Meteorological observations from 56 stations were linked to provincial yields during 2008-2024 using nonlinear, lagged, elasticity and climate-risk analyses. Simple correlations were generally weak. Nonlinear models suggested approximate temperature turning points of 28.0 C for durian, 27.3 C for mangosteen and 26.6 C for coffee, with humidity turning points around 76-78%. The authors correctly frame these as statistical response indicators, not fixed physiological optima.
Key findings
- Simple crop-climate correlations were mostly weak. • Statistical turning points differed by crop. • Lagged models suggested multi-year climate associations for perennial crops.
Why this matters globally
The results can inform climate-risk planning, but turning points should not guide individual farms without soil, cultivar, irrigation and management data.
Thai researcher contribution
A KMITL researcher integrates climate and yield data across Thailand’s six agroclimatic regions.
Limitations to consider
Provincial yields and station data introduce aggregation bias; observational models are noncausal and thresholds are region-composition sensitive.