Content uploaded by Savita Patwardhan
Author content
All content in this area was uploaded by Savita Patwardhan on Apr 07, 2015
Content may be subject to copyright.
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
334
*For correspondence. (e-mail: kolli@tropmet.res.in)
High-resolution climate change scenarios for
India for the 21st century
K. Rupa Kumar*, A. K. Sahai, K. Krishna Kumar, S. K. Patwardhan,
P. K. Mishra, J. V. Revadekar, K. Kamala and G. B. Pant
Indian Institute of Tropical Meteorology, Pune 411 008, India
A state-of-art regional climate modelling system, known
as PRECIS (Providing Regional Climates for Impacts
Studies) developed by the Hadley Centre for Climate
Prediction and Research, is applied for India to develop
high-resolution climate change scenarios. The present-
day simulation (1961–1990) with PRECIS is evaluated,
including an examination of the impact of enhanced
resolution and an identification of biases. The RCM is
able to resolve features on finer scales than those re-
solved by the GCM, particularly those related to im-
proved resolution of the topography. The most notable
advantage of using the RCM is a more realistic repre-
sentation of the spatial patterns of summer monsoon
rainfall such as the maximum along the windward side
of the Western Ghats. There are notable quantitative
biases in precipitation over some regions, mainly due
to similar biases in the driving GCM. PRECIS simula-
tions under scenarios of increasing greenhouse gas
concentrations and sulphate aerosols indicate marked
increase in both rainfall and temperature towards the
end of the 21st century. Surface air temperature and
rainfall show similar patterns of projected changes
under A2 and B2 scenarios, but the B2 scenario shows
slightly lower magnitudes of the projected change.
The warming is monotonously widespread over the
country, but there are substantial spatial differences
in the projected rainfall changes. West central India
shows maximum expected increase in rainfall. Extremes
in maximum and minimum temperatures are also ex-
pected to increase into the future, but the night tem-
peratures are increasing faster than the day temperatures.
Extreme precipitation shows substantial increases
over a large area, and particularly over the west coast
of India and west central India.
Keywords: Regional climate model, downscaling, re-
gional climate projections, Indian climate, Indian monsoon.
HUMAN activities since the beginning of the industrial revo-
lution have led to unprecedented changes in the chemical
composition of the earth’s atmosphere. We now have credible
evidence to show that such changes have the potential to
influence earth’s climate
1
, though it is difficult to clearly
delineate the characteristics of climate change associated
with natural and anthropogenic forcings due to complex
interactions within the climate system. Although meteoro-
logical data compiled over the past century suggest that the
earth is warming, there are significant differences at re-
gional levels. Climate variations and change, caused by
external forcings, may be partly predictable, particularly
on the larger (e.g. continental, global) spatial scales. Because
human activities, such as the emission of greenhouse gases
or land-use change, do result in external forcing, it is be-
lieved that the large-scale aspects of human-induced climate
change are also partly predictable. However, the ability to
actually do so is limited because we cannot accurately
predict population change, economic policy, technological
development, and other relevant characteristics of future
human activity. In practice, therefore, one has to rely on care-
fully constructed scenarios of human behaviour and deter-
mine climate projections on the basis of such scenarios.
The Third Assessment Report (TAR) of the IPCC (Inter-
governmental Panel on Climate Change) notes that the
current versions of atmosphere–ocean general circulation
models (AOGCMs) have generally well simulated the
features of the present-day climate at the large and conti-
nental scale
1
. Though the large inter-model differences on
a regional scale with the consequent uncertainties is a cause
for concern, an encouraging sign is that the AOGCMs
have been showing steady improvement over the recent
past. The effects of climate change are expected to be
greatest in the developing world, especially in countries
reliant on primary production as a major source of income.
One of the high priorities for narrowing gaps between cur-
rent knowledge and policymaking needs is the quantita-
tive assessment of the sensitivity, adaptive capacity and
vulnerability to climate change, particularly in terms of
the major agro-economic indicators in the developing
countries. To systematically pursue such assessments, the
most fundamental requirement is the availability of reli-
able estimates of future climatic patterns on the regional
scale, which can be readily used by different impact assess-
ment groups. This needs a systematic validation of the
climate model simulations and development of suitable
regional climate change scenarios, estimations of the as-
sociated uncertainties.
Climate scenarios often make use of climate projections
(descriptions of the modelled response of the climate system
to scenarios of greenhouse gas and aerosol concentrations
or some other hypothetical forcings on the climatic com-
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
335
ponents), by manipulating model outputs and combining
them with observed climate data. Our current level of un-
derstanding of the components of the climate system and
their interactions has reached an advanced stage, with the
availability of a hierarchy of coupled ocean–atmosphere–
sea–ice–land–surface models to provide indicators of global
response as well as possible regional patterns of climate
change. A variety of experiments has been performed by
different modelling groups in the world, to simulate the
expected climate change patterns under different emission
scenarios prepared under IPCC coordination. Prominent
among the scenarios are the IS92a and SRES
2
, for which
extensive model simulated data are made available to the
climate change research community by the IPCC Data
Distribution Centre (
http://www.ipccddc.cru.uea.ac.uk/).
