Forest landscape aesthetic quality model (FLAQM): A comparative study on landscape modelling using regression analysis and artificial neural networks

Today, the landscape aesthetic quality assessment is more technical and quantitative in environmental management. We aimed at developing artificial neural network (ANN) modelling and multiple regression (MLR) analysis approaches to predict the perceptional aesthetic quality of forest landscapes. The methodology, followed in this paper, can be divided into six distinct parts: (i) selection of representative study sites, (ii) mapping of landscape units, (iii) quantification of naturalness indicators, (iv) visibility analysis, (v) assessment of human perceptions, (vi) ANN and MLR modelling and sensitivity analysis. The results of ANN modelling, especially its high accuracy (R2 = 0.871) in comparison with MLR results (R2 = 0.782), introduced the forest landscape aesthetic quality model (FLAQM) as a comparative model for an assessment of forest landscape aesthetic quality. According to sensitivity analysis, the values of livestock density, tree harvesting, virgin forest, animal grazing, and tree richness were identified as the most significant variables which influence FLAQM. FLAQM can be used to compare the classes of aesthetic quality of forests.