Predictability of the Indian Summer Monsoon onset through an analysis of variations in surface air temperature and relative humidity during the pre-monsoon season

The prediction of the Indian Summer monsoon (ISM) onset is one of the vital questions for the Indian subcontinent, as well as for areas directly or indirectly affected by the ISM. In previous studies, the areas used for ISM-onset prediction were often too large (or too small), or did not include all necessary information for the ISM-onset forecasting. Here, we present recent findings that suggest that a climate network approach may help to provide better definitions for areas used for ISM-onset prediction and an overall better ISM-onset prediction. Our analysis focuses on the following domains: North West Pakistan (NP) and the Eastern Ghats (EG) as they have been identified to include important pre-monsoon information for predicting ISM onset dates. Specifically, we focus on the analysis of surface air temperature and relative humidity in both areas that allows us to derive temporal trends and to estimate the ISM onset. We propose an approach, which allows to determine ISM onset in advance in 67% of all considered years. Our proposed approach is less effective during the anomalous years, which are associated with weak/strong monsoons, e.g. El-Nino, La-Nina or positive Indian Ocean Dipole events. ISM onset is predicted for 23 out of 27 normal monsoon years (85%) during the past 6 decades. In addition, we show that time series analysis in both areas during the pre-monsoon period reveals indicators whether the forthcoming ISM will be normal or weaker/stronger.