LAUSR

Present study has assessed different sources of uncertainties in multi-model precipitation projection using Global Climate Models (GCMs) from coupled model inter-comparison project phase five (CMIP5) experiment over seven homogeneous rainfall zones of India namely North Mountainous India (NMI), Northwest India (NWI), North Central India (NCI), Northeast India (NEI), West Peninsular India (WPI), East Peninsular India (EPI), and South Peninsular India (SPI). A relatively new method has been employed to separate out internal variability and climate change signal from precipitation time series before quantifying the uncertainties. In our method, signal has been defined as dynamic trend instead of considering a fixed trend line. Three different types of weighting namely equal weighting, independence based weighting and performance based weighting have been employed to assess the uncertainties of GCM projection over different zones. It has been found that ensemble with performance based weighting has produced smaller inter-model uncertainty but the patterns of temporal evolution of uncertainties have been quite irregular compared to other two ensembles. On the other hand, it has been noticed that bias correction using quantile mapping can effectively reduce the range of uncertainty in a systematic way. It has been observed that inter-model uncertainties over NEI has been relatively lower compared to other zones indicating more robust projection over this zone. A dynamic threshold on signal-to-internal variability ratio (S/I) has been used for estimating time of emergence (TOE) at 95% confidence level over each zone. TOE would be earlier in case of NEI and late in NMI. However no zone may experience TOE in first half of the present century.