Get reliable rangeland science

The impact of livestock and tourism on vegetation includes a reduction in biodiversity and in some instances, species extinction. To assess these stressor-effect relationships and provide a management tool for protected rangelands of Iran's Lar National Park, we created a multilayer perceptron (MLP) artificial neural network model to forecast vegetation diversity in relation to human activities. Recreation and restricted zones, representing areas with the highest and lowest human impact, were chosen as sampling sites. Vegetation diversity, indicated by the number of species, was recorded in 210 samp le plots. Additionally, twelve landform and soil variables were documented and utilized in developing the model. Sensitivity analyses revealed that the intensity of human activity (in four classes of livestock and tourism population) and soil moisture were the most critical inputs affecting the MLP. The MLP demonstrated strong performance, with R2 values of 0.91 for training, 0.83 for validation, and 0.88 for test datasets. A graphical user interface was created to integrate the MLP model into an environmental decision support system for protected rangelands managers, allowing them to predict impacts and formulate proactive plans to manage human activities affecting vegetation diversity.