and FPAR
(fractional absorbed photosynthetically active radiation),
as well as the interaction of precipitation with the surface
through interception and retention,
with resulting improvements in model photosynthesis,
net primary production, trace gas and hydrologic fluxes.
VCL visits,
on average,
the same 1 cell at the equator every two weeks,
with more frequent revisits away from the equator.
The exact ground track through any cell is essentially random,
being a function of orbital drag and the monthly reboost
required to keep the satellite at altitude.
The number of visits to a 2 km cell,
globally averaged over all cells,
is approximately 10 during the
two-year mission
(with more frequent visits away from the equator,
and less frequent visits at the equator).
VCL science goals place further constraints on the measurements,
achieved through sensor
design, orbit configuration and mission lifetime.
Among these constraints are:
laser footprint of 25 m to adequately resolve
vegetation height and structure;
continuous along track returns to adequately characterize
horizontal spatial variability and vertical height distributions;
5 beams separated by 2 km across-track spacing to achieve
sufficient global sampling and exact positioning and pointing;
precessing, inclined orbital configuration such that the sensor
passes within the same 1 @times@ 1 block along the equator
every two weeks,
allowing for (a) seasonal refresh of canopy structure and,
(b) global 2 km coverage every 6 months, minimizing the effects of clouds.
Canopy measurements are achieved through analysis of laser pulse
returns (waveforms) and require that return signals
include the canopy top,
adjacent structural elements and,
eventually, the underlying ground.
Waveform analyses based on extensive airborne and spaceborne
laser altimetry have revealed the need for
footprint sizes on the order of 1 to 2
canopy diameters.
This guarantees a resulting reflection from the very
top of each canopy within the sampled area as well as sufficient intra-
and inter-tree gaps required to image the underlying ground.
Dynamic range
in the receiver as well as sufficient laser output energy are required to
detect small ( < 1%) weighted returns from the canopy top and,
in dense canopies,
the underlying ground.
So, potentially, even the most dense
canopies still reveal their heights and sub-canopy topography.
Smaller footprints under-represent the canopy structure
especially with respect to true height.
This
results from the reduced probability of sampling the top of
the canopy:
smaller footprints most often hit the shoulders of the canopy.
Conversely,
with larger footprints (beyond a few tens of meters),
similar to those proposed for space-borne missions of the future,
such as the Geoscience Laser Altimeter Sensor (GLAS) and Shuttle
Laser Altimeter (SLA) series,
the precent of the total return that is
contributed by the canopy top is greatly reduced
making height measurement inaccurate.
The footprint size selected for VCL will be especially suitable
for high biomass tropical forests where canopy diameters can reach
10 to 25 m.
The medium size footprint of VCL
also reduces any surface slope spreading of the ground return,
most notably
in the case where the surface slope, canopy height, and canopy thickness
combine to convolve the canopy return with the ground return.
A 65° orbital inclination was selected to
provide near-complete coverage of vegetation
areas of interest.
Such an inclination samples 98% of closed canopy forest,
and gives at least 92% coverage for all the major types
of close-canopy forest,
and at least 89% for all but one type of
open woodland.
The orbit and 2 km beam spacing allows VCL to cover the entire earth
between 65 N and S in less than 6 months,
so that over an 18 month mission VCL will visit the same general
2 km @times@ 2 km area 3-4 times near the equator
(with higher repeats at higher latitudes).
Cloud probability studies have shown that this leads to a 85-90% chance
of obtaining at least one clear pass through the area.
VCL acquires data 10 times over an average 2 km @times@ 2 km cell
during a 2 year mission,
accumulating 19.1 km lf laser ground tracks.
This samples 9.4% of an average cell's area,
or after adjusting for cloud cover,
4.7% of the land area between 65° N and S.
VCL's mission duration is driven by requirements for coverage
in the tropics,
where there are the greatest uncertainties in present land cover data.
The more frequent coverage where orbital paths converge at higher
latitudes yields measurements of canopy heights and
median intercepts during the summer growing season,
even where deciduous trees are in leaf for but a few months a year.
Baseline Instrumentation
The MBLA is comprised of five laser transmitters in a single altimeter
instrument.
