Spatial analysis of tree species occurrence: an example of Mount Kenya region

Abstract

Ecologists often seek to predict species distributions of environmental variables. Statistical models are useful for making predictions about occurrence of species based on the variables derived from remote sensing or geographic information system. We used five environmental variables from 265 plots sampled on farms around Mount Kenya region and species inventories colleted for five years (1999-2004) by ICRAF. We modeled the occurrence of the two species: Croton megalocarpus and Cupressus lusitanica using logistic regression model and then abundance was modeled by negative binomial model. The aim of the study was to establish the relationship between the environmental factors and occurrence/ abundance of the tree species on farm and using R functions to map them. "Generally all the variables did poorly in predicting the presence-absence and abundance of the two species. The results suggested that the presence absence for the two species w~spest explained by vegetation class only. For the abundance, the results suggested that Croton megalocarpus abundance was best explained by Vegetation class whereas Cupressus lusitanica abundance was best explained by agro ecological zones. Moreover the findkigs of the study were that occurrence and abundance of the two species were not affected by rainfall, temperature and altitude. ·Sensitivity,.and specificity were affected by the species occurrence and not the abundance.