Articles
DEVELOPING A PREDICTION MODEL FOR SEASONAL HARVEST VOLUMES FOR THE PEACH PRESERVE INDUSTRY IN SOUTH AFRICA
Article number
1068_11
Pages
99 – 104
Language
English
Abstract
The local fruit industry depends heavily on exportation of products.
Profit margins are small, even in good seasons.
This is largely due to the costs of contract factory labor and supplies, like tins and sugar, which need to be in stock at the beginning of the season.
If the crops are down from the estimated, and orders placed based on these data, supplies would need to be carried until the next season, and labor would still need to be paid, regardless of the crop volume, which impacts on the profit margin.
Thus, the contribution of an accurate crop estimate, early in the season, for planning in the factory is evident.
At present, technical advisors provide the crops estimates.
As the accuracy depends heavily on the experience of these advisors, a need arose to supplement their expertise with a prediction model, to ensure that this prediction will still be available and accurate in future, should the expertise become unavailable.
A model was developed using historical crop volumes and various climatic parameters, in a multi linear regression.
Then, it was validated against the following seasons crop, as well as the crop estimates of the technical advisors.
The accuracy of the prediction model was comparative to the technical estimate, and more accurate under specific climatic conditions that occurred irregularly.
The model was applied commercially, during 2010/11.
Profit margins are small, even in good seasons.
This is largely due to the costs of contract factory labor and supplies, like tins and sugar, which need to be in stock at the beginning of the season.
If the crops are down from the estimated, and orders placed based on these data, supplies would need to be carried until the next season, and labor would still need to be paid, regardless of the crop volume, which impacts on the profit margin.
Thus, the contribution of an accurate crop estimate, early in the season, for planning in the factory is evident.
At present, technical advisors provide the crops estimates.
As the accuracy depends heavily on the experience of these advisors, a need arose to supplement their expertise with a prediction model, to ensure that this prediction will still be available and accurate in future, should the expertise become unavailable.
A model was developed using historical crop volumes and various climatic parameters, in a multi linear regression.
Then, it was validated against the following seasons crop, as well as the crop estimates of the technical advisors.
The accuracy of the prediction model was comparative to the technical estimate, and more accurate under specific climatic conditions that occurred irregularly.
The model was applied commercially, during 2010/11.
Authors
E. Lötze, O. Bergh
Keywords
chill units, climate, crop yield, heat units, weather
Online Articles (34)