Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method

In Malaysia, coconut is one of the commercialized crops, and has become the most important industrial crop. Furthermore, Johor has the largest coconut plantation area in Peninsular Malaysia, which is 14,931.6 hectare. Johor was also the largest producer of coconuts in Malaysia in 2018, with the prod...

Full description

Saved in:

Bibliographic Details
Main Author:	Md. Sabtu, Norraisha
Format:	Thesis
Language:	English
Published:	2021
Subjects:	G Geography (General)
Online Access:	http://eprints.utm.my/id/eprint/100135/1/NorraishaMdSabtuMFABU2021.pdf
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my-utm-ep.100135
record_format	uketd_dc
spelling	my-utm-ep.1001352023-03-29T06:49:44Z Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method 2021 Md. Sabtu, Norraisha G Geography (General) In Malaysia, coconut is one of the commercialized crops, and has become the most important industrial crop. Furthermore, Johor has the largest coconut plantation area in Peninsular Malaysia, which is 14,931.6 hectare. Johor was also the largest producer of coconuts in Malaysia in 2018, with the production number amounting to 112,122 hectares. However, the increasing population of Red Palm Weevil (RPW) in Malaysia is very worrying. RPW is known to be a very dangerous pest on coconut trees, especially found in the Middle East. This study was conducted to model the incident potential of RPW in coconut plantations. The first objective of this study is to examine temporal factors of RPW population. The second objective is to develop a model that can predict the occurrence population of RPW on a coconut plantation based on significant environmental variables. The third objective is to validate the model that has been developed in the second objective. The fourth objective is to produce a susceptibility map of RPW population. This study has been divided into eight phases. This phase includes a site visit to the coconut plantation attacked by RPW, and an interview session conducted with the plant biosecurity officer to address objective 1. A model linear model (GLM) was used to model events based on poisson regression. The model was then validated using root mean square error (RMSE) and mean absolute error (MAE) to address objectives 2 and 3. Finally, the location of the appropriate RPW population infestation was described using weighted Overlay analysis (WOA). Based on the model, five statistically significant variables were determined, which contributed to the total RPW. The variables are humidity, rainfall, wind direction, distance of the trap from the river and road. Based on the value of R squared, the most significant variable was distance from the river (R=5.8, p<0.5), followed by humidity (R=2.9, p <0.5), wind direction (R=1, p<0.5), rain (R=0.2, p<0.5), and distance from the road (R=0.1, p <0.5). This study shows that the places with the lowest probability of being infected by RPW in Mersing are Kg. Semaloi and Tg. Resang, because the two trap locations are far from roads and rivers. Meanwhile, the locations with the highest probability of being attacked by RPW are Kg. Penyabong and certain areas in Kg. Sungai Berbatu, Kg. Belukar Juling, and Kg. Semanyir, because all locations are close to roads and rivers. In conclusion, this model predicts the presence of RPW with errors of 4 (RMSE) and 3 (MAE) for each trap. This model can be developed further, which will assist the authorities in planning and providing an early warning system to identify areas that are likely to be attacked with RPW. 2021 Thesis http://eprints.utm.my/id/eprint/100135/ http://eprints.utm.my/id/eprint/100135/1/NorraishaMdSabtuMFABU2021.pdf application/pdf en public http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:150037 masters Universiti Teknologi Malaysia, Faculty of Built Environment & Surveying Faculty of Built Environment & Surveying
institution	Universiti Teknologi Malaysia
collection	UTM Institutional Repository
language	English
topic	G Geography (General)
spellingShingle	G Geography (General) Md. Sabtu, Norraisha Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method
description	In Malaysia, coconut is one of the commercialized crops, and has become the most important industrial crop. Furthermore, Johor has the largest coconut plantation area in Peninsular Malaysia, which is 14,931.6 hectare. Johor was also the largest producer of coconuts in Malaysia in 2018, with the production number amounting to 112,122 hectares. However, the increasing population of Red Palm Weevil (RPW) in Malaysia is very worrying. RPW is known to be a very dangerous pest on coconut trees, especially found in the Middle East. This study was conducted to model the incident potential of RPW in coconut plantations. The first objective of this study is to examine temporal factors of RPW population. The second objective is to develop a model that can predict the occurrence population of RPW on a coconut plantation based on significant environmental variables. The third objective is to validate the model that has been developed in the second objective. The fourth objective is to produce a susceptibility map of RPW population. This study has been divided into eight phases. This phase includes a site visit to the coconut plantation attacked by RPW, and an interview session conducted with the plant biosecurity officer to address objective 1. A model linear model (GLM) was used to model events based on poisson regression. The model was then validated using root mean square error (RMSE) and mean absolute error (MAE) to address objectives 2 and 3. Finally, the location of the appropriate RPW population infestation was described using weighted Overlay analysis (WOA). Based on the model, five statistically significant variables were determined, which contributed to the total RPW. The variables are humidity, rainfall, wind direction, distance of the trap from the river and road. Based on the value of R squared, the most significant variable was distance from the river (R=5.8, p<0.5), followed by humidity (R=2.9, p <0.5), wind direction (R=1, p<0.5), rain (R=0.2, p<0.5), and distance from the road (R=0.1, p <0.5). This study shows that the places with the lowest probability of being infected by RPW in Mersing are Kg. Semaloi and Tg. Resang, because the two trap locations are far from roads and rivers. Meanwhile, the locations with the highest probability of being attacked by RPW are Kg. Penyabong and certain areas in Kg. Sungai Berbatu, Kg. Belukar Juling, and Kg. Semanyir, because all locations are close to roads and rivers. In conclusion, this model predicts the presence of RPW with errors of 4 (RMSE) and 3 (MAE) for each trap. This model can be developed further, which will assist the authorities in planning and providing an early warning system to identify areas that are likely to be attacked with RPW.
format	Thesis
qualification_level	Master's degree
author	Md. Sabtu, Norraisha
author_facet	Md. Sabtu, Norraisha
author_sort	Md. Sabtu, Norraisha
title	Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method
title_short	Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method
title_full	Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method
title_fullStr	Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method
title_full_unstemmed	Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method
title_sort	predictive model for red palm weevil population in coconut using environmental variables and regression method
granting_institution	Universiti Teknologi Malaysia, Faculty of Built Environment & Surveying
granting_department	Faculty of Built Environment & Surveying
publishDate	2021
url	http://eprints.utm.my/id/eprint/100135/1/NorraishaMdSabtuMFABU2021.pdf
_version_	1776100660298645504

Predictive model for Red Palm Weevil population in coconut using environmental variables and regression method

Similar Items