Transcript Climate Change Impacts On Rice Farming Systems in North

Climate Change Impacts On Rice Farming
Systems in North Western Sri Lanka
Lareef Zubair, S.P. Nissanka, S.Ariyawansha, A.P.Keerthipala, B.R.V. Punyawardhene, V. Ralapanawe,
W.M.W. Weerakoon,. K.D.N.Weerasinghe, P. Wickramagamage, P. Agalawatte, S.C. Chandrasekara, A.L.C.DeSilva,
C.M. Navaratne, J. Gunaratna, D.I. Herath, R.M. Herath, A. Karunaratne, S. Ratnayake, P. Samaratunga,
K.Sanmuganathan, J. Vishwanathan E. Wijekoon, Z. Yahiya. ARP - Daniel Wallach
Sri Lanka AgMIP team at Launch
Workshop in Colombo.
1. Introduction
Locations of meteorological stations and studied
farming systems (Centered at 07.52N, 80.43E).
3. Climate Analysis
Current (Black line and stars) and future (box-and-whiskers)
monthly and seasonal temperature (top left) and precipitation
(bottom left) for Batalagoda in the mid-20th Century.
Corresponding annual and seasonal estimates are at right.
Current (square) and future (letter key for GCM’s) mean Yala
(left) and Maha (right) estimates for rainfall and temperature.
The climatology of rainfall for
Batalagoda in Kurunegala depicts
bimodal rain peaks from April to June
and from October to December. Rice
growing seasons (Yala and Maha)
commence with these rainy seasons.
The climate projection is for the midcentury (2040-2069) based on 20
GCMs under higher concentrations
(RCP8.5). The future projections
following protocols of AgMIP shows
an increase in temperature by 1.5-2.8
⁰C for Yala and 1.1 to 2.4 ⁰C for Maha
and an increase in rainfall by -0.5 – 2
mm/day during Yala and -0.1 to 2.6
mm/day during Maha. The rainfall in
the future period is similar to the
historical period except for September
to December when the models project
a significant increase.
5. Crop Analysis (Yield Simulations)
Goals -To understand the impact of climate on
current farming systems under climate change,
with and without adaptation, crop simulations
were undertaken for each system. Inputs:
Weather data: from observations and GCMs;
Fertilizer and water application: gathered
through the farm survey, Soil: Parameters
were estimated from the relevant fields.
Results: Yields simulated in DSSAT using
DOME input files and AgMIP IT tools are
shown. The yield based on two GCMs lead to
lower yields than the baseline by (<10%) while
that based on the other three models show a
small increase. This variation is consistent with
the rainfall variation among the models.
Comparison with other farming systems:
The University of Peradeniya team finds drops
in yield for Maha and particularly Yala for all
GCM’s. Impact of Adaptation: This work
shows that the use of shorter duration varieties
and changes in the planting dates helped
mitigate losses due to climate change.
Schematic of Rice Dominated Agricultural system that was simulated.
organic fertilizer use is limited
Home garden system is well
•established
with high diversity
Calibration of Crop Models: Genetic coefficients
were estimated using multiple methods (sensitivity
analysis and GLUE) based on experimental
observations from multiple field and variety trials.
Left: Comparison of observed and simulated phenology for
Bg 300 - days to Panicle Initiation, to heading and to harvest
maturity. Right: Observed and Simulated Yields for
experiments and 27 variety trials for Yala and Maha.
Comparison of observed and simulated anthesis day for
nine rice cultivars (Left), comparison of observed and
simulated yields at harvest maturity (Right)
Calibration for APSIM Work presented
by S. Nissanka at the University of
Peradeniya on the calibration of APSIM
models is shown on the left. The
comparison of simulated and observed
phenological parameters and yield of the
cultivar Bg 300 is shown to the left. This
used 27 data series including 3 detailed
experiments and 24 variety trial at
Batalagoda and Mahailuppallama for
both Yala and Maha seasons.
