Application of Artificial Intelligence to Planning Sustainable Construction: an Ecological Quality Assessment of Soil

Accurately assessing environmental impact is important for planning sustainable mega-construction projects. Soil ecology plays a significant role in global ecosystems. However, methods for classifying and assessing soil ecological quality have not been well established. To enable such an evaluation, mapping quantitative information on soil ecological quality is required. Thus, the present study was undertaken in order to develop a predictive model for soil ecological quality using an Artificial Intelligence (AI) algorithm and to evaluate the applicability of soil ecological quality maps for planning eco-friendly road construction. Soil microbial diversity as an indicator of soil ecological quality was measured in 196 soil samples in a forest region. A Decision Tree (DT) model was optimized by correlating the field-measured microbial diversity data with existing soil property data (7 soil properties from the Agricultural Soil Information System of Korea). Based on accuracy and coherence analysis, the Ordinary Pairwise Partitioning (OPP) model classifying into 4 groups was selected to predict soil ecological quality. A soil ecological quality map for the forest area was constructed using the OPP model. In order to evaluate the applicability of the soil ecological quality map for planning eco-friendly road construction, optimal eco-road routes were simulated with varying weights assigned to soil ecological quality among various environmental and ecological factors. When the weight was greater than 14%, the soil ecological quality significantly influenced the determination of the route. These results from environmental informatics support the validity of using soil ecological quality information when planning for sustainable construction.