Current rice models underestimate yield losses from short-term heat stresses

Area of Science:

• Agricultural Science
• Climate Change Impact Assessment
• Crop Modeling

Background:

• Extreme climatic events, particularly heat waves, pose significant threats to rice (Oryza sativa) production.
• Accurate crop modeling is essential for predicting rice yield under future climate change scenarios.

Purpose of the Study:

• To evaluate the accuracy of 14 rice growth models in predicting grain yield under heat stress.
• To identify improvements for crop models to better simulate the effects of short-term extreme heat stress (SEHS) on rice.

Main Methods:

• Phytotron experiments were conducted over four years with four heat treatment levels applied post-flowering.
• Fourteen rice growth models were tested against experimental data.
• Grain-setting rate response to temperature (TRF_GS) functions in eight models were analyzed and adjusted.

Main Results:

• All evaluated models significantly underestimated the negative impact of heat stress on rice grain yield.
• Adjusting the effective periods of TRF_GS improved model performance, particularly for models simulating cumulative daily temperature effects.
• An alternative method using heating-degree days and stage-dependent heat sensitivity parameters reduced prediction uncertainty more effectively than heat dose alone.

Conclusions:

• Existing rice crop models have limitations in predicting yield under variable climatic conditions, especially SEHS.
• There is an urgent need for improved grain-setting functions that account for stage-dependent heat sensitivity.
• Model ensembles did not improve accuracy for predicting grain yield under heat stress.