A knowledge-based prediction model of Verticillium wilt on potato and its use for rational crop rotation

详细信息	查看全文 | 推荐本文 |

• 作者：Y. Cohen^a ; ^{yafitush@volcani.agri.gov.il} ; E. Goldstein^a ; ^b ; A. Hetzroni^a ; I. Lensky^b ; U. Zig^c ; L. Tsror (Lahkim)^d
• 关键词：Multi-criteria analysis ; Pairwise comparison ; GIS ; SDSS ; Verticillium dahliae ; Solanum tuberosum
• 刊名：Computers and Electronics in Agriculture
• 出版年：2012
• 期刊代码：27_01681699
• 类别：abs
• 出版时间：July, 2012
• 卷：85
• 期：Complete
• 页码：112-122
• 文件大小：1611 K

摘要

Verticillium dahliae is the major causal agent of potato early dying (PED) syndrome, characterized by stunting, chlorosis and wilting, premature senescence and early plant dying. It is a common practice to reduce the risk of Verticillium wilt (VW) by applying a rational crop rotation. A knowledge based prediction model for VW was developed and validated. It was based on experimental data and practical management experience, and utilized a knowledge-based approach to acquire the expert knowledge. The potential contribution of this approach was demonstrated in the process of knowledge acquisition and in the model development. The experts identified eight major factors that affect disease development (in descending order of importance): inoculum level in the soil, cultivar susceptibility, fumigation history, frequency of susceptible crops within a crop rotation, growing season, fallow seasons within a crop rotation, infection level in the tubers and soil type. The procedure used for selecting the factors and the method used for extracting their relative weights (pairwise comparison) was proved to be useful as the factor and their relative weights replicated the conclusions of studies from other sites. Additionally, the prediction model was found to be 80%accurate. The prediction model was integrated in a spatial decision support system (SDSS), built under ArcGIS 10, that advises the farmer on plot allocation for potato growing with minimal risk of VW. The SDSS interface and its outputs are presented using real-world data. Effective use of the SDSS by farmers requires the construction of a historical database that includes values of the abovementioned factors and a user-friendly interface for routine updating.