While global atmosphere–ocean coupled models have
provided good representations of the planetary scale fea-
tures, their application to regional studies is limited by
their coarse resolution (~300 km). For example, these
models do not contain realistic topographical features like
the Western Ghats along the west coast of India, and con-
sequently fail to reproduce their predominant influence
on the peninsular monsoon rainfall patterns
3–5
. Developing
high resolution models on a global scale is not only compu-
tationally prohibitively expensive for climate change
simulations, but also suffers from the errors due to inade-
quate representation of high-resolution climate processes
worldwide. It is in this context that regional climate models
(RCMs) provide an opportunity to dynamically down-
scale global model simulations to superimpose the re-
gional detail of specified regions. Developing high-
resolution climate change scenarios helps in: (i) a realis-
tic simulation of the current climate by taking into ac-
count fine-scale features of the terrain etc.; (ii) more
detailed predictions of future climate changes taking into
account local features and responses; (iii) representation
of the smaller islands and their unique features; (iv) bet-
ter simulation and prediction of extreme climatic events;
and (v) generation of detailed regional data to drive other
region-specific models analysing local-scale impacts
6
.
In the present paper, we present the results of regional
climate model simulations for India, based on the second
generation Hadley Centre regional climate model known
as PRECIS (Providing Regional Climates for Impacts
Studies). An evaluation of the model skills and biases is
made by comparing with observed precipitation and tem-
perature patterns with those in the baseline simulation.
Future scenarios of the regional climate towards the end
of the 21st century are also presented.
Data and analysis
The regional climate model PRECIS
PRECIS is an atmospheric and land surface model of lim-
ited area and high resolution, which is locatable over any
part of the globe. Dynamical flow, the atmospheric sulphur
cycle, clouds and precipitation, radiative processes, the
land surface and the deep soil are all formulated, while
the boundary conditions at the limits of the model’s do-
main are required to be specified. PRECIS is forced at its
lateral boundaries by the simulations of a high-resolution
global model (HadAM3H) with a horizontal resolution of
150 km × 150 km, in the so-called ‘time slice’ experi-
ments. HadAM3H is an atmosphere-only GCM which has
been derived from the atmospheric component
7
of
HadCM3, the Hadley Centre’s state-of-the-art coupled
model which has a horizontal resolution of 3.75° longi-
tude by 2.5° latitude. The idea is that a high resolution
atmosphere-only GCM can be used to obtain an improved
regional-level simulation over specific periods of interest
identified from the coupled model integration, as it is
computationally too expensive to run these high resolu-
tion GCMs themselves over century-long integrations. In
the present experiments, two time slices, namely 1961–90
and 2071–2100, have been selected from 240-year long
transient simulations
8
(1860–2100) with HadCM3. Observed
time-dependent fields of SST and sea-ice (HadISST1 data-
set
9
) are used as lower boundary conditions in the control
simulation with HadAM3H. In the climate change ex-
periments, the HadCM3 SST anomaly is added to the ob-
served data to use as the lower boundary forcing. Time-
dependent greenhouse gas and aerosol concentrations are
the same as those in the corresponding HadCM3 time
slice, and initial atmospheric and land surface conditions
are interpolated from HadCM3. HadAM3H has been fa-
voured over HadCM3 for driving the RCMs, since it has a
higher resolution and exhibits an improved control climate,
especially with respect to the positioning of the storm
tracks of the Northern Hemisphere
10,11
. Ensembles of
three baseline simulations for the period 1961–1990,
three simulations for the A2 future scenario (2071–2100)
and one simulation for the B2 future scenario (2071–
2100) have been run with HadAM3H and assessed
12
.
PRECIS has been configured for a domain extending
from about 1.5°N to 38°N and 56°E to 103°E. In the
choice of an RCM domain, it is desirable to select a do-
main that is both large enough so that the regional model
can develop its own internal regional-scale circulations,
but not too large that the climate of the RCM deviates
significantly from the GCM in the centre of the domain.
In the present case, it is presumed that the domain chosen
is adequate to include almost all the regional-scale circu-
lation mechanisms. Using a similar domain, the RCM
was demonstrated to provide a realistic representation of
the intraseasonal variability of the Indian summer mon-
soon, responding to both the global forcing via the lateral
boundary conditions and independent internal dynamics
13
.
The horizontal resolution is 1.24° latitude × 1.88° lon-
gitude in the driving GCM (HadAM3H) and 0.44° ×
0.44° in the RCM (PRECIS). With a nominal resolution
of 50 km versus 150 km, the RCM provides a more real-
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
336
istic representation of orographic features over South
Asia. A more complete description of PRECIS is pro-
vided by Jones et al.
14
. A one-way nesting procedure is
used, with lateral boundary conditions for the RCM being
specified by the GCM. The lateral boundary coupling oc-
curs across a linearly-weighted 4-point buffer zone at
each vertical level.
