numerically represent the global climate system, and
simulate historical and future climate projections. The
most recent Fifth Assessment Report of IPCC (AR5;
IPCC 2013) was based on multiple AOGCM outputs
that participated in the fifth phase of the Coupled
Model Intercomparison Project (CMIP5; Taylor et al.
2012) of the World Climate Research Program
(WCRP). The CMIP5 AOGCMs projected distinct
increases in temperature over South Asia during the
twenty-first century, especially during the winter season (Christensen et al. 2013). These AOGCMs with
coarse horizontal resolution (*100 km) were assessed
to have good skill in simulating the regional
synoptic-scale circulation pattern and smoothly varying climate variables like temperature. However, the
temperature biases were assessed to be larger in few
specific regions, particularly at high elevations over
the Himalayas (Flato et al. 2013). Also, the assessment
of a wide range of regional climate processes and
features that are important for capturing the complexity of the Indian summer monsoon rainfall indicated that the performance of the individual AOGCMs
varied in the CMIP5 historical experiments depending
on which aspect of a model simulation was evaluated
(Singh et al. 2017).
The recent developments to generate highresolution regional-scale climate information by
downscaling the CMIP5 AOGCM based global-scale
climate change projections using statistical (i.e.
empirical) and dynamical (i.e. regional climate modelling) methods are used to assess the future changes
in temperature over India in Sect. 2.3.
The statistical downscaling approach derives
empirical relationships linking large-scale atmospheric
variables (predictors) and local/regional climate variables (predictands). These relationships are then
applied to equivalent predictors from AOGCMs.
The NASA Earth Exchange (NEX) Global Daily
Downscaled Projections (GDDP) dataset uses the
Bias-Correction Spatial Disaggregation (BCSD)
method (Thrasher et al. 2012) to correct the systematic
bias of the CMIP5 AOGCM daily maximum and
minimum temperature historical data through comparisons performed against the Global Meteorological
Forcing Dataset (GMFD; Sheffield et al. 2006), and
spatially interpolates the adjusted AOGCM data to the
finer resolution grid of the 0.25° GMFD data.
The BCSD approach used in generating this downscaled dataset inherently assumes that the relative
spatial patterns in temperature observed from 1950
through 2005 will remain constant for future climate
change under the RCP4.5 and RCP8.5 emission
scenarios. The limitation of the NEX-GDDP dataset is
that other than the higher spatial resolution and
bias-correction this dataset does not add information
beyond what is contained in the original CMIP5 scenarios, and preserves the frequency of periods of
anomalously high and low temperature (i.e. extreme
events) within each individual CMIP5 scenario.
The dynamical downscaling derives regional climate information using physical–dynamical relationships by embedding a high-resolution regional climate
model (RCM) within a coarse-resolution AOGCM.
The WCRP regional activity Coordinated Regional
climate Downscaling Experiment (CORDEX; http://
www.cordex.org/) has generated an ensemble of
regional climate change projections for South Asia
with a high spatial resolution (50 km) by dynamically
downscaling several CMIP5 AOGCM outputs using
multiple RCMs. Section 2.3 assess the future changes
in the annual mean, maximum and minimum surface
air temperature over India using the CORDEX South
Asia dynamically downscaled historical simulations
and future projections of climate change till the end of
the twenty-first century available from the CORDEX
data archives on the Earth System Grid Federation
(ESGF). This multi-RCM ensemble consists of six
simulations with IITM-RegCM4 RCM and ten simulations with SMHI-RCA4 RCM, respectively, under
the future RCP4.5 and RCP8.5 emission scenarios,
and five simulations with SMHI-RCA4 RCM under
the future RCP2.6 emission scenario (see more details
in Table 2.6). These dynamically downscaled CMIP5
future temperature projections for India are also
compared in Sect. 2.3 with the NEX-GDDP statistically downscaled daily maximum and minimum temperature projections under the RCP4.5 and RCP8.5
emission scenarios available from the NEX-GDDP
data archives for the 10 CMIP5 host models that were
used to provide lateral and ocean surface boundary
conditions for the CORDEX South Asia RCMs (see
Table 2.6).
The performance of the CORDEX South Asia
multi-RCM historical temperature simulations have
been evaluated in several studies (e.g. Mishra 2015;
Sanjay et al. 2017a, b; Nengker et al. 2018; Hasson
et al. 2018). These dynamically downscaled RCM
simulations showed added value relative to their
driving CMIP5 AOGCMs in simulating the climatological seasonal and annual spatial patterns of surface
air temperature over the South Asia land region, and
the climatological amplitude and phase of the annual
cycle of monthly mean temperature over central India
(Sanjay et al. 2017a). The spatial pattern of
2 Temperature Changes in India
31
simulate historical and future climate projections. The
most recent Fifth Assessment Report of IPCC (AR5;
IPCC 2013) was based on multiple AOGCM outputs
that participated in the fifth phase of the Coupled
Model Intercomparison Project (CMIP5; Taylor et al.
