Quantifying Dispersal of Aster Leafhopper

& Spatially Explicit Simulation of

Aster Yellows Epidemics in Vegetable

Production Areas

Casey W. Hoy & Liyang Zhou
Department of Entomology
Ohio Agricultural Research and Development Center

Aster yellows caused up to 100% loss of lettuce crops in Ohio this year. From May 1 to June 20, 15 lettuce fields in Celeryville, Ohio, were sampled twice weekly. Immigration, estimated by the increase in adult leafhoppers/m2 throughout the production area, peaked in early June. Leafhopper population density was consistently higher in lettuce than in other crops, such as escarole, endive, onions, spinach, potatoes.

Flight speed and duration was measured in the laboratory by gluing leafhoppers to flight mills designed for small insects (developed by R. A. J. Taylor). Of 350 leafhoppers attached to flight mills, 37 flew. All flight experiments took place between 15:00 and 20:00 under a light level of 5µE/m2/s. Flight statistics were compared among three treatment groups, the treatments having been applied before the leafhoppers were attached to the flight mill: fluorescent dust marking, rabbit protein marking, and untreated control. No significant differences among treatments were observed in the proportion of leafhoppers flying. Variance in distance flown (number of revolutions on the flight mill), flying speed and flying time was large for all three groups. For example, one leafhopper flew continuously for more than 7 hours, 16,850 revolutions or 7919.5 m at 0.3m/s, whereas most leafhoppers flew tens or hundreds of revolutions. However, results to date indicate that more females flew on the flight mills than males, and females tend to fly longer distances and times. The results of flight mill observation also have demonstrated that aster leafhoppers, particularly females, can fly for long periods of time.

From June to September, four mark-release-recapture (MRR) experiments were conducted. Recapture was by vacuum insect net 24 h after release in 4 m2 areas centered on a regular grid pattern, the spacing of which varied among experiments. Two marking methods, fluorescent dust (Rocket Red, Day-Glo Corp.) and rabbit immunoglobulin detected by ELISA, were compared. No significant differences were observed between the two and fluorescent dust was much easier to detect in recaptured leafhoppers. In each experiment, more females were captured 24 hr after the release than males, but more marked males were recaptured than marked females. More males than females also were captured on sticky window traps, made of 8 x 11 in clear acetate sheets and with 6 such traps mounted on a wood frame from ground level to 3 m. Similar differences in level of activity, particularly interplant movement, by male aster leafhoppers have been observed in previous laboratory and field studies. Aster leafhoppers were captured at 3 m, although most were captured below 3 m. Sticky traps located at the edge of 0.25 acre untreated lettuce plots had more leafhoppers on the side facing away from than on the side facing the plot, suggesting more movement into than out of the plots.

Leafhopper captures in grid patterns 24 h after the releases were analyzed by geostatistics. Exponential and spherical models provided better fit for semivariograms than linear models. Spatial correlation in leafhopper distributions was high, particularly for lettuce crops. Correlation range was greater for releases in fields that contained more lettuce and smaller for a field where different crop types were planted alternatively. Spatial heterogeneity in host crops, therefore, influences leafhopper distribution.

Proportion of dispersing leafhoppers was modeled as a function of distance from a dispersal point with the Gumbel and normal probability density functions for dispersal parallel with and perpendicular to the wind direction, respectively (data at left and below). Kolmogorov goodness-of-fit tests were used to evaluate these probability distribution models, which fit the data well. To estimate probability distributions that describe dispersal without wind, MRR data for recaptures within 50 m of the release point were used. All of the MRR data was used to estimate leafhopper dispersal under the wind speeds measured during our MRR experiments, averaging 9-13 mph.

The product of the Gumbel and normal probability density functions was used to estimate dispersal patterns in all directions from a single point, given distances parallel with and perpendicular to the wind from the point (fig. above). Based on point dispersal functions and field shape (width and length with respect to wind direction) we estimated the proportion of leafhoppers leaving lettuce fields and new probability density functions for the proportion dispersing as a function of distance from the edge of the field. These new functions were needed to simulate dispersal using Arc/Info GIS software. The software can provide distances from grid cells to a given polygon, but only to the nearest cell in the polygon (in our case, distance from grid cells outside a source lettuce field to the nearest cell at the field edge only). Regardless of the shape of the polygon (field), therefore, only the distance to the nearest edge can be calculated by ArcGrid. Both the proportion of leafhoppers leaving a given field and the proportion dispersing as a function of distance from the edge of the field were calculated by integrating point dispersal functions from the field edge toward the center of the field, up to distances where no dispersal was expected. The integrated probabilities of dispersal as a function of distance from the field edge, for all possible starting points within the field, were used to estimate the parameters of new Gumbel and normal probability density functions, for directions parallel with a perpendicular to the wind, respectively, unique to that particular field shape. Gumbel and normal distribution parameters were estimated for four field shapes, so that parameters could be interpolated for other field shapes and various wind speeds.

The calculated dispersal parameters, as a function of wind speed, wind direction, and field shape as described above, is being used in a spatially explicit simulation of aster yellows epidemics. The simulation links an object oriented model of aster leafhopper population dynamics and aster yellows development within fields with a GIS containing spatially referenced information about each of the fields in the model (figure below). Calculations described above for parameters describing the pattern of aster leafhopper dispersal are performed in C++, as are calculations needed to describe crop, leafhopper, and disease development. Dispersal among fields is being simulated with an Arc/Info AML (Arc Macro Language) file that makes use of the spatial analysis tools in Arc/Grid. A global positioning system was used to measure lettuce field positions and place them on a base map of the production area. Aster yellows incidence was estimated in each of these fields at harvest, for comparison with simulation results. This is the most detailed mechanistic simulation of insect dispersal as a function of patch shape and pattern, and most thorough integration of simulation with GIS that we have seen to date in the crop protection arena. The techniques developed should be useful at a range of spatial scales and for various migration and dispersal related simulations.

Simulation results so far support previous empirical and simulation results at much smaller spatial scales. The incidence of aster yellows is reduced when leafhopper dispersal is reduced. When dispersal of aster leafhoppers is high among lettuce fields, predicted aster yellows incidence can be very high. We observe the highest aster yellows incidence in Ohio production areas in lettuce fields that are planted adjacent to relatively mature lettuce fields with aster yellows symptoms. Field shape and spatial pattern (distance between fields) has an effect on aster yellows incidence. Ongoing simulation studies are exploring practical strategies for arranging plantings to reduce aster yellows incidence and more efficient targeting of insecticide applications to control the leafhopper vectors.

Migration and Dispersal Bibliography

Beanland, L., C. W. Hoy, S. A. Miller, and L. R. Nault. 1999. Leafhopper (Homoptera: Cicadellidae) transmission of aster yellows phytoplasma: does gender matter? Environ. Entomol. 28: 1101-1106.

Beanland, L., C. W. Hoy, S. A. Miller, and L. R. Nault. 2000. Influence of aster yellows phytoplasma on the fitness of aster leafhopper (Homoptera: Cicadellidae). Ann. Entomol. Soc. Am. 93: 271-276.

Hoy, C. W., T. T. Vaughn and D. A. East. 2000. Increasing the effectiveness of spring trap crops for Leptinotarsa decemlineata. Entomol. Exp. Appl. 96: 193-204.