The RCM is driven at its lateral boundaries by relaxing
surface pressure (p*), the horizontal wind components (u
and v on the 19 model levels) and cloud-conserved tem-
perature and moisture variables (theta and qt on the 19
model levels) towards values interpolated in time from data
saved every 6 hours from the GCM integration. Orographic
heights in the RCM are equal to those in the GCM in the
4-point buffer zone, as well as the 4 rows/columns imme-
diately within it (referred to as the 8-point buffer zone).
Various surface boundary forcing fields are required to
drive the RCM. Prescribed observed SSTs are obtained
by temporal interpolation from monthly mean fields of
the Hadley Centre 1° × 1° resolution HadISST1 dataset
9
.
Future SSTs are obtained by adding the HadCM3 SST
anomaly to the observed data. The land–sea mask and
surface topography are derived from the US Navy 10-min
resolution dataset, and spatially varying vegetation and
soil properties for the land surface scheme are prescribed
from the 1° × 1° climatology of Wilson and Hendersen-
Sellers
15
. Initial conditions for the RCM (including at-
mospheric prognostic variables, soil and canopy moisture
contents, deep soil temperatures and snow amount) are
interpolated from the GCM timestep corresponding to the
start date of the RCM simulation.
Simulations using PRECIS have been performed to
generate the climate for present (1961–1990) and a future
period (2071–2100) for two different socio-economic
scenarios both characterized by regionally focused devel-
opment but with priority to economic issues in one (A2
scenario) and to environmental issues in the other (B2
scenario). Three-member ensembles have been simulated
for the baseline (1961–1990) and one simulation each has
been made for A2 and B2 scenarios towards the end of
the 21st century (2071–2100). The model simulations are
performed with and without including sulphur cycle, to
understand the role of regional patterns of sulphate aero-
sols in climate change. However, the effect of black car-
bon (soot) has not been included in the simulation
experiments. Using the model output from these experi-
ments, high-resolution climate change scenarios have
been developed for various surface and upper air parame-
ters of critical importance to the impact assessments.
Observed data
The basic data for the evaluation of baseline simulations
of PRECIS have been taken from the Climatic Research
Unit (CRU), University of East Anglia global gridded
data set
16
. For additional evaluation of the Indian part of
the domain, the regional mean rainfall and surface tem-
perature series from the homogeneous monthly data sets
developed by the Indian Institute of Tropical Meteorol-
ogy (IITM;
http://www.tropmet.res.in) have been used.
Details of these data sets are discussed by Pant and Rupa
Kumar
17
.
Analysis of extremes
Climate change studies over the past few decades have
mostly focused on the changes in mean values. However,
it is now being increasingly recognized that the manifes-
tations of such changes in the occurrence of extreme
weather and climatic events, particularly on the regional
and local scales, are of paramount importance in assessing
the socio-economic impacts of climate change. Analyses
of extremes requires daily data, primarily on surface tem-
peratures and precipitation. Some global analyses of the
extremes have been made with the available data
18,19
, but
a clear picture of such changes with regional detail is yet
to emerge. A major issue that comes up while assessing
changes in extremes is to objectively define and quantify
the various types of extremes in weather elements. The
joint World Meteorological Organization Commission for
Climatology (CCl)/World Climate Research Programme
(WCRP) project on climate variability and predictability
(CLIVAR) Expert team on Climate Change Detection,
Monitoring and Indices (ETCCDMI) coordinated efforts
to develop a suite of relevant indices and enable their
global analyses through regional participation. In the pre-
sent study, we use some of the indices developed by the
ETCCDMI to evaluate the simulations of PRECIS in rep-
resenting the extremes during the baseline period, and
then examine the future projections of the extremes. The
indices considered in this context are: (i) highest daily
maximum temperature, (ii) lowest daily minimum tem-
perature, (iii) highest 1-day precipitation and (iv) highest
5-day precipitation. The observed indices have been cal-
culated based on daily data during the period 1961–90
over a well-spread network of 40 stations for the
temperatures and 147 stations for precipitation. Identical
parameters for the regional model simulations are also
worked out, both for the baseline period and the future
scenarios.
Global model projections of the climate over India
Assessments of the relative skills of a range of atmosphere–
ocean coupled general circulation models (AOGCMs) are
available concerning their ability to simulate the broad
features of present-day observed surface climatological
features. The large-scale tropical precipitation patterns in
winter (DJF) and summer (JJAS) seasons, as simulated by
several AOGCMs models have been examined earlier
3–5,20
.
