An indicator of forest dynamics using a shifting landscape mosaic

摘要: Abstract The composition of a landscape is fundamental indicator in land-cover pattern assessments. objective this paper was to evaluate called ‘landscape mosaic’ as framework for interpreting dynamics over 9-year period 360,000 km2 study area the southern United States. classified land parcel into one 19 possible mosaic classes according proportions natural, developed, and agriculture types surrounding 4.41-ha neighborhood. Using maps from remote sensing, mosaics were calculated each 0.09-ha pixel 1996 2005. Mosaic transition matrices estimated change data then used develop two Markov chain models. A “landscape mosaic” model temporal shifting mosaic, based on probability all pixels. “forest security” same, except that states defined by both pixel, which allowed forest context mosaic. In security model, overall percentage decreased 33% 2005 17% at steady-state, there little relative distribution existing among classes. contrast, steady-state reached later, indicated maximum 10% total available forest. implication depended ultimately contained forest, not within those mosaics.