Search

Coordinated Regional Climate Downscaling Experiment (CORDEX) has developed high-resolution, dynamically downscaled data across the world. This paper investigates the spatio-temporal correspondence between CORDEX data for South Asia (CORDEX-SA) and observed precipitation data to assess its reliability. The entire Indian mainland is considered as study region. Both monthly and seasonal analyses are undertaken along with the extreme magnitudes during historical (1971–2005) and future (2006–2100) periods. The outputs of two regional climate models (RCMs), viz. regional climate model, version-4.4 (RegCM4) and Rossby Centre regional Atmospheric model, version-4 (RCA4), participating in CORDEX-SA program, are considered for medium (RCP4.5) and high (RCP8.5) emission scenarios. The CORDEX-SA is found to well represents in central India, but significant bias in the mean precipitation (± 200 mm) is noticed at northern, north-eastern, southern and coastal regions. The inferior performance of both the RCMs is noticed over high rainfall regions. Specifically, more than 70% area is found to have mean bias more than ± 100 mm. Performance of RCA4 is better in North and Himalayan regions, where 67% of area is found to be within the aforementioned threshold, which is 23.64% in case of RegCM4. Performance of RegCM4 is marginally better in the peninsular India as compared with RCA4. In case of extreme magnitudes, the insignificant correspondance between CORDEX-SA data and observed precipitation is noticed over the many parts of India. Results of this study are important to check the applicability of CORDEX data in various impact assessment studies. [ABSTRACT FROM AUTHOR]

It is important to estimate a high-resolution spatiotemporal distribution of temperature rise under climate change for effective adaptation and mitigation planning. This paper studies the threshold crossing time (TCT) for eight Indian warming targets (IWTs) between 1.5 and 5 °C based on regional climate models (RCMs) participating in The Coordinated Regional Downscaling Experiment (CORDEX)—South Asia project. To evaluate the performance of CORDEX experiments in simulating the near-surface mean temperature, the model performance index (MPI) is used which represents that the CanESM experiment performs best for simulating Tmean over India. All India average Tmean projected by the multi-model ensemble (MME) represents that 1.5 °C TCT is about 2033 ± 3 and 2028 ± 3 in RCP4.5 and RCP8.5 scenarios, respectively. During the twenty-first century, Indian subcontinent can experience five IWTs (1.5–3.5 °C) under RCP4.5 and all eight under scenario RCP8.5. The fastest warming region in India is the western Himalayan region (North India) and north-west region (Indian Thar Desert). [ABSTRACT FROM AUTHOR]

Given the general notion that dynamical downscaling leads to added accuracy in both historical simulations as well as climate change projections, this paper investigates its validity over India using historical data (1975-2005) from the CORDEX models and their driving global climate models (GCMs) from Coupled Model Intercomparison Project Phase 5 (CMIP5), and comparing them against observed temperature and rainfall. We find that downscaling invariably leads to an improvement in the spatial pattern of surface air temperature, but compared to the driving GCMs, the errors in magnitude after downscaling are even worse in some cases. In regard to JJAS rainfall simulations, the CMIP5 driving GCMs are found to be superior to their dynamically downscaled counterparts both in terms of spatial patterns as well as magnitude of errors. Both CMIP5 driving GCMs as well as the CORDEX models underestimate rainfall during JJAS; however, negative bias in CORDEX models is worse. Unlike the driving CMIP5 GCMs, their dynamically downscaled counterparts simulate an early onset followed by a slow and late withdrawal of the Indian summer monsoon rainfall. The frequency of occurrence of rainfall intensities is simulated well by both sets of models in the lower intensity regime (0-20 mm/day); however, for higher intensities, the driving CMIP5 GCMs underestimate whereas the CORDEX models overestimate. [ABSTRACT FROM AUTHOR]