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
337
While most models simulate the general migration of
tropical rain belts from the austral summer to the boreal
summer, some of them miss the rainfall maximum in the
tropical Pacific Ocean. Apart from this, in the Indian
monsoon context, the observed maximum rainfall during
the monsoon season along the west coast of India and the
north Bay of Bengal and adjoining northeast India is not
quite realistically simulated in many models excepting
HadCM3 and CSIRO and to some extent in DKRZ. This
may possibly be linked to the coarse resolution of the
models as the heavy rainfall over these regions is gener-
ally in association with the steep orography. However,
the annual cycle in the simulated precipitation averaged
over the South Asian region (land and sea) showed re-
markably similar patterns with the observed, though there
are substantial quantitative biases. The annual surface air
temperature patterns over the South Asian region when
compared with the observed fields also show a general
match of gross features. The models capture the gross
features of the monsoon in terms of low rainfall amount
coupled with high variability over northwest India. How-
ever, some of the finer details of regional significance are
not represented in some of the models; for instance,
ECHAM4 fails to reproduce the rainfall minimum in the
rain shadow region over eastern peninsula, while HadCM2
seems to underestimate the rainfall over the Indo-
Gangetic plains
4
. The simulated monsoon rainfall patterns
in these models differ from the observed patterns in some
respects probably due to the obvious coarse resolution of
the AOGCMs. The horizontal and vertical resolutions of
the atmosphere in the AOGCMs appear to be strongly re-
lated to the skill of the models on regional scale.
Coupled atmosphere–ocean general circulation models
indicate general warming and enhanced rainfall over India
in a greenhouse gas increase scenario, the changes be-
coming particularly conspicuous after the 2040s
3–5,20,21
.
Inclusion of projected increases in sulphate aerosols
damps the sensitivity of the regional climate to green-
house gas increase, but the effect of the greenhouse gases
still dominates the projected changes
3
. While there is
considerable consensus in temperature projections, there
is some disagreement among the models on rainfall
changes
3–5
. In a study with four different GCMs, Douville
et al.
22
found a significant spread in the summer monsoon
precipitation anomalies despite a general weakening of
the monsoon circulation. They concluded that, for decades
to come, the increase in the atmospheric water content
could be more important than the increase in the land–sea
thermal gradient for understanding the evolution of the
monsoon precipitation. They found that the monsoon sen-
sitivity to CO
2
doubling is not only related to changes in
the horizontal transport of water vapour, but also to
changes in the precipitation efficiency, which depends on
soil moisture. Therefore, the treatment of land surface
hydrology in the GCMs is a critical factor in determining
monsoon sensitivity.
Considering all the land-points in India according to
the resolution of each AOGCM, the arithmetic averages
of rainfall and temperature fields are worked out to gen-
erate country-level (all-India) monthly data for the entire
duration of model simulations and for different experi-
ments (Figure 1). GHG simulations with IS92a scenarios
show marked increase in both rainfall and temperature by
the end of 21st century relative to the baseline. There is a
considerable spread among the models in the magnitudes
of both precipitation and temperature projections, but
more conspicuously in the case of summer monsoon rain-
fall. The increase in rainfall from the baseline period
(1961–90) to the end of 21st century ranges between 15
and 40% among the models. In case of mean annual tem-
perature, the increase is much is of the order of 3 to 6ºC.
At a glance one can realize that the change in rainfall in
these two increased greenhouse gas simulations is not as
high as that noted earlier in IS92a scenarios. Compared to
A2 scenario, the B2 simulations show much subdued
trends into the future. The temperature however shows
comparable increasing trends in IS92a and A2 but B2
shows slightly lower trends.
Most models project enhanced precipitation during the
monsoon season, particularly over the northwestern parts
of India. However, the magnitudes of projected change
differ considerably from one model to the other. There is
very little or no change noted in the monsoon rainfall
over a major part of peninsular India. As far as the tem-
perature trends into the future are concerned, all the mod-
els show positive trends indicating widespread warming
into the future. Examination of the spatial patterns of an-
nual temperature changes in the two future time slices for
different models indicates that the warming is more pro-
nounced over the northern parts of India. The different
models/experiments generally indicate the increase of
temperature to be of the order of 2–5°C across the country.
The warming is generally higher in IS92a scenario runs
compared to A2 and B2 simulations. Also, the warming is
more pronounced during winter and post-monsoon months
compared to the rest of the year. Interestingly, this is a
conspicuous feature of the observed temperature trends
from the instrumental data analyses over India
4,5
.
High-resolution GCMs are beginning to provide a more
realistic representation of the extremes in daily precipitation
during the Indian summer monsoon season, allowing the
development of more reliable projections of short-duration
precipitation characteristics such as extremes
23,24
. However,
such projections are available only for a couple of models
for a limited number of scenarios.
Evaluation of PRECIS simulations of Indian climate
The high-resolution regional simulations generated using
PRECIS have been studied in detail to evaluate the model
skills in representing the regional climatological features,
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
338
Figure 1. Present (CTL) and future (IS92a-GHG, SRES-A2 and SRES-B2) simulations of all-India mean sur-
face air temperature and summer monsoon rainfall, based on five global atmosphere–ocean general circulation
models.
