Development of a low-cost network of webcams for monitoring plant phenology in a Chihuahuan Desert shrubland
Phenology is the study of the timing of natural events such as the schedule of plant life-cycle events like bud burst, flowering, seed set and senescence. Many factors control plant phenological dynamics including climate. As such, phenology has been shown in published literature to be one of the best biological indicators of climate variability and change. Recent advantages in digital photography and digital image processing have opened new opportunities for inexpensive, repeatable and automated capture of plant and landscape phenological dynamics. This technology, however, remains poorly studied in terms of its capacity to depict the development of different plant phenophases, and correlate with indices typically used on satellite platforms for remotely sensing land surface properties such as above ground green plant biomass. The overarching goal of this thesis is to develop and test a network of phenocams on the Jornada Experimental Range in a northern Chihuahuan Desert shrubland ecosystem. This study links: (1) field based phenophase monitoring (2) output from a network of webcams, and (3) spectral indices from a Robotic tram system that measures hyperspectral reflectance. Phenophase development of five dominant plant species differed between species and between years for the same species. The timing of bud break and the presence of green leaves for dominant shrub species closely matched trends in landscape phenology derived from a network of four Microsoft Vx7000 webcams, which were found to be the most cost effective and spectrally sound camera of four different models tested. Greenness indices derived for determination of landscape phenological trends were calculated with new software co-developed with graduate students in the Department of Computer Science and UTEP's Cyber-ShARE Center of Excellence. Greenness indices derived from the webcams correlated well with the normalized difference vegetation index (NDVI) derived for two shrub species (Prosopis glandulosa, honey mesquite; and Larrea tridentata, creosote bush) from a robotic tram system that was used to measure hyperspectral reflectance with a Unispec dual channel spectrometer on a weekly basis for 110 sampling points along a 110 meter elevated tramline. Correlations were poor for bare ground and two species of graminoids). Although observations were not sustained over multiple years, thereby preventing isolation of specific climatic conditions controlling phenology in this ecosystem, the delay in 2011 phenophase development documented for all species is likely to be a result of drier and warmer conditions experienced during the 2011 growing season and an extreme freeze event which occurred at the study site in early 2011. This study has established a baseline set of phenological measurements and tested and developed a relatively unique assemblage of research infrastructure that will be sustained beyond the study period of this thesis. Based on findings that species responded differently in their phenophase development, and the same species demonstrated a hypothesized response to climate variability; it is likely that sustained phenological observations will facilitate isolation of climatic factors controlling phenophase development and dynamics of landscape phenology. These findings may have strong implications to forecasting the future state of northern Chihuahuan Desert ecosystems, and ecosystem properties and processes such as biodiversity and carbon balance.
Gonzalez-Alonso, Libia, "Development of a low-cost network of webcams for monitoring plant phenology in a Chihuahuan Desert shrubland" (2011). ETD Collection for University of Texas, El Paso. AAI1503722.