You are here
Harvard Forest Symposium Abstract 2011
- Title: Digital repeat photography for phenological research: statistical methodology and camera choice
- Primary Author: Koen Hufkens (Harvard University)
- Additional Authors: Koen Hufkens (Harvard University); Andrew Richardson (Northern Arizona University); Oliver Sonnentag (Harvard College (Harvard University)); Cory TesheraSterne (Mount Holyoke College); Adam Young (SUNY College of Environmental Science and Forestry)
- Abstract:
Phenological research has been improved by continuous automated monitoring of vegetation canopies using digital cameras and webcams. Most cameras used for this purpose have a native capture system in the red (R) - green (G) - blue (B) color space, which can be used for simple visual inspection but also for separate extraction of color information as RGB digital numbers that allow for quantitative analysis of vegetation status. In this research we compare image archives obtained at Harvard Forest (fall 2010) with different digital cameras including indoor and outdoor webcams, game- and plant-cams, digital point-and-shoot and single-lens reflex cameras. The goal was to assess the impact of image quality on color information for phenological research. In addition, using image archives from eleven different forest ecosystems in the United States and adjacent Canada, we developed a protocol based on moving window quantiles to extract daily time series of canopy greenness that maximizes phenological information content but that suppresses the influence diurnal, seasonal and weather-related changes in illumination intensity. Our results suggest that image quality and thus camera choice is of secondary importance compared to the technique used to extract robust time series of canopy greenness for phenological research.
- Research Category: Forest-Atmosphere Exchange
