A REGRESSION-MODEL FOR THE SPATIAL-DISTRIBUTION OF RED-CROWN CRANE INYANCHENG BIOSPHERE RESERVE, CHINA

A logistic regression of principal components, in combination with spa tial autocorrelation analysis, was developed to simulate and predict t he distribution of species. The model was tested by the matching coeff icient (m(1) + m(2))/n, which was obtained by comparing the observed a nd the predicted presence/absence data of species. The results from th e case study for red-crown crane (Grus japonensis) in Yancheng Biosphe re Reserve, Eastern China, showed that the matching coefficient of the model was as high as 91.13%. (C) 1997 Elsevier Science B.V.