2012) of the World Climate Research Program
(WCRP). The CMIP5 AOGCMs projected distinct
increases in temperature over South Asia during the
twenty-first century, especially during the winter season (Christensen et al. 2013). These AOGCMs with
coarse horizontal resolution (*100 km) were assessed
to have good skill in simulating the regional
synoptic-scale circulation pattern and smoothly varying climate variables like temperature. However, the
temperature biases were assessed to be larger in few
specific regions, particularly at high elevations over
the Himalayas (Flato et al. 2013). Also, the assessment
of a wide range of regional climate processes and
features that are important for capturing the complexity of the Indian summer monsoon rainfall indicated that the performance of the individual AOGCMs
varied in the CMIP5 historical experiments depending
on which aspect of a model simulation was evaluated
(Singh et al. 2017).
The recent developments to generate highresolution regional-scale climate information by
downscaling the CMIP5 AOGCM based global-scale
climate change projections using statistical (i.e.
empirical) and dynamical (i.e. regional climate modelling) methods are used to assess the future changes
in temperature over India in Sect. 2.3.
The statistical downscaling approach derives
empirical relationships linking large-scale atmospheric
variables (predictors) and local/regional climate variables (predictands). These relationships are then
applied to equivalent predictors from AOGCMs.
The NASA Earth Exchange (NEX) Global Daily
Downscaled Projections (GDDP) dataset uses the
Bias-Correction Spatial Disaggregation (BCSD)
method (Thrasher et al. 2012) to correct the systematic
bias of the CMIP5 AOGCM daily maximum and
minimum temperature historical data through comparisons performed against the Global Meteorological
Forcing Dataset (GMFD; Sheffield et al. 2006), and
spatially interpolates the adjusted AOGCM data to the
finer resolution grid of the 0.25° GMFD data.
The BCSD approach used in generating this downscaled dataset inherently assumes that the relative
spatial patterns in temperature observed from 1950
through 2005 will remain constant for future climate
change under the RCP4.5 and RCP8.5 emission
scenarios. The limitation of the NEX-GDDP dataset is
that other than the higher spatial resolution and
bias-correction this dataset does not add information
beyond what is contained in the original CMIP5 scenarios, and preserves the frequency of periods of
anomalously high and low temperature (i.e. extreme
events) within each individual CMIP5 scenario.
The dynamical downscaling derives regional climate information using physical–dynamical relationships by embedding a high-resolution regional climate
model (RCM) within a coarse-resolution AOGCM.
The WCRP regional activity Coordinated Regional
climate Downscaling Experiment (CORDEX; http://
www.cordex.org/) has generated an ensemble of
regional climate change projections for South Asia
with a high spatial resolution (50 km) by dynamically
downscaling several CMIP5 AOGCM outputs using
multiple RCMs. Section 2.3 assess the future changes
in the annual mean, maximum and minimum surface
air temperature over India using the CORDEX South
Asia dynamically downscaled historical simulations
and future projections of climate change till the end of
the twenty-first century available from the CORDEX
data archives on the Earth System Grid Federation
(ESGF). This multi-RCM ensemble consists of six
simulations with IITM-RegCM4 RCM and ten simulations with SMHI-RCA4 RCM, respectively, under
the future RCP4.5 and RCP8.5 emission scenarios,
and five simulations with SMHI-RCA4 RCM under
the future RCP2.6 emission scenario (see more details
in Table 2.6). These dynamically downscaled CMIP5
future temperature projections for India are also
compared in Sect. 2.3 with the NEX-GDDP statistically downscaled daily maximum and minimum temperature projections under the RCP4.5 and RCP8.5
emission scenarios available from the NEX-GDDP
data archives for the 10 CMIP5 host models that were
used to provide lateral and ocean surface boundary
conditions for the CORDEX South Asia RCMs (see
Table 2.6).
The performance of the CORDEX South Asia
multi-RCM historical temperature simulations have
been evaluated in several studies (e.g. Mishra 2015;
Sanjay et al. 2017a, b; Nengker et al. 2018; Hasson
et al. 2018). These dynamically downscaled RCM
simulations showed added value relative to their
driving CMIP5 AOGCMs in simulating the climatological seasonal and annual spatial patterns of surface
air temperature over the South Asia land region, and
the climatological amplitude and phase of the annual
cycle of monthly mean temperature over central India
(Sanjay et al. 2017a). The spatial pattern of
2 Temperature Changes in India
31