Abstract

This paper develops predictive (or correlative) models for the date of first catch of the com earworm, Helicoverpa zea, as a basis for identifying biotic and abiotic factors that influence dispersal and migration. Data described in Goodenough et al. (1988, J. Econ. Entomol.) on the catch of H. zea gathered at over 150 sites predominantly in the central U.S. from 1983 to 1986 are analyzed. The dependent variables, Y1 and Y2, are date of first meaningful catch and date when cumulative catch exceeds 5, respectively; the independent variables are latitude, longitude and elevation of the site. Outstanding among the findings are the following :

1) There is no statistical evidence based on all the data that the slopes of the simple linear regression models of Y2 on latitude differ among the four years. The common slope estimate is 8.11 days/degree, the intercepts differ by as many as 16 days, and the combined model has r2 = 0.69.

2) There is no statistical evidence based on the data in the central U.S. that the partial slopes of the multiple regression models of Y2 on latitude and longitude differ among the four years. The common partial slope estimates are 7.36 and -1.27 days/degree, the intercepts differ by as many as 17 days, and the combined model has R 2= 0.69. Second order terms are not significant .

3) An exploratory analysis using GIS mapping software suggests that elevation is also a significant predictor variable. The suggestion is confirmed in multiple regression models for both Y1 and Y22= 0.71 and 0.72 respectively. The intercepts differ by as many as 20 and 17 days, respectively, over the four years .

These results imply that the time of fi!'st appearance at any location in the central U.S. could be predicted once the date of first appearance in South Texas is ascertained. They also demonstrate the utility of analyzing residuals using GIS mapping software. Research is in progress to investigate other possible predictor variables including soil moisture, soil temperature and precipitation .

Creative Commons License


This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
Apr 23rd, 10:00 AM

PREDICTING THE DATE OF FIRST CATCH OF THE CORN EARWORM, HELICOVERPA ZEA, IN CENTRAL U.S.

This paper develops predictive (or correlative) models for the date of first catch of the com earworm, Helicoverpa zea, as a basis for identifying biotic and abiotic factors that influence dispersal and migration. Data described in Goodenough et al. (1988, J. Econ. Entomol.) on the catch of H. zea gathered at over 150 sites predominantly in the central U.S. from 1983 to 1986 are analyzed. The dependent variables, Y1 and Y2, are date of first meaningful catch and date when cumulative catch exceeds 5, respectively; the independent variables are latitude, longitude and elevation of the site. Outstanding among the findings are the following :

1) There is no statistical evidence based on all the data that the slopes of the simple linear regression models of Y2 on latitude differ among the four years. The common slope estimate is 8.11 days/degree, the intercepts differ by as many as 16 days, and the combined model has r2 = 0.69.

2) There is no statistical evidence based on the data in the central U.S. that the partial slopes of the multiple regression models of Y2 on latitude and longitude differ among the four years. The common partial slope estimates are 7.36 and -1.27 days/degree, the intercepts differ by as many as 17 days, and the combined model has R 2= 0.69. Second order terms are not significant .

3) An exploratory analysis using GIS mapping software suggests that elevation is also a significant predictor variable. The suggestion is confirmed in multiple regression models for both Y1 and Y22= 0.71 and 0.72 respectively. The intercepts differ by as many as 20 and 17 days, respectively, over the four years .

These results imply that the time of fi!'st appearance at any location in the central U.S. could be predicted once the date of first appearance in South Texas is ascertained. They also demonstrate the utility of analyzing residuals using GIS mapping software. Research is in progress to investigate other possible predictor variables including soil moisture, soil temperature and precipitation .