ABSTRACT
Rainfall forecasting can be categorized into two types: short term and long term. Short-term forecasting has higher accuracy due to the low probability of sudden changes in the system. Apart from the agricultural sector, long-term rainfall forecasting also has advantages over other sectors. The Indian Meteorological Department (IMD) uses empirical and dynamic approaches for building long-term forecasting models.
Machine learning approaches are data driven, but until now, they haven’t been incorporated to build long-term rainfall forecasting models by the IMD. In this chapter, Curl Analytics uses machine learning and time series techniques to forecast rainfall one and three months ahead of the current time, providing good accuracy.
