the major findings related to climate and phenology connections, some of the
limitations of current work, and how the advances in LSP research will help
address these issues and lead to greater utility.
The earliest documented work we could find pertaining to remote sensing and
phenology was a progress report from Texas A&M University on a project
sponsored by NASA’s Goddard Space Flight Center in the early 1970s (Rouse
et al. 1974). That study concludes that satellite data can provide a quantitative
description of vegetation conditions as phenological indicators for seasonal and
climate effects. It is interesting to note this study was based on ‘‘Landsat 1’’
(originally called ERTS-1, Earth Resources Technology Satellite 1) Multispectral
Sensor (MSS) data. The satellite was launched on July 23, 1972 and the report was
written in the fall of 1973. While this is an impressive turnaround time, it is also an
indication that, at least for some researchers, satellites clearly offered an approach
to connect vegetation growth to climate drivers. Many similar (relatively) local
studies using Landsat and other 10–30 m spatial resolution sensors have been done
since then. However, to better understand climate and vegetation phenology
connections, LSP studies needed to expand to the continental and global scales.
Within a few years of the earliest demonstration of how AVHRR could be used
to monitor vegetation health (Gray and McCrary 1981, Schneider et al. 1981,
Townshend and Tucker 1981), researchers started evaluating LSP from that sensor
at regional scales far larger than then current Landsat studies. As an example of
some of the earliest work on continental LSP, Justice et al. (1986) used AVHRR to
monitor 1 year of phenology in Kenya. Particular emphasis was placed on quantifying the phenology of the Acacia Commiphora bushlands. Considerable variation was found and explained through the high spatial variability in the distribution
of rainfall and the resulting green-up of the vegetation. They explored the relationship between rainfall and NDVI using meteorological stations existing within
the bushland, which shows that the early AVHRR work also considered the
relationship between satellite-observed LSP and annual weather patterns (Justice
et al. 1986; Justice 1986).
Building on this and other studies demonstrating the utility of AVHRR to
monitor vegetation for larger areas over time, much of which was led by the
Global Inventory Monitoring and Modeling Studies (GIMMS) group at NASA/
Goddard Space Flight Center (GSFC) (Justice 1986), NOAA and NASA initiated
the AHVRR pathfinder program to produce global, 8 km NDVI. The pathfinder
data were processed using the best available methods at that time to produce a
consistent time series of data; including cross-satellite calibration, navigation
using an orbital model and updated ephemerides, and correction for Rayleigh
scattering. The data were made openly available to the community as both daily
and composite data. Analysis of this initial global time series provided insight into
terrestrial processes, seasonal and annual variability, and methods for handling
large volume data sets and facilitated land surface phenology (James et al. .). As
such, it truly did establish a path for subsequent global land vegetation products
from MODIS, MERIS, and the Visible Infrared Imaging Radiometer Suite
(VIIRS).
116
J. M. Hanes et al.
limitations of current work, and how the advances in LSP research will help
address these issues and lead to greater utility.
The earliest documented work we could find pertaining to remote sensing and
phenology was a progress report from Texas A&M University on a project
sponsored by NASA’s Goddard Space Flight Center in the early 1970s (Rouse
et al. 1974). That study concludes that satellite data can provide a quantitative
description of vegetation conditions as phenological indicators for seasonal and
climate effects. It is interesting to note this study was based on ‘‘Landsat 1’’
(originally called ERTS-1, Earth Resources Technology Satellite 1) Multispectral
Sensor (MSS) data. The satellite was launched on July 23, 1972 and the report was
written in the fall of 1973. While this is an impressive turnaround time, it is also an
indication that, at least for some researchers, satellites clearly offered an approach
to connect vegetation growth to climate drivers. Many similar (relatively) local
studies using Landsat and other 10–30 m spatial resolution sensors have been done
since then. However, to better understand climate and vegetation phenology
connections, LSP studies needed to expand to the continental and global scales.
Within a few years of the earliest demonstration of how AVHRR could be used
to monitor vegetation health (Gray and McCrary 1981, Schneider et al. 1981,
Townshend and Tucker 1981), researchers started evaluating LSP from that sensor
at regional scales far larger than then current Landsat studies. As an example of
some of the earliest work on continental LSP, Justice et al. (1986) used AVHRR to
monitor 1 year of phenology in Kenya. Particular emphasis was placed on quantifying the phenology of the Acacia Commiphora bushlands. Considerable variation was found and explained through the high spatial variability in the distribution
of rainfall and the resulting green-up of the vegetation. They explored the relationship between rainfall and NDVI using meteorological stations existing within
the bushland, which shows that the early AVHRR work also considered the
relationship between satellite-observed LSP and annual weather patterns (Justice
et al. 1986; Justice 1986).
Building on this and other studies demonstrating the utility of AVHRR to
monitor vegetation for larger areas over time, much of which was led by the
Global Inventory Monitoring and Modeling Studies (GIMMS) group at NASA/
Goddard Space Flight Center (GSFC) (Justice 1986), NOAA and NASA initiated
the AHVRR pathfinder program to produce global, 8 km NDVI. The pathfinder
data were processed using the best available methods at that time to produce a
consistent time series of data; including cross-satellite calibration, navigation
using an orbital model and updated ephemerides, and correction for Rayleigh
scattering. The data were made openly available to the community as both daily
and composite data. Analysis of this initial global time series provided insight into
terrestrial processes, seasonal and annual variability, and methods for handling
large volume data sets and facilitated land surface phenology (James et al. .). As
such, it truly did establish a path for subsequent global land vegetation products
from MODIS, MERIS, and the Visible Infrared Imaging Radiometer Suite
(VIIRS).
116
J. M. Hanes et al.