12/20/2013
ASA- Special Session on AgMIP – 4th November-2013 at Tampa, Florida
Integrated Climate Change Assessment
through Linking Crop Simulation with
Economic Modeling – Preliminary results from
Indo-Gangetic Basin
Nataraja Subash, Babooji Gangwar, Harbir Singh Guillermo Baigorria,
Anup Das, Andy McDonald, Balwinder Singh, Abeed Hussain
Chaudhary, Rajendra Dorai & Dileepkumar Guntuku
1
12/20/2013
Climate change and IGB
 Climate change impacts are increasingly visible in IGB with
greater variability of the monsoon
 An increase in the occurrence of extreme weather events
such as heat waves and intense precipitation that affect
agricultural production drastically and thereby the food
security and livelihoods of many small and marginal
farmers, particularly in the more stress-prone regions of
the central and eastern IGB
 Earlier estimates indicated that if the current trends
continue until 2050, the yields of irrigated crops in IGB are
projected to decrease significantly – maize by 17 %, wheat
by 12 % and rice by 10 % - as a result of climate change
induced water stress (Aggarwal et al. 2009)
ASA- Special Session AgMIP – 4th Nov-2013
2
12/20/2013
Indo-Gangetic Basin – Food
Basket of South Asia
Low productivity (Rice-Wheat 4-5 t/ha)
Poor investment in infrastructure
Medium-high precipitation (1000-2000 to > 2000 mm)
High potential for cold water fisheries and livestock
Degradation of Land and water resources
Low human capital - high out-migration
Downstream environmental constraints
ASA- Special Session AgMIP – 4th Nov-2013
3
12/20/2013
Low Productivity (R-W:4-8 t/ha) - Food
deficit region
Low investment in infrastructure
Medium - High rainfall (1000-2000 to >
2000 mm)
Underutilization of ground water (< 20
%)
Very few developed irrigation network
High risk of flooding, poor drainage
and moderate drought
Out-migration of laborers
High Productivity (R-W: 8-12 t/ha) - Food surplus region
High investment in infrastructure
Higher inputs of agro-chemicals
Low - Medium rainfall (500-1000 to 1000-2000 mm)
Over exploitation of ground water (>80 %)
Well developed irrigated network
Severe to moderate drought prone areas
In-migration of labour
ASA- Special Session AgMIP – 4th Nov-2013
4
12/20/2013
Location of Study sites
5
12/20/2013
Study site
6
12/20/2013
Farming Systems of Study area
ORGANIC WASTE
CROPS
ORGANIC WASTE
VEGETABLES
FRUIT/ORCHARDS
HOUSE HOLD
DUNG (FUEL)
LIVESTOCK
7
12/20/2013
Understanding Current Agro-climatic
Variability
Wheat Yield vs Mean Growing Season Mean maximum
Temperature
Examining wheat
production and
observed baseline
weather in Modipuram,
Meerut District, India
8
12/20/2013
Comparison between simulated (APSIM
& DSSAT) and actual wheat yield
9
12/20/2013
Tmax (OC)
Future Scenarios- Temperature
43
39
35
31
27
23
19
3.8 OC
3.0 OC
Tmin (OC)
J
F
M
A
M
J
J
A
S
O
N
D
34
30
26
22
18
14
10
6
All GCMs predicted
higher monthly mean
maximum and minimum
temperatures during the
mid-century period
2040-2069 under RCP8.5
compared to baseline
(1980-2010). All the five
targeted GCMs
predicted more or less
same nature of
projections.
Baseline (blue line and dots)
and future (box-and-whiskers)
monthly and seasonal mean
maximum and minimum
temperature for Modipuram,
India, in the 2050s under
RCP8.5. The stars (different
colors) represents 5 “target”
GCMs (CCSM4, GFDL-ESM2M,
J
F
M
A
M
J
J
A
S
O
N
D
HadGEM2-ES, MIROC5, MPI-ESMMR).
