Rats and rice: Belief network models of rodent control in the rice fields of Cambodia

Smith, C., Russell, I. and King, C. (2005). Rats and rice: Belief network models of rodent control in the rice fields of Cambodia. In: Andre Zerger and Robert M. Argent, MODSIM 2005 International Congress on Modelling and Simulation. MODSIM05 International Congress on Modelling and Simulation: Advances and Applications for Management and Decision Making, Melbourne, Australia, (449-455). 12-15 December 2005.


Author Smith, C.
Russell, I.
King, C.
Title of paper Rats and rice: Belief network models of rodent control in the rice fields of Cambodia
Conference name MODSIM05 International Congress on Modelling and Simulation: Advances and Applications for Management and Decision Making
Conference location Melbourne, Australia
Conference dates 12-15 December 2005
Proceedings title MODSIM 2005 International Congress on Modelling and Simulation
Journal name MODSIM05 - International Congress on Modelling and Simulation: Advances and Applications for Management and Decision Making, Proceedings
Place of Publication Canberra, Australia
Publisher CSIRO Sustainable Ecosystems
Publication Year 2005
Sub-type Fully published paper
ISBN 0975840002.
Editor Andre Zerger
Robert M. Argent
Start page 449
End page 455
Total pages 7
Language eng
Formatted Abstract/Summary
Rodents have long been the most significant pests of rice in Southeast Asia. Under traditional, smallholder rice farming systems, rodents generally cause chronic production losses in the order of 5-10% per annum. However, chronic yield losses of 15-30% are not unusual due to increases in annual cropping frequency (ACIAR 2001).

Many farmers have adopted chemical methods of rodent control (rodenticides), such as acute poisons and anticoagulants, to reduce crop damage. However, these carry significant environmental and human health risks. In Cambodia, rodents are a food source for many rural communities and human poisoning can occur if rodents that have ingested poisons are consumed.

In response to this, an Australian Centre for International Agricultural Research (ACIAR) project titled “Farmer-based Adaptive Rodent Management, Extension and Research System in Cambodia” (referred to as the FARMERS project in this paper) has been trialing a physical method of rodent control called the Community Trap Barrier System or TBS (ACIAR 2001). TBS consists of a lure crop and trapping mechanism that attracts and traps rats from surrounding areas before the main rice crop is planted. This helps to reduce the rat population and provides a halo of protection that can potentially protect a large area of crop if groups of smallholder rice farmers cooperate in its implementation.

Over the past four years, Cambodian rice farmers, researchers and extensionists partaking in the FARMERS project have developed a good understanding of the ecological, technical and socio-economic constraints of Community TBS. However, this knowledge is scattered among several people. System modelling can provide a means for capturing and integrating this knowledge and using it for decision support. However, effective participation of all stakeholders in the model development process is essential to ensure that the models are relevant and culturally appropriate by incorporating local and experiential knowledge, and to ensure that model development provides a learning mechanism for local communities.

This paper describes a participatory modelling process used to develop a model describing factors influencing the effectiveness of TBS, and a benefitcost model of various rodent control options available to rice farmers in Cambodia. Rice farmers from the Samrong Commune, and researchers and extensionists from the Cambodian Agricultural Research and Development Institute (CARDI) and the Office of Agricultural Extension (OAE) Kampong Cham, were involved. Participatory research techniques were used to capture their understanding of factors believed to influence TBS effectiveness and of factors considered when selecting a particular rodent control method. Bayesian Belief Networks (BBNs) were then used to model TBS effectiveness and the benefit-cost of alternative rodent control options.

The outcomes suggest that BBNs provide a useful framework for the participatory development of decision support tools. The particular advantages of BBNs are that they facilitate the integration of knowledge and data from diverse sources, can easily combine biophysical, economic and social variables (either quantitative or qualitative) and deal explicitly with uncertainty. They also provide a good tool for communicating management system understanding and behaviour among stakeholders.
Subjects 0701 Agriculture, Land and Farm Management
Keyword Bayesian belief networks
Decision support tools
Participatory modelling
Participatory research
Cambodia
Q-Index Code E1

 
Versions
Version Filter Type
Citation counts: Scopus Citation Count Cited 0 times in Scopus Article
Google Scholar Search Google Scholar
Created: Tue, 02 Mar 2010, 21:32:09 EST by June Temby on behalf of Faculty Of Nat Resources, Agric & Veterinary Sc