Harnessing Digital Twin Technology for Detailed Simulation and Analysis of Climate Impact on Agricultural Systems

The forecast of climate variability and unpredictability has become more and more difficult and its implications for agricultural systems and food security are well known to pose serious threats. The robust supply of food production ensured by the tradition farming practices based on the historical data and experience becomes unsustainable as the agricultural sector grew. For the understanding of impacts of climate change on agriculture, dynamic tools are necessary to simulate and better understand the effects. This research addresses the above issue by developing a comprehensive simulation model built upon Digital Twin Technology to predict the impact of different scenarios for the climate on the crop yields. The model incorporates real time data collection and well developed, yet still new, algorithms like random forest to enable real time analysis, as well as realistic prediction using system simulation models to yield highly useful and actionable insights. The study considers a climate change scenario with different degrees of climate variability by comparing the alternatives maximizing yields with damage caused by climate changes. The results show increase of more than 90% in prediction accuracy of yield with proof of operational robustness of the framework for decision making in agriculture. This approach can provide major promise to develop the capability to adjust agricultural practices to a very quick climate change, establishing more sustainable and robust food production systems.