Classification and delineating natural forest canopy density using FCD model (Case study: Shafarud area of Guilan)

Mapping of natural forests in mountainous areas is difficult and expensive. For this propose, satellite remote sensing data is an appropriate solution. In this study, satellite remote sensing ETM+ data was used to develop map of forest canopy cover density, applying FCD model at Shafarud basin of Guilan province of Iran. Forest canopy cover density map was developed, using different density classes, including: 5-25, 25-50, 50-75, 75-100 and control (without forest cover). In the method of forest canopy cover mapping, using FCD model, four indices, including plant vegetation, soil, shadow and temperature were applied on Landsat ETM+ image data, considering an appropriate threshold. Then advanced shadow index and shadow index were uniscaled and calculated. Plant density index and plant canopy cover density map were developed and achieved by percentage, based on FCD model. The accuracy of forest density classification map was estimated, using a map based on 100% Orto-photo-mosaic of aerial figures, drawing a 10 x 10 mm. network and appropriate enlargement of the photos. The developed figure was classified, using method of maximum probability. The overall accuracy and Kappa index were 71 % and 0.61, respectively. The results showed that the developed map, using FCD model and Landsat ETM+ dta, was close to natural land data and facts. Considering the erroe's matris at seven processes of FCD calculation, showed that the model was insufficient for moderate located, thinned and semi-dense canopy covers, whereas it was sufficient and well detected by the model at high located, very thinned canopy covers (<5%) or stands without canopy cover.