Each laser beam
operates at the 1064 nm fundamental wavelength of the neodymium-doped
yttrium aluminum garnet (Nd:YAG) solid-state laser, are arranged in a
pentagon inside a 20 mrad telescope circular field-of-view that is
centered
on nadir as illustrated in Figure 3.
The optical telescope is 0.9 m in diameter composed of beryllium.
For the VCL orbital
altitude of 400 km the across-track separation between adjacent tracks is
2
km.
VCL makes simultaneous measurements of range to the surface by synchronous
triggering of the 5 laser pulse transmitters and detection with a single
telescope that is staring at nadir and is equipped with multiple silicon
avalanche photo diode detectors in its focal plane.
Individual laser footprints are 25 m in diameter and are
contiguous along-track, commensurate with the best DTED Level 2 topography
mapping resolution and LANDSAT Thematic Mapper pixel resolution. Surface
echoes from the 5 beams are digitized in the MBLA electronics at 250
Megasamples per sec to achieve the required sub-meter vertical resolution
in the vegetation canopy and permit pulse centroid correction of the range
measurement.
The MBLA pulsed laser transmitter modules are based on high power Nd:YAG
and employ the Q-switching technique to concentrate laser energy in a
short
pulse.
Each of these laser modules produces a laser pulse of 5 nsec
duration at the rate of 290 pps. Laser pulse energy of 10 mJ per pulse
will
be sufficient to establish a link performance for the MBLA instrument that
results in 95% probability of detection of the Earth's surface under clear
atmospheric conditions and permits surface lidar investigations.
Mission Overview
The priniciple goal of the Vegetation Canopy Lidar
is the characterization of the three-dimensional structure of
the earth: in particular, canopy vertical and horizontal structure and
land surface topography. The VCL mission has two main science objectives:
VCL measurements are used to derive a variety
science data products including canopy heights,
canopy vertical distribution, and ground elevations gridded monthly
at 1° resolution and and every 6 months at 2 km resolution,
as well as a 2 km fractional forest cover product.
These products are created at the VCL Data Center at the University
of Maryland and archived and distributed through a collaboration with
the EROS Data Center.
Motivation
The primary focus of the VCL mission is to provide quantitative
description of landcover and global productivity,
one of five science priorities of MTPE (Harris et al., 1996).
Landcover change or dynamics
is directly linked to several important environmental issues,
including climatic change and variability,
biotic erosion (loss of biodiversity),
and sustainable landuse.
Landcover is a first order component of general circulation models.
Its status and dynamics have a direct feedbacck on the climate system
by determining the boundary conditions of the
exchange of momentum, energy and mass between the atmosphere and
the land.
Landcover also exerts an indirect or trace gas feedback
because it represents a significant and dynamic pool in the
carbon cycle.
Similarly,
landcover dynamics are the principal driver of biotic erosion.
It is changes in landcover structure and composition that in turn,
determines changes in habitat suitability.
Any predictive understanding of
biotic erosion,
a necessary prerequisite for slowing and reversing that loss,
will be underpinned by a quantitative comprehension and representation of
current landcover status and the regime of disturbance by landuse.
SCIENTIFIC RATIONALE
Forests comprise 29% of the world's terrestrial sruface
and,
as indication by the seasonal changes in atmospheric CO2
(Denning et al. 1995),
these systems represent key dynamic components in the global carbon
cycle.
Besides harboring the bulk of the Earth's biodiversity (Erwin 1995),
forest canopies are the primary interfce between
terrestrial ecosystems and the atmosphere,
accounting for greater than 50% of the annual
CO2 flux (Potter et al 1993).
Knowledge of canopy structure is required for modeling processes
such as photosynthesis,
energy transfer,
and evapotranspiration at local to global scales.
Landcover characterization for terrestrial ecosystem modeling and prediction
Figure 2 . Profile of canopy height, canopy density and
subcanopy topography from an airborne laser altimeter (SLICER)
over a forest near Salisbury Maryland.
The individual waveform contains multiple distinct returns,
The the independence of canopy top and ground topography,
as well as the ability to detect the ground below the canopy.
