Artificial Intelligence-Driven Economic Threshold Prediction for Sustainable Pest Management in Agroecosystems

Abstract

Economic Threshold (ET) and Economic Injury Level (EIL) are fundamental concepts in Integrated Pest Management (IPM), providing scientific guidance for determining when pest control measures become economically justified. However, conventional threshold estimation methods often rely on empirical observations, fixed decision rules, and historical field experiments, limiting their ability to respond to dynamic agroecosystems characterized by changing climatic conditions, pest population variability, and fluctuating economic factors. Recent advances in Artificial Intelligence (AI) have created new opportunities for developing intelligent, adaptive, and data-driven approaches for predicting economic thresholds with greater accuracy and efficiency. This review examines the current state of AI applications in predicting economic thresholds for pest management across diverse agroecosystems. The paper first discusses the evolution of economic threshold-based pest management and reviews existing literature on conventional and AI-assisted approaches. Subsequently, recent developments in machine learning, deep learning, reinforcement learning, explainable artificial intelligence, and hybrid AI models are critically analyzed with respect to their capability to forecast pest populations, assess crop damage, estimate yield losses, and support threshold-based decision-making. The review further highlights practical applications of AI in cereal, oilseed, horticultural, plantation, greenhouse, and precision agriculture systems. Current challenges, including limited availability of standardized datasets, model interpretability, regional variability, and climate-induced uncertainty, are also discussed, together with emerging research opportunities involving multimodal data fusion, edge AI, digital twins, and autonomous decision support systems.