Table 1. Characteristics of observed and PRECIS-simulated (baseline, A2 and B2 scenarios without sulphur cycle) seasonal and
annual all-India mean rainfall and mean temperature
Rainfall (mm) Mean temperature (°C)
JF MAM JJAS OND Annual JF MAM JJAS OND Annual
Means
Observed 21 93 839 120 1073 20.3 28.1 27.5 22.8 25.2
Baseline 36 195 939 149 1319 15.6 27.2 24.8 16.6 21.8
A2 49 280 1114 185 1628 20.4 30.9 28.3 21.3 25.9
B2 50 256 1078 171 1554 18.9 29.4 27.5 20.1 24.7
Standard deviations
Observed 11.3 21.0 94.4 27.1 110.0 0.5 0.5 0.3 0.5 0.3
Baseline 14.4 40.5 56.6 31.4 81.4 0.8 0.6 0.4 0.4 0.3
A2 18.0 59.3 87.1 45.0 195.9 0.9 1.1 0.7 0.8 0.7
B2 16.6 49.2 84.9 46.3 119.1 0.8 0.8 0.4 0.6 0.3
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
339
especially the summer monsoon characteristics. The mean
monsoon rainfall for the baseline (1961–1990), simulated
by PRECIS is 939 mm with a standard deviation of
57 mm (Table 1). All India Summer Monsoon Rainfall
(AISMR) based on 306 stations, averaged over the period
of 1871–1990, has been reported to be 852 mm, with a
standard deviation of 84 mm
17
. While the model seems to
have overestimated the all India summer monsoon rain-
fall, it has underestimated the variability. However, the
observed rainfall quantity is considerably linked to the
raingauge network used, and therefore the noted biases
may be partly due to the procedures used in determining
the observed spatially averaged rainfall quantities.
The spatial patterns of seasonal rainfall as simulated by
PRECIS for the baseline period, in comparison with the
observed (gridded data based on the CRU data set) as
well as the driving global models HadCM3 and HadAM3,
are shown in Figure 2. The rainfall maximum over west
coast of India and the rain-shadow region in the south-
eastern peninsula are well simulated by the model. The
seasonal precipitation patterns in the baseline simulation
are quite similar to those observed, indicating that the
baseline simulations provide an adequate representation
of present-day conditions. However, there do exist some
Figure 2. Observed and simulated (baseline) patterns of summer
monsoon precipitation (mm/day) for HadCM3, HadAM3H and
PRECIS. The shaded area indicates rainfall above 12 mm/day in all the
panels.
quantitative biases in the spatial patterns. A conspicuous
bias is the considerably higher than observed monsoon
precipitation over east central India in the baseline simu-
lation. However, this bias also exists in HadAM3, which
indicates that the regional model inherits some of the bi-
ases in the driving global model. The temperature maxi-
mum over northwest India during pre-monsoon season is
well represented in the model (not shown). The monsoon
characteristics like monsoon trough and monsoon circula-
tion are also well represented by the model (not shown).
The mean annual cycles of the simulated all-India
monthly rainfall and surface air temperature are shown in
Figure 3. The model does reproduce the annual cycle of
rainfall reasonably well. However, there appears to be a
significant positive bias in the rainfall during the onset
phase of the monsoon. The model produces excess pre-
cipitation during the two transitional months of May and
June. Further examination of the HadAM3 simulations of
precipitation (not shown) indicates that this bias also ex-
ists in the driving GCM, and therefore was inherited by
the regional model PRECIS. The annual cycle in the sur-
face air temperature, having the highest temperature during
the pre-monsoon months followed by an abrupt fall dur-
ing the monsoon months is well represented by PRECIS.
There appears to be some cold bias in the model through-
out the year, particularly in the seasons other than spring.
However, these biases could be partly due to the network
chosen for the observed temperatures.
RCMs afford us to investigate another important facet
of global climate change in terms of the extreme climatic
events, which are most prominently seen on smaller space
scales. Therefore, it is interesting to assess the ability of
PRECIS in representing the observed patterns of the ex-
tremes in precipitation and temperature over India. For
this purpose, the spatial patterns of highest maximum
Figure 3. Observed and PRECIS-simulated (baseline, 1961–1990)
mean annual cycles of all-India mean precipitation and surface air tem-
perature.
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
340
temperatures and lowest minimum temperatures observed
during the baseline period 1961–1990, have been com-
pared with the simulations of PRECIS (not shown). The
model simulates considerably more intense warm events
over a large part of north India. While part of such biases
could be due to the limited network of stations and damp-
ing due to gridding algorithm used to derive the observed
patterns, there is a reasonable agreement between the
model and the observed patterns in terms of the spatial
location of the extreme maximum temperatures over north
central India. In terms of minimum temperatures also, the
extremes on the colder side are more severe in the model
simulations, but the spatial patterns are quite similar. The
precipitation extremes, in terms of the 1-day and 5-day
highest values, in the simulations are generally less severe
over a large part of the country (not shown). However,
the strong positive bias noticed in the mean seasonal pre-
cipitation over the eastern peninsula is reflected in the ex-
tremes also, in terms of both 1-day and 5-day highest
precipitation.
High-resolution climate change scenarios for India
using PRECIS
The mean annual cycles of all-India mean precipitation
and surface air temperatures for A2 and B2 scenarios are
presented in Figure 4. These indicate a general increase in
precipitation and temperature, for the country as a whole.
