Resumo

TADEO NOBLE, Alfredo E.; DE LOS SANTOS POSADAS, Héctor M.; ANGELES PEREZ, Gregorio  e  TORRES PEREZ, Jorge A.. Cluster sampling for forest management at Ejido Noh Bec, Quintana Roo. Rev. mex. de cienc. forestales [online]. 2014, vol.5, n.25, pp.64-83. ISSN 2007-1132.

Several sampling estimators are proposed for the inventory data analysis and re-measurement for the Ejido Noh Bec's forests. A total of 302 clusters in 4 strata established was established 2009 in seasonal evergreen tropical forest. The performance of stratified ratio of means (SRM), stratified mean of ratio (SMR) and stratified random sampling (SRS) was compared. A ratio estimator was used to update the inventory by re-measuring 22 clusters in 2012, and from this estimate the annual periodic growth (APG) and annual percentage growth (APG %). The APG and the APG % were 3.721 m3 ha^-1 yr^-1 and 3.2 %, respectively. When the primary sampling units (PSU) are of equal size, SRM and SMR produce similar mean, standard error and accuracy values. When the PSUs are of unequal size, better accuracy is obtained using SRM. We found that the SRS estimators are more accurate than cluster sampling using SRM and SMR. Cluster sampling is not recommended for forest inventories for management purposes when calculating annual allowable cut in small areas since its cost is higher and less accurate compared with the traditional PSUs cluster independent sampling.

Palavras-chave : Annual periodic growth; stratified ratio of means; cluster sampling; stratified mean of ratio; Quintana Roo; primary sampling units.