Calibration of DSSAT Graphs shown on
the left depict the comparison of the
simulated and observed anthesis dates
derived from the GLUE methodology for
one set of experiments. The crop was
not under water or nitrogen stress. The
comparison is good for anthesis days
after planting and for other phenological
parameters as well. The comparison of
observed and simulated yields too are
satisfactory. There remains a need for
further review.
Distribution of Relative yields and Net farm returns: System 2 (Future) for the 5 GCMs
Box-and-whiskers (Top) Probability of exceedance
(Bottom) of mean yields simulated using DSSAT for
baseline, CCSM4, GFDL-ESM2M, HadGEM2-ES,
MIROC5 and MPI-ESM-MR climate scenarios
8. Conclusions
Findings
Simulated yields
Per Farm
As a percentage of
Poverty (%)
Percent
(Kg/Ha)
Relative (Rs/Season) mean farm returns(%)
RCP
Gainers
yields
With Climate
(%)
Baseline Future
Gains Losses Gains Losses Net Loss
Current
change
8.5
4071
3450
0.85
1077 7266.6 2.53 17.08
14.55
22.14
38.4
46.4
(1297) (1278) (0.13)
8.5
4071
3471
0.85
1141 7105.1 2.68 16.70
14.02
23.27
38.8
44.5
(1297) (1306) (0.15)
8.5
machinery is significant while
The Economic analysis is based on a farm survey carried in the Kurunegala district, for
2011/2012 Maha season. The population represents rice farms cultivating 3-4.5 month rice
varieties. For the selected sample of 31 rice farms, average farm and household sizes range
from 0.2-1.0 ha and 2- 6 persons respectively. TOA-MD model was used to analyze the socioeconomic impacts of climate change for five GCMs under Climate change without adaptation
scenario. DSSAT based crop simulations were used for the analysis. The economic and crop
variables were parameterized according to AGMIP procedures.
Climate Impact Assessment
TOA-MD results show contrasting results
for the five GCMs selected for the study.
Whereas, CCSM4 and GFDL-ESM2M
show more losers due to the effect of
climate, HadGEM2-ES, MIROC5 and MPICGCM3 GCMs show higher yields, hence
more gainers. Assessment with adaptation
shall follow as shall analysis for other
farming systems.
Climate Impact Assessment Summary table (Standard Deviations in parenthesis)
MPICGCM3
depending on water availability
•Use of inorganic fertilizers and
6. Economic Analysis
7. Synthesis Diagrams
HadGEM2 8.5
-ES
8.5
MIROC5
the main activities and salient
features Rice and other field crops (i.e.
•Chilli,
onions, pulses, etc) are
grown in seasonal rotation
4. Crop Analysis
CCSM4
GFDLESM2M
Rice farming systems vary primarily with climate, the extent of irrigation, other
environmental and socio-economic considerations. We have worked in several rice
dominated farming systems including in the Kurunegala district. This district has a paddy
sown extent of around 66,500 ha in the Maha season commencing in October (average
rainfall and temperature: 5.5 mm/day and 28.8 oC) and 41,500 ha in the Yala season
commencing around May (average rainfall and temperature: 3.1 mm/day and 29.1 oC). The
climate is “intermediate” in Sri Lanka terms and there is partial irrigation from perennial
rivers (Kala Oya, Deduru Oya), 25 major irrigation schemes and over a thousand small
village tanks.
Prototypical farming system
The schematic to the left shows
Our goal is to quantify the socio-economic impacts of climate change on rice dominated
farming systems incorporating state-of-the-art climate products as well as crop and
agricultural trade model improvements in selected rice farming systems and assess
adaptation options. The climate projections are based on 5 relatively skillful (over the
region) Global Climate Models in the CMIP5 archive; the crop models used were DSSAT
and APSIM; and the Economic modeling used the Tradeoff Modeling – Minimum Data
software (TOA-MD). Consideration of future projections and adaption overs
“Representative Agricultural Pathways” (RAPs) used the DevRAP framework.
