Determining the Optimal Spatial Resolution of Remotely Sensed Data for the Detection of Sirex Noctilio Infestations in Pine Plantations in Kwazulu-Natal, South Africa

Abstract

Sirex noctilio is causing considerable mortality in commercial pine plantations in KwaZulu-Natal, South Africa. The ability to remotely detect variable (for example, low, medium and high) S.noctilio infestation levels remains crucial for monitoring of the actual spread of the disease and for the effective deployment of suppression activities. Although high resolution image data can detect and monitor S.noctilio infestations there are no guidelines to the appropriate spatial resolutions that are suitable for detection and monitoring purposes. This study examines the use of minimum variance to analyze S.noctilio infestations in an effort to determine an optimal spatial resolution of remotely sensed data for forest health monitoring purposes. High resolution (0.5 m) image data was collected using a four band airborne sensor and infestation levels were derived using the normalized difference vegetation index (NDVI) and Gaussian maximum likelihood classifier. It was determined that the appropriate spatial resolution for the detection and monitoring of S.noctilio infestations as estimated by the minimum variance of sub samples narrowly differed based on the level of localized infestations present in the study area. Pixel sizes larger than 2.3 m will not provide adequate information for high infestation levels, while using pixel sizes smaller than the 1.75 m for detecting low to medium infestation levels will yield inappropriate results. The results of this study establish the necessary spatial resolution guidelines needed for the operational detection and monitoring of S.noctilio.