Use of a geographic information system to explore spatial variation in pathogen virulence and the implications for biological control of locusts and grasshoppers
Article first published online: 29 MAR 2007
2007 The Royal Entomological Society
Agricultural and Forest Entomology
Volume 9, Issue 3, pages 201–208, August 2007
How to Cite
Klass, J. I., Blanford, S. and Thomas, M. B. (2007), Use of a geographic information system to explore spatial variation in pathogen virulence and the implications for biological control of locusts and grasshoppers. Agricultural and Forest Entomology, 9: 201–208. doi: 10.1111/j.1461-9563.2007.00336.x
- Issue published online: 24 APR 2007
- Article first published online: 29 MAR 2007
- Accepted 24 December 2006First published online 29 March 2007
- Body temperature model;
- Dociostaurus maroccanus;
- insect–pathogen interaction;
- Locustana pardalina;
- Metarhizium anisopliae var. acridum;
- Nomadacris septemfasciata;
- Oedaleus senegalensis;
1 In a previous study, we developed a model to predict the effects of temperature on performance of a fungus-based biopesticide for controlling locusts and grasshoppers. Currently, the model is limited to predicting rate of mortality after a spray application at site-specific locations. The aim of the present study is to enhance the utility of this model by linking it with meteorological station data in a geographic information system (GIS) framework to investigate the spatial variation in the performance of the biopesticide.
2 The model provides maps that define spatial variation in pathogen virulence (measured as LT90 for a treated population) across different regions. The model was used to explore the variation in biopesticide performance against four economically important pest species: Moroccan locust Dociostaurus maroccanus in Spain; brown locust Locustana pardalina in South Africa; red locust Nomadacris septemfasciata in Zambia and; Senegalese grasshopper Oedaleus senegalensis in Niger.
3 Model outputs for the different species were partially validated against data from field trials. The models provided good estimates of time to 90% mortality for five out of six independent comparisons. There was also good agreement between the spatial model and equivalent output from the site-specific model.
4 Simulations of virulence against N. septemfasciata in Zambia indicated very uniform, rapid mortality with LT90 throughout the country generally less than 11 days. Pathogen-induced mortality of O. senegalensis in Niger was predicted to be slightly slower and more variable with mortality fastest in the southern regions (< 15 days) and slowing to the north of the country (16–20 days). For both L. pardalina in South Africa and D. maroccanus in Spain, the model revealed highly variable patterns of mortality with LT90 ranging from < 15 days in some areas to > 30 days in others.
5 The implications of these different patterns of variability for the development of optimum use strategies for the various species and the basic understanding of the ecology and evolution of insect–pathogen interactions are discussed.