Landcover characterization for climate modeling
Direct measurements of canopy height, canopy vertical and spatial
structure,
and ground topography would allow determination of landcover
properties critical to GCMs, SVATs and mesoscale models,
whether used for climate modeling or numerical weather forecasting.
There have been efforts to include complex land surface parameterization
schemes both on-line and off-line in these models,
as well as to understand the effects of various schemes on
model outputs (e.g. see the Project for Intercomparison of Landsurface
Parameterization Schemes - PILPS - activities).
The determination of bulk aerodynamic parameters
that control the transfer of energy,
mass and momentum between the atmosphere and the surface,
roughness length, zero-plane displacement,
and canopy and ground resistances,
is generally regarded as a major source
of uncertainty.
For example,
in SiB2 (Sellers et al. 1996) 1 @times@ 1 maps of
the bulk aerodynamic parameters are found by first using
a global landcover classification.
Mean values of
eight canopy and ground morphological and physical parameters
(canopy top height,
canopy base height,
ground roughness length,
leaf-area density inflection height,
leaf width and length,
leaf-angle distribution factor and leaf area index)
are assigned from the literature based on landcover type.
These variables are then used with estimates of LAI from
satellite data to derive the bulk parameters at 1 @times@ 1
spatial resolution.
This procedure can only be seen as inadequate
given the sparsity of literature values for canopy variables
and their gross spatial and categorical generalization.
Globally distributed topographic control points
The strong scientifc need for accurate,
global topographic data bases
has led to recent progress in
limited release of portions of the Defense Mapping Agency Digital Terrain
Elevation Data (DTED) Level 1 data that has 90 spatial (horizontal) pixel
resolution and 16 m vertical accuracy.
Despite this progress the
scientific benefit for global MTPE studies is not nearly realized.
The existing DTED Level 1 data will not be released for the entire Earth and
space-based imaging sensors now in orbit and planned in the EOS-AM1 era
(1998-2003) require global topography at DTED Level 2
(30 m spatial, 16 m vertical)
for full realization of their science potential in land
cover/global productivity, short-term climate modeling, and natural hazard
studies.
The Shuttle Radar Topography Mission (SRTM) has been
announced to address these needs.
Since IFSAR data,
or the
conventional photogrammetric data sets,
are only relative in their
measurement of surface elevation,
direct measurements are needed to
"control" the vertical dimension of the topographic image.
Estimates of
TCPs needed for a global DTED Level 2 are well
in excess of 100,000,000. Only a limited number of these TCPs can be
provided by ground-based radar targets and GPS receivers, the remainder
will have to be estimated from existing maps and digital elevation models
that do not routinely achieve the 1 meter level of vertical accuracy and
do not have a common reference frame.
The VCL surface elevation measurements will have a common, global
reference frame and are "direct" rather than inferred. Furthermore, only
VCL will address the vegetation cover issues that limit all present mapping
techniques to tens-of-meters rms in forested areas.
The VCL Mission, by
virtue of its primary vegetation canopy measurements,
will provide billions of
sub-canopy surface elevation points.
MISSION DESCRIPTION
VCL is scheduled for launch in early 2000 on board a Pegasus XL
launch vehicle.
The VCL mission will be conducted by means of a small satellite carrying
the MBLA instrument in a 400 km orbit of 65°
inclination with a two-year nominal lifetime.
This will provide
sufficient coverage of the Earth to characterize the vegetation canopy
structure on a global basis during two complete growing seasons and produce a
global reference grid of land topography.
Because of increased atmospheric drag caused by the solar maximum
during the mission,
monthly reboosts are required to maintain nominal orbital altitude.
Command and control of the spacecraft during operation,
as well as data processing will take place at the University
of Maryland.
Distribution and archiving of VCL data products will be performed
by the EROS Data Center.
Table 1 lists VCL data characteristics and quality.
Measurement Objectives and Requirements
To address its major science goals,
the VCL mission has 3 direct measurement objectives listed above,
all at sub-meter vertical accuracies:
(1) canopy top height;
(2) the vertical distribution of nadir intercepted surfaces;
and,
(3) surface topography elevation,
including sub-canopy topography.
Figure 3 .
VCL mission concept.