Watershed condition inventory of the Hoopa Valley Indian Reservation utilizing Landsat digital data

Landsat digital data was analyzed using EDITOR software to map land condition and vegetative cover of the Hoopa Valley Indian Reservation in northern California. Guided clustering was employed to identify the maximum number of spectral classes within the study area. The following vegetation cover categories were mapped: conifer, mixed (conifer, hardwood and brush), hardwood and brush. The brush category was further divided into seven categories of percent crown closure (10% to 80%). Water and snow were also identified and mapped. Evaluation of the classification was completed using binomial approximation within a cluster sampling design. The overall mean probability of correct classification for omission and commission was 0.916 and 0.914, respectively. The major conclusions of this study are (1) Guided clustering techniques can be used to obtained accurate detailed information on vegetative cover in rugged mountainous terrain when analyzed using guided clustering techniques, and (2) successful classification requires reliable and adequate ground control information, an interactive computer system with clustering capabilities and adequate time for re-training to increase classification accuracy.