Abstract

Forests and agroforestry systems are increasingly vulnerable to climate-driven stressors such as drought, heatwaves, wildfires, and pest outbreaks, all of which alter ecosystem structure, productivity, and carbon dynamics. Traditional field-based monitoring approaches are spatially limited and insufficient to capture the rapid pace, scale, and complexity of these changes. Recent advances in Artificial Intelligence (AI) and Remote Sensing (RS) now enable high-resolution, multi-temporal monitoring of ecological processes across large and heterogeneous landscapes. This review synthesizes current AI-RS innovations relevant to climate-smart agronomy and forestry. Drawing upon peer-reviewed literature covering multispectral, hyperspectral, LiDAR, SAR, UAV, and IoT sensing, alongside Machine Learning (ML) and Deep Learning (DL) architectures such as CNNs, RNNs, transformers, and anomaly-detection models, we identify eight thematic areas where AI-RS integration is transforming environmental assessment: Structural monitoring, productivity and carbon modelling, resilience evaluation, disturbance forecasting, early-warning systems, multi-sensor fusion, cloud-based platforms, and research challenges. AI-RS fusion enhances forest monitoring by improving biomass estimation, species differentiation, canopy structural mapping, and phenological analysis, while multi-sensor combinations, especially LiDAR-optical and SAR-optical, significantly strengthen detection of degradation, drought signals, fire susceptibility, and regeneration patterns. AI-powered early-warning systems detect drought, fire, and pest risks weeks before observable symptoms, and predictive analytics enable scenario modelling for climate adaptation planning. Persistent challenges include limited ground-truth datasets, reduced model transferability across biomes, computational constraints, interpretability issues in DL models, and socio-ecological barriers to adoption. Nonetheless, AI-enabled RS provides a scalable, data-rich foundation for climate-smart forestry and agroforestry by supporting accurate carbon accounting, early threat detection, and adaptive management. Addressing existing gaps in data, transparency, and governance will be crucial for fully operationalizing these technologies and strengthening resilience in rapidly changing landscapes.

Keywords:Forestry; Climate change; Agroforestry; Artificial intelligence; Remote sensing

Abbreviations: AI: Artificial Intelligence; AI-RS: Artificial Intelligence-Remote Sensing; AGB: Aboveground Biomass; CNN: Convolutional Neural Network; CSF: Climate-Smart Forestry; CHM: Canopy Height Model; DL: Deep Learning; DSS: Decision Support System; eDNA: Environmental DNA; ET: Evapotranspiration; EWS: Early-Warning System; GEDI: Global Ecosystem Dynamics Investigation; GIS: Geographic Information System; GLAD: Global Land Analysis and Discovery; IoT: Internet of Things; LAI: Leaf Area Index; LiDAR: Light Detection and Ranging; LSTM: Long Short-Term Memory (Network); ML: Machine Learning; MRV: Monitoring, Reporting and Verification; MSI: Moisture Stress Index; NDVI: Normalized Difference Vegetation Index; NDWI: Normalized Difference Water Index; NISAR: NASA-ISRO Synthetic Aperture Radar; REDD: Reducing Emissions from Deforestation and Forest Degradation; RNN: Recurrent Neural Network; RS: Remote Sensing; SAR: Synthetic Aperture Radar; SDM: Species Distribution Model; SHAP: SHapley Additive exPlanations; SMAP: Soil Moisture Active Passive; SOC: Soil Organic Carbon; TCN: Temporal Convolutional Network; UAV: Unmanned Aerial Vehicle; VIIRS: Visible Infrared Imaging Radiometer Suite; WUE: Water-Use Efficiency; XAI: Explainable Artificial Intelligence