HR: 09:15h
AN: B41E-04 [Abstracts]
TI: Challenges for Characterizing Land Surface Phenology
AU: * Reed, B C
EM: reed@usgs.gov
AF: USGS, Flagstaff Science Center
2255 North Gemini Drive, Flagstaff, AZ 86001
United States
AB:
Considerable progress in developing phenology metrics from satellite-based multispectral sensors has been made over the past
25 years. A number of approaches using time-series vegetation index (VI) data collected at a synoptic scale of observation
have been used to produce metrics that correspond to different phenological stages of vegetation, such as the start of the
growing season, end of growing season, length of growing season, and others. However, there are a number of challenges in
this field that must be faced head-on in order to advance the field of land surface phenology, such as 1) addressing
non-vegetation related impacts on the satellite signal (e.g., effects of variable soil moisture, plant litter, atmospheric
contamination); 2) characterization of evergreen environments; 3) identification and characterization of sparsely vegetated
areas with poorly defined seasonality, and 4) identification and characterization of multiple growing seasons.
Progress toward meeting these challenges include clarifying the requirements of the phenology user community (decision
makers, modeling community, etc.) to produce meaningful satellite estimates, creative time-series processing methods to
remove non-vegetative impacts on the signal (such as temporal VI smoothing, cloud screening, and snow/ice masking), and a
more thorough understanding of the ecology of the communities that are being characterized. It is clear that the growing
season has different meaning for different ecosystems (e.g., croplands, shrublands, and evergreen forests).
Perhaps the most important challenges facing remote sensing phenology studies are 1) identifying the surface biophysical
characteristics that are being estimated by the various approaches, and 2) scaling ground observations of phenology to the
scale observed by the near-daily sensors (250 to 1000m ground resolution). A pilot study comparing satellite phenology
estimates to CO2 flux data gathered by the USDA-ARS Agriflux network in the western US grasslands show a strong agreement
regarding the timing of the growing season. Further investigations where remote sensing phenology measures are compared to
the timing of these and other directly measured fluxes need to be undertaken for improved understanding and validation.
Recent progress in forming a National Phenology Network (NPN), which will collect multiple levels of detailed phenology data
across the country, promises to support large area scaling efforts by collecting ground observations of indicator species and
other naturally occurring plants across the country that can be compared to satellite-scale observations.
DE: 0429 Climate dynamics (1620)
DE: 0438 Diel, seasonal, and annual cycles (4227)
DE: 0476 Plant ecology (1851)
DE: 0480 Remote sensing
SC: Biogeosciences [B]
MN: Fall Meeting 2005