MODELING OF GEOMETRIC ANISOTROPIC SEMIVARIOGRAM (CASE STUDY: RODENTS ATTACK DATA ON OIL PALM PLANTATION IN CENTRAL SULAWESI)

• Main Author: Tanny, Michael
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/42430
• Institution: Institut Teknologi Bandung

Semivariogram is a variance diagram used to quantify spatial relationships between observations. Anisotropic semivariogram is a semivariogram that depends on both direction and distance. Anisotropic semivariogram, based on range and sill values, can be divided into geometric and zonal anisotropy. Geometry transformation, namely rotation and scaling, is applied to lag distances calculation. These calculations result into new distances to reduce directional semivariograms into one semivariogram with standardizes range of 1. An oil palm company in Morowali, Central Sulawesi conducts monthly monitoring to count the quantity of rodent attacks. This monitoring is carried out on several blocks, with each block having 225-375 sample points. Rodent attacks are recorded with 3 types of attacks based on the period of the attack, namely old attacks, new, and combinations of both. The results of the estimated range and sill values based on the results of the experimental semivariogram calculation: maximum and minimum range of 34.5 and 6, and sill of 9.4. Thus, these data are suitable to be modeled with anisotropic geometry semivariogram. The theoretical models used are spherical, exponential and Gaussian. These models match with data after being validated using test statistics Q1 and Q2 with Q1 vary from 0.0391 to 0.2242 and Q2vary from 0.5474 to 0.6418.