Examining flower thrips distribution in southern highbush blueberries utilizing geostatistical methods to improve monitoring techniques

During 2007, Florida produced 3.54 million kg of fresh market blueberries at an average of $2.27 per kg. Despite this high price, blueberry production is threatened by high populations of thrips during flowering. White sticky traps are used by many growers and researchers to monitor thrips populations. To develop an effective sampling program, traps must be spaced so that samples are spatially independent from each other. Using geostatistics, semivariograms can be constructed and fit with an appropriate model. The range, the distance at which spatial independence is reached, can then be determined. Kriging uses the variogram models to interpolate values at unsampled locations. The purposes of this study were to determine the optimum trap spacing to monitor flower thrips in southern highbush blueberries and to examine their distribution using kriging. One hundred white sticky traps were distributed throughout a regular grid, with an additional 30 traps placed randomly throughout the gridded area. Traps were changed out weekly over a period of 3 weeks and the number of thrips per trap was recorded. TerraSeer STIS was used to construct and analyze semivariograms and to perform ordinary kriging. Only the variogram from the second week of the study had both a low nugget value (0.14) and a low MSS error (0.037). The range of this variogram was 11 m. Ordinary kriging showed two major hot spots, one in the southwest corner of the study area and one in the center towards the eastern edge of the study area.