Assessment of the future environmental carrying capacity using machine learning algorithms

Abstract

Globally, ecological overshoot has become more prevalent. Enhancing biocapacity has become critical to resolving ecological demand overshoot in sustainable urban development. However, most of the prior research has focused on minimizing environmental carrying capacity (ECC) while ignoring the potential of carbon footprint, population growth, and land cover (LULC). This research assessed the past and projected changes in bio-capacity and bio-productivity dynamics in Khulna City due to land use and land cover (LULC) transformations spanning from 2000 to 2035. Support Vector Machine algorithms were utilized for the classification of LULC, and LULC, bio-capacity and bio-productivity predictions were made using Cellular Automata-Artificial Neural Network models. Results revealed that built-up and cropland area expansion led to an increase in carbon emissions by 43,000 tons/year and bio-productive land by 1100 gha, while a decrease in bio-capacity from 0.09 gha to 0.06 gha occurred during 2000–2020. The prediction shows that by 2021, population growth and urban growth will exceed Khulna City's bio-capacity, and by 2035, bio-capacity (without built-up cover) will decrease to 0.00 gha. The R2 values (–0.67 and –0.91) indicate the strong negative influence of population growth and urbanization on the optimization capacity of soil surfaces. The study demonstrates that Khulna's present urban growth and population growth will result in irreversible ecological collapse, with dire consequences for humans in thenear future.However, the findings facilitate the potential for the decision makers including policymakers, planners, and environmentalists to enhance local land use practices, thereby addressing CO2 emissions and their associated consequences meriting further study.