VIEW ARTICLE

Techniques

Analysis of the Spatial Pattern of Plant Pathogens and Diseased Plants Using Geostatistics. D. O. Chellemi, Former graduate research assistant, Department of Plant Pathology, University of Hawaii-CTAHR, Honolulu 96822, Present address: Plant Pathology Department, University of California, Davis 95616; K. G. Rohrbach(2), R. S. Yost(3), and R. M. Sonoda(4). (2)Assistant director, College of Tropical Agriculture and Human Resources, University of Hawaii-CTAHR, Honolulu 96822; (3)Associate professor, Department of Agronomy and Soil Science, University of Hawaii-CTAHR, Honolulu 96822; (4)Professor, University of Florida, Institute of Food and Agricultural Sciences, Agricultural Research and Education Center, Ft. Pierce 33454. Phytopathology 78:221-226. Accepted for publication 28 July 1987. Copyright 1988 The American Phytopathological Society. DOI: 10.1094/Phyto-78-221.

Geostatistical techniques were used to examine the spatial variability among members of a population of plant pathogens and diseased plants in simulated and actual field conditions. Variability was determined by measuring the average of the squared differences between samples over a series of distances and directions with semivariograms. A random pattern with a variance-to-mean (v/m) ratio of 1.01 and a Moran I statistic of 0.019 was used to simulate propagule counts from a hypothetical field and resulted in semivariograms with constant values, indicating propagule counts from each quadrat were independent of neighboring quadrats. An aggregated pattern with a v/m ratio of 2.37 and a Moran I statistic of 0.907 resulted in a linear semivariogram (r² = 0.95) with a slope of 1.22, indicating the variability between propagule counts increased linearly as the distance between quadrats increased. Application to disease incidence data was demonstrated by measuring the variability in patterns of dead pepper seedlings created by amending soil with toxic levels of copper. Semivariograms were constant for random patterns of dead seedlings but showed a nonlinear increase in variability for aggregated patterns of dead seedlings. Two field plots were monitored to determine the variability of initial inoculum of Phytophthora nicotianae var. parasitica and its relation to the incidence of pineapple heart rot. Inoculum had a v/m ratio of 5.22 and 27.26 and a Moran I statistic of 0.044 and 0.309 for fields 1 and 2, respectively. Semivariograms revealed that aggregation of inoculum was not homogeneous but varied over four directions analyzed. However, semivariograms for disease incidence showed that aggregation of diseased plants was homogeneous for the same four directions.