Applications of artificial intelligence in forest health surveillance and management

Forests provide vital ecological, economic, and social services, yet their sustainability is increasingly threatened by deforestation, climate variability, and biological stressors such as pests and diseases. Despite the proliferation of Artificial Intelligence (AI) applications for environmental monitoring, there remains a critical knowledge gap in how diverse AI methods spanning machine learning, deep learning, and IoT-based systems can be operationally integrated for real-time and near-real-time forest health surveillance. This study systematically reviews Artificial Intelligence (AI) applications for real-time and near-real-time forest health monitoring, synthesizing evidence from 79 peer-reviewed studies published between 2010 and 2025. Using a PRISMA-guided methodology, data were extracted on study context, data sources, AI algorithms, computational frameworks, and performance indicators. The analysis identifies three dominant AI application domains: (1) disease and pest detection, (2) deforestation and illegal logging surveillance, and (3) wildfire risk prediction and damage assessment. Machine learning (ML) and deep learning (DL) models, including Random Forest, Support Vector Machines, Convolutional Neural Networks (CNNs), and hybrid CNN-LSTM frameworks, demonstrated strong performance, achieving accuracy levels above 90% in most case studies. The review highlights the increasing role of multimodal data fusion from UAVs, LiDAR, IoT sensors, and satellite imagery in improving early detection accuracy. However, challenges persist, including limited generalization across biomes, high data and computational demands, and poor interpretability of complex models. The emerging integration of Explainable AI (XAI) techniques such as SHAP and Grad-CAM provides new pathways for transparency and ecological relevance. The study concludes that advancing multi-hazard, interoperable AI frameworks supported by robust data governance and cross-disciplinary collaboration will be critical for developing adaptive, scalable, and trustworthy forest monitoring systems to enhance global forest resilience and climate mitigation.