A2 and B2 scenarios show similar patterns, with B2
showing slightly reduced magnitudes.
Spatial patterns of rainfall change indicate maximum
increase over west coast and northeast India for both A2
and B2 scenarios (Figure 5). PRECIS estimates 20% rise
in all India summer monsoon rainfall in future scenarios
Figure 4. Mean annual cycles of all-India mean precipitation and sur-
face air temperature for the baseline period (1961–1990) and the future
scenarios (2071–2100) of A2 and B2.
as compared to present. Rise in rainfall is seen over all
states except Punjab, Rajasthan and Tamil Nadu, which
show slight decrease in precipitation in the future scenar-
ios (Figure 6).
PRECIS simulation for 2071–2100 indicates an all-
round warming over Indian subcontinent associated with
increasing greenhouse gas concentrations (Figure 7). The
annual mean surface air temperature rise by the end of the
century ranges from 3 to 5ºC in A2 scenario, whereas the
rise lies between 2.5 and 4ºC in the B2 scenario. The
warming seems to be more pronounced over the northern
parts of India.
Another important aspect of PRECIS simulations is the
role of sulphur cycle. From a general comparison of the
simulations performed with and without sulphur cycle
switched on in the regional model, it appears that there is
no marked change in the simulations either in terms of
rainfall or in terms of surface air temperature (data not
shown). It may be noted here that, as the driving GCM al-
ready has the sulphate aerosols included, the LBCs do
contain the associated large-scale signals. Therefore, the
results suggest that the regional sulphur cycle as consid-
ered by the model has no major impact on the scenarios
derived.
Climate change scenarios over India for 2020s,
2050s and 2080s
As noted earlier, the PRECIS simulations are carried out
for only two time slices, viz. 1961–1990 and 2071–2100,
as the LBCs were available only for these periods. In order
to generate scenario products for periods intermediate be-
tween the two available simulations created for the 1961–
90 (the control) and 2071–2100 (the perturbed) periods, a
simple solution suggested by D. B. Stephenson and C. A.
T. Ferro (pers. commun.
d.b.stephenson@reading.ac.uk)
has been adopted.
It is assumed here that the probability distribution of
the values for a given parameter changes simply due to
anthropogenic trends in location (e.g. mean) and scale
(e.g. standard deviation). In other words, it is assumed
that the shape of the probability distribution stays con-
stant under increased greenhouse forcing. This assump-
tion appears to work moderately well for regional climate
simulations of daily minimum and maximum tempera-
tures over Europe
25
. One of the simplest assumptions to
make about the trends in location and scale is that they
are linearly dependent on the total radiative forcing. In
other words, the location and scale are linear functions of
the logarithm of the equivalent CO
2
concentration as
specified by the IPCC SRES scenarios. This is a more
justifiable assumption than assuming linear-in-time
trends and can be used to obtain intermediate simulations
for different SRES scenarios.
In the present study, the above approach is used to es-
timate the changes in all-India and state wise means of
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
341
Figure 5. Projected changes in summer monsoon precipitation and surface air temperature towards the end of 21st century, for
A2 and B2 scenarios.
the seasonal precipitation and seasonal mean temperature
at 1.5 m for the time slices 2011–2040 and 2041–2070
representing the climatic conditions during 2020s and
2050s respectively. The average CO
2
concentrations for
the different time slices, viz. 1961–1990, 2011–2040,
2041–2070 and 2071–2100 are calculated and the radia-
tive forcing functions based on these average values are
used as interpolating factors to generate the scenarios for
2020s and 2050s. These intermediate scenarios, while be-
ing qualitatively similar to the scenarios for the 2080s,
provide some preliminary quantitative estimates for use
in impact assessments in the near term. The authors may
be contacted for further details on these scenarios.
Future changes in extreme temperatures and
precipitation
Figure 7 shows the A2 scenarios of extremes in tempera-
ture and precipitation, in terms of the highest maximum
temperature, lowest minimum temperature and 1-day/5-
day highest precipitation, towards the end of the 21st cen-
tury. The all-round warming seen in the seasonal mean
temperatures is reflected in the extreme temperatures
also, and both the days and nights are getting warmer in
the future scenario. The projected changes in the ex-
tremes are similar in the case of B2 scenario also, though
with slightly reduced magnitudes (not shown). Interest-
ingly, the night temperatures seem to be increasing at
much higher rates than the day temperatures. While the
lowest minimum temperatures are expected to be warmer
by more than 5°C over most parts of the country, the
highest maximum temperatures show an increase of only
2°C. Even during the recent observational period, there is
mounting evidence that the minimum temperatures are
increasing more rapidly than the maximum temperatures
not only over India
26
, but also across several regions in
the world
27
. In terms of extreme precipitation, there is a
general increase in both 1-day and 5-day extremes. In
particular, there is a marked increase in the severe rainfall
activities over an extensive area covering the Western
Ghats and northwestern peninsular India including Maha-
rashtra and the adjoining parts of Andhra Pradesh, Madhya
Pradesh and Karnataka.