10
12/20/2013
Rainfall (mm)
Future scenarios - Rainfall
300
200
100
0
J
F
M
A
M
J
J
Month
A
S
O
N
D
Baseline (blue line and dots) and future (box-andwhiskers) monthly mean precipitation for
Modipuram, India, in the 2050s under RCP8.5. The
stars (different colors) represents 5 “target” GCMs
(CCSM4, GFDL-ESM2M, HadGEM2-ES, MIROC5, MPI-ESM-MR).
11
12/20/2013
Uncertainty in GCM Ensemble of
Climate Change Projections
Uncertainties in maximum
and minimum temperature
Wheat growing season (December-April) mean maximum and
minimum temperature projected by 20 CMIP5 climate models
(denoted by letters A-T) for Modipuram, Meerut District, India, in the
2050s under the high-emissions RCP8.5 scenario. The red square
represents baseline conditions. The pink letters shows the GCMs
used for AgMIP home stretch.
12
12/20/2013
Crop Modeling- Sentinel data
source and treatment details
Long term data on nutrient management experiment
(1993-2010)
Years
: 2007-08 (calibration) & 2008-09 (validation)
Soil data : Profile-wise (0-150 cm) bulk
density, OC, NO3, NH4, EC & pH, LL15,
DUL, SAT and Soil texture
Crop data
: Phenology, LAI, and Biomass
partitioning at different phenology, Grain and straw
yield
Variety : PBW343
N:P:K (kg/ha) : 120-60-40
Irrigation: 5 irrigations : CRI, PI, Anthesis, Milking &
Dough
13
12/20/2013
Genetic coefficients for
APSIM and DSSAT
APSIM Genetic coefficients
DSSAT Genetic coefficients
P1V P1D
P5 G1
G2
G3
PHINT
64
748 21
32
1.1
100
74
14
12/20/2013
Crop Modeling
Observed variability in the farm data
 Crop Models used – APSIM & DSSAT
 Crop - Wheat
 Farm survey data of 76 rice-wheat farms
 Wide variability in dates of sowing - 17th
October to 3rd January
 Date of Harvest – 10th April - 17th May
 Five cultivars – PBW223, PBW243,WL502,
PBW343, UP232
 No. of irrigations – 3,4 & 5
 Variability in N, P and K applications
15
12/20/2013
Assumptions made for simulations
 Single cultivar – PBW343 – Potential yields of
5 varieties are almost same
 Irrigation depth – 5 cm
 Plant density, Plant spacing – as per
recommendations
 Soil parameters for 7 farms analyzed and
incorporated to the nearby farms
16
12/20/2013
CDF- Comparison of APSIM and DSSAT
simulated wheat yield and observed farm yield
Cumulative Probability
1.2
Observed Farm Survey
APSIM Simulated
DSSAT Simulated
1.0
0.8
0.6
0.4
0.2
0.0
2000
3000
4000
5000
6000
7000
Simulated (APSIM & DSSAT) and
Observed farm survey wheat yield
17
12/20/2013
30-year Mean Yield Change (%)
Projections of climate impacts on
wheat yield
30
25
20
15
10
5
0
-5
-10
APSIM
DSSAT
Mean Climate Changes Only
30
25
E 20
I 15
K 10
O 5
R 0
-5
-10
E
I
K
O
R
APSIM
DSSAT
Mean and Variability Changes
Box-and-whisker plot showing distribution of mean yield changes
(%) from 20 GCM-based scenarios using (right) mean-only “delta”
scenarios, and (left) scenarios including both mean and variability
changes. It also compares the results from APSIM & DSSAT
simulation outputs. The five models that are the focus of the core
simulations are represented as different color stars.
18
12/20/2013
30-year Mean Yield Change (%)
Projections of climate impacts on wheat yield –
Comparison between incorporating increase of CO2 and
without increase of CO2 effect (mean climate change only)
0
DSSAT
APSIM
-10
30
20
-20
10
-30
0
-40
-50
DSSAT
APSIM
-10
Without increase of CO2
With increase of CO2
Box-and-whisker plot showing distribution of mean yield changes (%)
from 20 GCM-based scenarios using mean-only “delta” scenarios. It
also compares the results from APSIM & DSSAT simulation outputs
with and without incorporating the CO2 effect. The five models that are
the focus of the core simulations are represented as different color
stars.
