Document: SAR-3-39-22

Image analysis in neighborhood studies of plant competition over time: Test of methods in populations of a small winter annual, Draba (Erophila)verna L.

EMERY, S.M.* and J.C.MULROY

Denison University, Granville, OH 43023 USA ¹

Abstract:
For plants and other sessile organisms, measurements of mean behavior may be insufficient to describe interactions among individuals. Consequently, ecologists have turned to methods such as neighborhood analysis to characterize the environment around each individual in a population. In competition studies, for instance, neighborhoods are defined by the size and distance of neighbors within some arbitrarily or experimentally defined "zone of influence." While this approach can provide a more precise analysis of plant-plant interactions, most studies have included a relatively small number of individuals, a single time interval, and/or a simplified experimental design. Image analysis allows acquisition of large data sets that can be analyzed across a number of time intervals and potential zones of influence. We explored the capabilities of serial photography and image analysis to measure neighborhood effects on individual plant growth over time using the small rosette-forming winter annual Draba verna L. (Brassicaceae). We took photographs from fixed overhead positions at weekly intervals throughout the life cycle in several experimental and natural populations of Draba verna. The public domain image analysis program, NIH Image, its companion program, Object-Image, and customized macros allowed characterization of plant size (leaf area) and location at each date. Automated measurement functions of NIH Image then allowed creation of a spreadsheet containing distance and size of all neighbors in relation to each individual at each sample date. In subsequent analyses, we were able to vary the size of zone of influence, method of characterizing neighborhood, and number and length of time intervals examined. Pre-processing of images, however, was necessary and time consuming. All images needed to be checked for accurate delimitation of individuals and some manual adjustments were necessary. Nevertheless, the relative ease of data capture and analysis and creation of large data sets offer many advantages for studies of plant competition over space and time.

Keywords:

This abstract is being presented at: 10:30 AM in session:
PLANT DEMOGRAPHY