CCSM4
Rice Farm in Kurunegala and
varieties of Rice Seeds
2. Farming Systems
Scope: While having undertaken work under the
AgMIP program across Sri Lanka for rice and
sugarcane dominated farming systems at six
institutions, we report selectively here on the
analysis for the North-Western Kurunegala district
for Rice. Importance: Rice is the staple food of 21
million Sri Lankans and is cultivated over 4,800 sq
km. Previous work shows the high sensitivity of
rice to climatic parameters such as rainfall,
streamflow, temperature, solar radiation and El
Nino states. Stakeholders at farm level and at
multiple administrative, resource management
scales have all expressed the need for better risk
management and adaptation for climate.
GCM
Photo
4071
(1297)
4433
(1372)
1.09
(0.19)
7226 1897.3 16.97
4.46
-12.52
70.46
45.6
33.6
4071
(1297)
4071
(1297)
4430
(1439)
4325
(1350)
1.09
(0.14)
1.06
(0.14)
6311 2176.7 14.83
5.12
-9.71
66.57
44.5
34.1
5308 2484.1 12.47
5.83
-6.63
62.19
43.1
35.0
• Climate, crop and economic models have been implemented with multiple data sets by
project personnel with adequate data even of variable depth, quality and accessibility.
• Climate projections show that a consistent clear increase in temperature in the midtwentieth century for a high concentration pathway. The rainfall shows an increase on
average for a preponderance of models, with the highest increase for early Maha(Oct-Nov).
• The crop models simulations showed relatively lower yields (of around10%) for three
GCM’s and similarly relatively higher yields for two GCM’s under these climate projections.
• The TOA-MD analysis showed that the number of gainers from climate change declined
and poverty levels increased for the three simulations based on the projections that showed
yield declines. The converse held for the models that showed yield increases.
However,
• Analysis on a different farming system shows that there was a drop in yields in both
seasons with it being particularly deleterious in Yala.
• In the latter case, use of shorter duration varieties and changes in the planting dates led to
recovery of losses.
Ongoing Work
• Analyzing the consequences of alternative Representative Agricultural Pathways (RAPs)
• Inter-comparison of results from different models, calibration methodologies and different
groups / farming systems.
• Review of skill over Sri Lanka of GCM's
Work on Sugarcane
• Climate and crop modeling has bee carried out and farm survey data has been collected.
Acknowledgements
We are grateful to our data providers in the Department of Agriculture, Meteorology, Mahaweli
Authority, Soils Science Society, Sugarcane Research Institute and individual researchers.
The support of. Alex Ruane, Sonali Mcdermid, John Antle, Roberto Valdivia, Ken Boote, Peter
Thoburn, Carolyn Mutter, Jim Jones, Cynthia Rosenzweig, Jerry Hatfield and others at AgMIP
has been essential and unstinted. The support of our Resource Person, Daniel Wallach and
the Regional Coordination team led by G. Dileepkumar has helped tremendously. The
support of Dakshina Murthy, S. Nedumaran, Lieven Claessens, Raja Reddy, Amor Ines, Eric
Justes, D. Ripoche, F. Ruget is gratefully acknowledged. The assistance of other colleagues
–S.K. Herath, M. Weerasekera, Yasas Harischandra, Y. Indrachapa, S.B. Weerakoon,
Navoda Mihraj, Thilina Abeysekera, W.R.S.S. Dharmaratne, and S.K Cyril is gratefully
acknowledged. We thank our colleagues at FECT, the Mahaweli Authority of Sri Lanka,
Department of Agriculture, Sugarcane Research Institute, Universities of Peradeniya,
Ruhuna, and Rajarata for their support. We are grateful to DfID and AgMIP including
ICRISAT, USDA, Columbia University for making this work possible.