Climate-Conditioned Artificial Intelligence for Crop Pest and Disease Prevention: From Recognition Benchmarks to Actionable Early Warning
Asia Pacific Journal of Management and Sustainable Development · 2026-09 · 14 (02) · 84-93
DOI: 10.70979/dvko4747 ↗Kurz gefasst
Fachartikel in der Zeitschrift Asia Pacific Journal of Management and Sustainable Development, 2026-09. Verfasst von Weidong Li und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Climate change is shifting pest phenology, pathogen infection windows, geographic suitability and the timing of crop-protection interventions. At the same time, computer vision, remote sensing and environmental sequence models are widening the reach of crop-health surveillance. The two literatures remain poorly connected. Climate-impact studies seldom test operational AI pipelines, whereas detection models are usually evaluated on static datasets with random train–test splits. This structured critical review examines peer-reviewed research published mainly between January 2021 and August 2026, retaining earlier studies only when they define a benchmark or mechanism used by recent work. We classify the evidence as direct climate-conditioned AI, enabling perception and sensing, or contextual ecological projection. Across these tiers, non-stationarity emerges as the central technical problem: climate change alters species ranges and class priors, modifies spectral and thermal signatures, shifts monitoring windows and can invalidate models trained on historical distributions. Meteorological time series, environmental deep learning, mechanistic–data hybrids and climate-scenario distribution models already support earlier or longer-range risk assessment. Cross-region tests, calibrated probabilities and measurements of decision impact remain scarce. We propose an architecture that links climate forcing, biological hazard, observation, forecasting, decision and outcome feedback, together with an evaluation ladder spanning lead time, spatial transfer, calibration, avoided loss, pesticide use and net benefit. Reliable prevention will depend less on marginal gains in recognition accuracy than on testing the whole system under changing biological and operational conditions.
Zitieren
APA
Li, W., & Ben, G. (2026). Climate-Conditioned Artificial Intelligence for Crop Pest and Disease Prevention: From Recognition Benchmarks to Actionable Early Warning. Asia Pacific Journal of Management and Sustainable Development, 14(02), 84-93. https://doi.org/10.70979/dvko4747
Vancouver
Li W, Ben G. Climate-Conditioned Artificial Intelligence for Crop Pest and Disease Prevention: From Recognition Benchmarks to Actionable Early Warning. Asia Pacific Journal of Management and Sustainable Development. 2026;14(02):84-93. doi:10.70979/dvko4747
BibTeX
@article{Li2026,
author = {Li, Weidong and Ben, Guozhi},
title = {Climate-Conditioned Artificial Intelligence for Crop Pest and Disease Prevention: From Recognition Benchmarks to Actionable Early Warning},
journal = {Asia Pacific Journal of Management and Sustainable Development},
year = {2026},
volume = {14},
number = {02},
pages = {84--93},
doi = {10.70979/dvko4747}
}
EUVIMED