19
12/20/2013
Interactions with StakeholdersMethodology/approach
IGB
Country level Agricultural policy/developmental
agencies ,Universities, Progressive farmers,
Researchers of organizations, NGOs
State
Level
District
Level
Institute
Level
Farm level
State agricultural development agencies,
Universities, Progressive farmers, Researchers
of organizations, NGOs
Officials of district agricultural
development agencies
Universities, Progressive
farmers, Researchers of
organizations in the district
Researchers/
Scientists
Farmers
Sensitization
presentations,
discussions and
development of
RAPS
Multidisciplinary interactive meetings
with researchers/ scientists at PDFSR
during personal interactions/ group
meetings
Interacted with 140 farmers of the district
during our farm survey and listed their views
and opinions about current climate variability
and how they visualize the climate change
going to affect the agriculture
20
12/20/2013
RAPs Narrative
Climate change has adverse impact on wheat
production.
Though government adopts long-term and
short-term policy measures, wheat production
costs increase substantially.
Imports are inadequate to meet domestic
demand and assured price support policy also is
inadequate to raise the wheat production to meet
domestic demand.
Hence, government liberalizes wheat imports,
invests in food chain logistics and boost R&D for
developing new wheat cultivars to boost domestic
production of wheat.
21
12/20/2013
GAINS AND LOSSES – Model Inter-comparison (5
Climate Scenarios) AND 2 CROP MODELS (APSIM &
DSSAT) – TOA-MD results
Climate
Scenario
Future
simulated
yield
(kg/farm)
Time
Predicted
averaged
Gainers
future yield
(%)
relative
(kg/farm)
yield
E
1879
1.01
1706
50.01
29.86
30.61
0.75
APSIM
E
3176
1.19
2007
58.09
41.85
27.02
-14.83
DSSAT
I
1688
0.91
1532
43.64
24.03
36.30
12.27
APSIM
I
3425
1.28
2164
60.92
46.72
25.83
-20.90
DSSAT
K
1737
0.93
1577
45.82
25.71
33.95
8.25
APSIM
K
3213
1.19
2009
58.44
42.58
27.04
-15.54
DSSAT
O
1819
0.98
1651
48.31
28.03
31.83
3.80
APSIM
O
3179
1.19
2009
58.09
41.97
27.17
-14.80
DSSAT
R
1743
0.94
1582
45.83
25.72
33.94
8.23
APSIM
R
3218
1.20
2033
58.61
42.70
26.77
-15.93
DSSAT
Gains
(%)
Net
Losses
Losses
(%)
(%)
Crop
Model
used
Base period yield (kg/farm) – 1688; Base period simulated (kg/farm) APSIM-1860;DSSAT-1672)
(E-CCSM4, I-GFDL-ESM2M, K-HadGEM2-ES, O-MIROC5, R-MPI-ESM-MR).
22
12/20/2013
Comparison of Gains and losses through
DSSAT and APSIM under 5-GCM projected
climate scenarios.
46.7 % gainers
58.8 % gainers
DSSAT simulated slightly higher optimistic
scenarios compared to APSIM during midcentury period 2040-2069 under RCP8.5.
23
12/20/2013
Mean net farm return (USD/farm)
Comparisons of Projected Mean net farm
returns under different GCM scenarios
through APSIM and DSSAT
24
12/20/2013
Identified tested adaptation
packages
Advancement of date of sowing –
one week to two weeks from the
present
Use of Short duration varieties –
maturity before 15th March
Balanced fertilizer application
Modification of first date of irrigation
25
12/20/2013
Conclusions
All the 20 GCMs predicted higher mean monthly
maximum and minimum temperature in all the
months during mid-century period 2040-2069 under
RCP8.5 compared to baseline period 1980-2010.
Overall DSSAT model simulated higher crop
yields compared to APSIM under CO2 fertilization
scenario.
TOA-MD results predict higher percentage of
gainers (58-61%) with DSSAT as compared to
APSIM (44-50%).
Overall, Climate change situation, APSIM
predicts losses (1-12%) but DSSAT shows gains
(15-21%) in mean net farm returns.
26
12/20/2013
Thank You
27