Summary and conclusions
The scenarios presented in this paper include more de-
tailed regional information (50 km × 50 km), and are very
useful for impact assessments in various sectors. This
paper includes only the basic aspects of the simula-
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
342
Figure 6
a. Baseline and future projections (2071–2100) of mean annual cycles of precipitation
for different states of India, as simulated by PRECIS.
a
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
343
Figure 6
b. Baseline and future projections (2071–2100) of mean annual cycles of precipitation
for different states of India, as simulated by PRECIS.
Figure 7. Future projections of extremes in daily temperatures (highest maximum and lowest minimum tem-
peratures; shaded areas indicate warming above 4°C) and precipitation (1-day and 5-day highest precipitation:
shaded areas indicate changes exceeding 40 mm for 1-day precipitation and 60 mm for 5-day precipitation).
b
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
344
tion results; the regional model output contains a large
number of additional parameters that can be obtained from
the authors for use in the impact assessment models.
While the scenarios presented in this study are indicative of
the expected range of rainfall and temperature changes, it
must be noted that the quantitative estimates still have
large uncertainties associated with them. The following
are some of the major conclusions based on the results
presented in this paper:
1. PRECIS shows good skill in depicting the surface
climate over the Indian region, particularly the oro-
graphic patterns of summer monsoon precipitation,
both in terms of mean and extremes.
2. A major bias involving overestimation of rainfall over
the eastern peninsula has been inherited by the re-
gional model from its parent model, indicating the
critical importance of the skills of driving GCMs in
representing the large-scale features.
3. Model simulations under scenarios of increasing
greenhouse gas concentrations and sulphate aerosols
indicate marked increase in both rainfall and tempera-
ture towards the end of the 21st century.
4. Surface air temperature as well as rainfall show simi-
lar patterns of projected changes under A2 and B2
scenarios, but the B2 scenario shows slightly lower
magnitudes of the projected change.
5. The warming is monotonously widespread over the
country, but there are substantial spatial differences in
the projected rainfall changes. West central India
shows maximum expected increase in rainfall.
6. Extremes in maximum and minimum temperatures are
also expected to increase into the future, but the night
temperatures are increasing faster than the day tem-
peratures. Extreme precipitation shows substantial in-
creases over a large area, particularly over the west
coast of India and west central India.
1. Houghton, J. T. et al. (eds), Climate Change 2001: The Scientific
Basis. Contribution of Working Group I to the Third Assessment
Report of the Intergovernmental Panel on Climate Change, Cam-
bridge University Press, 2001, pp. 881.
2. Special Report on Emission Scenarios (SRES), A special report of
Working Group III of the Intergovernmental Panel on Climate
Change, Cambridge University Press, 2000, pp. 599.
3. Rupa Kumar, K. and Ashrit, R. G., Regional aspects of global cli-
mate change simulations: Validation and assessment of climate re-
sponse of the Indian monsoon region to transient increase of
greenhouse gases and sulphate aerosols. Mausam, 2001, 52, 229–
244.
4. Rupa Kumar, K., Krishna Kumar, K., Ashrit, R. G., Patwardhan,
S. K. and Pant, G. B., Climate change in India: Observations and
model projections. In Climate Change and India (eds Shukla, P.
R., Sharma, S. K. and Ramana, P. V.), Tata McGraw-Hill Ltd,
New Delhi, 2002, pp. 24–75.
5. Rupa Kumar, K., Krishna Kumar, K., Prasanna, V., Kamala, K.,
Deshpande, N. R., Patwardhan, S. K. and Pant, G. B., Future cli-
mate scenarios. In Climate Change and India: Vulnerability As-
sessment and Adaptation (eds Shukla, P. R., Subodh K. Sharma,
Ravindranath, N. H., Amit Garg and Sumana Bhattacharya),
Universities Press, Hyderabad, 2003, pp. 69–127.
6. Noguer, M., Richard Jones, David Hassell, Debbie Hudson, Simon
Wilson, Geoff Jenkins and John Mitchell, Workbook on Generat-
ing High Resolution Climate Change Scenarios using PRECIS,
Hadley Centre for Climate Prediction and Research, Met Office,
Bracknell, UK, 2002, pp. 43.
7. Gordon, C. et al., The simulation of SST, sea ice extents and
ocean heat transport in a version of the Hadley Center coupled
model without flux adjustments. Climate Dynamics, 2000, 16,
147–168.
8. Johns, T. C. et al., Anthropogenic climate change for 1860 to 2100
simulated with the HadCM3 model under updated emissions sce-
narios. Hadley Centre Technical Note No. 22, Hadley Centre for
Climate Prediction and Research, UK Meteorological Office,
2001.
9. Rayner, N. A. et al., Global analyses of SST, sea ice and night ma-
rine air temperature since the late nineteenth century. J. Geophys.
Res., 2002, 108, No. D14, 4407 10.1029/2002JD002670.
