Assessing Weather-Yield Relationships in Rice at Local Scale Using Data Mining Approaches

Area of Science:

• Agricultural Science
• Data Science
• Climate Science

Background:

• Climate variability poses significant challenges for farmers, impacting crop yields.
• Farmers are increasingly collecting observational data on crop performance.
• New tools are needed to help agriculture adapt to changing environmental conditions.

Purpose of the Study:

• To apply data mining techniques to identify best practices for adapting to climate variability.
• To analyze the relationship between climate factors and crop yield variability using observational data.
• To provide site-specific recommendations for cultivar selection and management.

Main Methods:

• Combined preexisting observational datasets of commercial harvest records with in situ daily weather series.
• Utilized Conditional Inference Forest and clustering techniques to analyze data.
• Assessed relationships between climatic factors and crop yield variability at local scales for specific cultivars and growth stages.

Main Results:

• Climatic factors explained 6% to 46% of spatiotemporal yield variability.
• Crop responses to weather were non-linear and cultivar-specific.
• Identified different weather patterns for irrigated and rainfed systems, with varying yield levels.

Conclusions:

• Data mining can leverage embedded knowledge from observational data for agricultural adaptation.
• Site-specific information on cultivar response to climate factors can be generated.
• The study supports on-farm management decisions for adapting to climate variability.