Previous View
 
APSnet Home
 
Phytopathology Home


VIEW ARTICLE

Ecology and Epidemiology

Spatial and Spatiotemporal Autocorrelation Analysis of Citrus Canker Epidemics in Citrus Nurseries and Groves in Argentina. T. R. Gottwald, Research plant pathologist, USDA, Agricultural Research Service, Horticultural Research Laboratory, Orlando, FL; K. M. Reynolds(2), C. L. Campbell(3), and L. W. Timmer(4). (2)Forest pathologist, USDA, Forest Service, Institute of Northern Forestry, Anchorage, AK; (3)Associate professor, North Carolina State University, Raleigh; (4)Professor, University of Florida, Citrus Education and Research Center, Lake Alfred. Phytopathology 82:843-851. Accepted for publication 16 April 1992. This article is in the public domain and not copyrightable. It may be freely reprinted with customary crediting of the source. The American Phytopathological Society, 1992. DOI: 10.1094/Phyto-82-843.

Spatial and spatiotemporal (ST) patterns of citrus canker were examined in three nurseries and two groves in Argentina. The center plant in each plot was inoculated with Xanthomonas campestris pv. citri, and disease was allowed to progress for two growing seasons. Disease assessments were made at about 21-day intervals. Final disease incidence was >90% in all three nurseries and reached 69 and 89% for orange (Citrus sinensis) and grapefruit (C. paradisi) groves, respectively. For nursery plots, each quadrat was represented by disease counts, i.e., the number of diseased leaves, in a six-plant row segment. For grove plots, each individual tree was considered a quadrat because of the large number of leaves per tree. Data from each assessment date were analyzed by spatial correlation analysis and by ST autocorrelation analysis. Changes in significantly correlated spatial lags closely followed the changes in the disease progress curves for each plot. Proximity patterns in all three nurseries changed little during the first three to four assessments and then became more complex, often with noncontiguous elements that indicated the formation of secondary foci. Noncontiguous elements remained until the last few assessments, when they eroded and the proximity patterns generally became larger and contiguous as the numerous foci coalesced. Disease incidence increased more rapidly in the grove plots than in the nursery plots. Spatial proximity patterns of disease for the grapefruit grove plot, corresponding to assessment dates immediately after a rainstorm with high winds, were elongated in the north-south direction. In contrast, spatial proximity patterns in the orange grove plot were more radially symmetrical until later in the epidemic, when they became more elongate in the north-south orientation and a few noncontiguous elements developed. ST autocorrelations and partial autocorrelations from the ST autocorrelation analysis of nurseries and groves were generally highest with a square proximity pattern. For citrus nurseries, ST autocorrelations and partial autocorrelations were consistent over time. ST autocorrelations decayed rapidly over spatial lags, but remained significant to four temporal lags. Therefore, the ST transfer function for citrus nurseries infected with citrus canker was represented by a ST autoregressive integrated moving-average (STARIMA) model, STARIMA(0,4,1,1). The ST partial autocorrelations were similar for both grove plots, indicating a similarity in the autoregressive components of each grove and, thus, a STARIMA model structure, but the two groves differed in inclusion of moving-average terms. For the orange grove, autocorrelations for the first temporal lag decayed slowly over the first three spatial lags, whereas the autocorrelation for the first temporal lag in the grapefruit grove decayed rapidly over spatial lags. Also, significant moving-average effects were estimated to extend to two temporal lags in the grapefruit grove data but to only one in the orange grove data. Thus, STARIMA model forms for the orange and grapefruit groves were estimated to be STARIMA(0,1,4,1) and STARIMA(0,2,1,2), respectively.