10. Jones, R. G. et al., Configuring new climate models for Europe
with improved climatologies via better representation of physical
processes. UKMO Second Year Report for MERCURE, EC Con-
tract ENV4.CT97.0485, 1999.
11. Stratton, R. A., Report on aspects of variability in high-resolution
versions of HadAM3. Hadley Centre Technical Note No. 53, Had-
ley Centre for Climate Prediction and Research, U.K. Meteoro-
logical Office, 2004, pp. 32.
12. Hudson, D. A. and Jones, R. G., Regional climate model simula-
tions of present-day and future climates of Southern Africa. Tech-
nical Note No. 39, Hadley Centre for Climate Prediction and
Research, UK, 2002, pp. 42.
13. Bhaskaran, B., Murphy, J. M. and Jones, R. G., Intraseasonal oscilla-
tion in the Indian summer monsoon simulated by global and nested
regional climate models. Mon. Wea. Rev., 1998, 126, 3124–3134.
14. Jones, R. et al., Generating high resolution climate change scenar-
ios using PRECIS, Hadley Centre for Climate Prediction and Re-
search, Met Office Hadley Centre, UK, 2004, pp. 40.
15. Wilson, M. F. and Hendersen-Sellers, A., A global archive of land
cover and soils data for use in general circulation models. J. Cli-
matol., 1985, 5, 119–143.
16. New, M., Lister, D., Hulme, M. and Makin, I., A high-resolution
data set of surface climate over global land areas. Climate Res.,
2002, 21, 1–25.
17. Pant, G. B. and Rupa Kumar, K., Climates of South Asia, John
Wiley & Sons, Chichester, 1997, pp. 320.
18. Groisman, P. Y. et al., Changes in the probability of heavy pre-
cipitation: Important indicators of climatic change. Climatic
Change, 1999, 42, 243–283.
19. Frich, P., Alexander, L. V., Della-Marta, P., Gleason, B., Haylock,
M., Klein Tank, A. M. G. and Peterson, T., Observed coherent
changes in climatic extremes during the second half of the twenti-
eth century. Clim. Res., 2002, 19, 193–212
20. Lal, M. and Harasawa, H., Future climate change scenarios for
Asia as inferred from selected coupled atmosphere–ocean global
climate models. J. Meteorol. Soc. Japan, 2001, 79, 219–227.
21. Lal, M. and Singh, S. K., Global warming and monsoon climate.
Mausam, 2001, 52, 245–262.
22. Douville, H., Royer, J-F., Polcher, J., Cox, P., Gedney, N., Ste-
phenson, D. and Valdes, P-J., Impact of CO
2
doubling on the
Asian summer monsoon: Robust versus model-dependent re-
sponses. J. Meteorol. Soc. Japan, 2000, 8, 421–439.
23. May, W., Simulation of the variability and extremes of daily rainfall
during the Indian summer monsoon for present and future times in
a global time-slice experiment. Clim. Dyn., 2004a, 22, 183–204.
24. May, W., Potential of future changes in the Indian summer mon-
soon due to greenhouse warming: analysis of mechanisms in a
global time-slice experiment. Clim. Dyn., 2004b, 22, 389–414;
SPECIAL SECTION: CLIMATE CHANGE AND INDIA
CURRENT SCIENCE, VOL. 90, NO. 3, 10 FEBRUARY 2006
345
Mooley, D. A. and Parthasarathy, B., Fluctuations of all-India
summer monsoon rainfall during 1871–1978. Climatic Change,
1984, 6, 287–301.
25. Ferro, C. A. T., Hannachi, A. and Stephenson, D. B., Simple non-
parametric techniques for exploring changing probability distribu-
tions of weather. J. Climate, 2005 (in press).
26. Kothawale, D. R. and Rupa Kumar, K., On the recent changes in
surface temperature trends over India. Geophys. Res. Lett., 2005
(in press).
27. Alexander, L. V. et al., Global observed changes in daily climate
extremes of temperature and precipitation. J. Geophys. Rev., 2005
(submitted).
ACKNOWLEDGEMENTS. This work is part of a Joint Indo-UK col-
laborative programme on climate change impacts in India. The authors
are grateful to the Department for Environment, Food and Rural Affairs
(DEFRA), Government of United Kingdom, for sponsoring this project
and the Ministry of Environment and Forests (MoEF), Government of
India, for coordinating its implementation. In particular, the active in-
terest and encouragement of Dr David Warrilow of DEFRA and Dr
Subodh Sharma of MoEF have been of great help. We are also grateful
to the Hadley Centre for Climate Prediction and Research, UK Mete-
orological Office, for making available regional models and their data
products required for this study. Messrs ERM India Private Ltd, the fa-
cilitating agency to manage the project implementation, and their CEO
and Vice-Chairman, Dr T. K. Moulik and Project Consultant, Dr Subrata
Bose, have taken excellent care of the various logistics. Computational/
visualization/documentation work for this project has been done using
the open source software RedHat Linux, Intel FORTRAN, GrADS,
xmGrace, OpenOffice, etc.
