Road impact assessment modelling on plants diversity in national parks using regression analysis in comparison with artificial intelligence

Increasing urban demand and population growth in cities have led to an increase in demand for developing new ways. Parchin-Pasdaran Road, which runs from the heart of Khojir National Park, is a big threat to this park. Despite these environmental threats, the development and creation of new highways is unavoidable. This research was carried out to study the effect of the road on Smith-Wilson evenness index and Simpson diversity index in Khojir National Park. The Land Management Units were created using the ArcGIS software. Using appropriate algorithm in artificial neural network structure and linear regression of species evenness and diversity was modelled. For modelling of species evenness and diversity, factors like bulk density, particle density, moisture content, porosity and distance from the road were used. Finally, considering that the amount of R-2 in artificial neural network method was statistically significant for Smith-Wilson and Simpson (0.54), (0.71) and in the regression method, respectively (0.25), (0.75), was obtained, the neural network model was selected as the optimal model. Based on the analysis of sensitivity analysis, humidity factors at 5 and 10 cm from the soil surface, the actual 5 cm particle density on the Smith-Wilson index and the porosity at 10 cm from the soil surface had the most effect on the Russian Simpson index.