[A-TT29] Machine Learning Techniques in Weather, Climate, Ocean, Hydrology and Disease Predictions

[ATT29-P04] Prediction of surface air maximum temperature over India one week ahead

*Venkata Ratnam Jayanthi¹、Swadhin Behera¹、Masami Nonaka¹、Patrick Martineau¹、Kalpesh Ravindra Patil¹ (1.Application Laboratory, JAMSTEC)

キーワード：heatwave, machine learning

In this study we used various machine learning models to predict the surface air maximum temperature one week ahead over India in the months March to June, the period when India experiences heat waves. The input attributes to the machine learning models are derived using lag correlation analysis between observed surface air maximum temperature anomalies and sea surface temperature as well as with soil moisture anomalies. Several model experiments were carried out by varying a) feature reduction techniques b) pre-processing methods, c) the number of neurons, and d) the activation function and solvers. The results indicate choosing the right combination of pre-processing methods and parameters is essential for predicting extreme temperatures over India using machine learning techniques.