A note on procedures used for accuracy assessment in land cover maps derived from AVHRR data

We present results from analyses conducted to evaluate the performance of a dvanced supervised classification algorithms (decision trees and neural net s) applied to AVHRR data to map regional land cover in Central America. Our results indicate that the sampling procedure used to stratify ground data into train and test sub-populations can substantially bias accuracy assessm ent results. In particular, we found spatial autocorrelation in test data t o inflate estimates of classification accuracy by up to 50 points. Results from evaluations performed using independent train and test data suggest th at the feature space provided by AVHRR NDVI data is poorly suited for most land cover mapping problems, with the exception of those involving highly g eneralized classes.