Using high-resolution satellite imagery and double sampling as a cost-effective means of collecting forest inventory data – the case of Hans Kanyinga Community Forest, Namibia

Abstract

The present study deals with the use of high-resolution satellite imagery in forest inventory of the open savanna woodlands of southern Africa. The study was carried out in Hans Kanyinga Community Forest, north-eastern Namibia. A two-phase (double) sampling design was used to estimate the variables of interest, namely stand volume, stand density and diameter distribution. QuickBird satellite images were used to extract auxiliary variables (image data), such as photogrammetric crown diameter and number of stems, using visual interpretation and measuring tools offered by Erdas 8.7 geographic imaging software. Field inventory data (terrestric data) collected in 2002 were used to obtain the terrestric data. Stepwise regression and correlation analysis between satellite imagery and field data was carried out and the results were moderate. The stand volume model explained 56%, the stand density model 81% and the diameter distribution model 43% of the variation. Cost-efficiency assessment of double sampling with regression estimators showed a considerable reduction in inventory costs by up to 24% compared to a traditional forest inventory method in Hans Kanyinga Community Forest. Double sampling was found to be cost-efficient when the proportion of variability of the models was higher than 27%. The inventory concept showed potential in providing information required both at the higher and lower levels of forest management. Follow up research is required.