Prakit Tongkum, Chairatna Nilnond, Theera Eksomtramage and Theerapong Juntaraniyom

pp. 717 - 726

Abstract

This study aimed to develop the simulation model for predicting fresh fruit bunch (FFB) yield of oil
palm through multiple linear regression analysis. Two experiments were conducted at the oil palm plantation of Agricultural and Technology College, Krabi province. Six-year-old Tenera hybrid palms were used
for the experiments. These palms were planted in Tha-sae soil series (Typic Paleudults; Fine loamy mixed)
with spacing of 9x9x9 m. In the first experiment, 151 Tenera palms were selected and marked randomly
throughout an area of plantation about 16 ha. For each selected palm, FFB yield and yield component
characters (FFB number and bunch weight) were recorded at every harvesting time for four consecutive
years (June 1993 to May 1997). The results showed that the FFB number and bunch weight could be used to
predict the FFB oil palm yield. In the second experiment, nine plots of Tenera hybrid palms were arranged.
The plot size was 0.48 ha and had twenty palms per plot for data collection for three consecutive years
(January 1994 to December 1996). These data included leaf nutrient (N, P, K, Mg and B) contents in the 17^th frond, the fresh fruit bunch (FFB) yield and the amount of rainfall. The results showed that N, P, K, Mg and
B contents in the leaves, the amount of rainfall and FFB yield in the previous year, together with the N, P, K,
Mg and B contents in the leaves (in the predicting year) could be used to predict the FFB oil palm yield.

Keywords

• oil palm
• Elaeis guineensis
• predicting fresh fruit bunch yield
• multiple linear regression