Identification of similarity amongst eucalyptus planted areas based on leaf-cutting ant nest sizes

TL;DR

This study uses data analysis techniques to identify similarities among eucalyptus regions based on leaf-cutting ant nest sizes, aiding pest management strategies.

Contribution

It applies Information Bottleneck and PCA methods to cluster regions by ant nest size distribution, offering a novel approach for pest control coordination.

Findings

IB effectively clusters regions with similar ant nest size distributions

PCA reduces data to a three-dimensional space for visualization

Results can inform targeted pest management strategies

Abstract

Techniques for leaf-cutting ant control have been investigated in literature due to the importance of the damage they cause to agriculture. The effectiveness of different forms of control is explored in researches aimed at identifying the balance between pest control efficiency and environmental damage caused by the treatments applied. Plantations with large territorial extensions, which can be contiguous or not, are usually subdivided into local administration that collects data to determine the frequencies of areas size occupied by ant nests. The purpose of this work is to build a relationship of similarities among different geographical regions using the frequency data of nests size occurrence by applying Information Bottleneck (IB) method and Principal Component Analysis (PCA). IB allows simultaneous clustering of each region with the ant nest size distribution, while PCA is used to reduce the variable dimensionalities into a three-dimensional representation of the results. The approach was applied to data of leaf-cutting ants Atta spp. (Hymenoptera: Formicidae) in cultivated Eucalyptus spp forests in Sao Paulo, Brazil. The results suggest information acquired by the method can help coordinate pest management, such as the allocation of baits, material and personnel.