Roots, shoots and leaves: the origins of GLAM(-JULES) - Andrew Challinor
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Institute for Atmospheric Science
Roots, shoots and leaves: the
origins of GLAM(-JULES)
Andrew Challinor
A.Challinor@see.leeds.ac.uk
with
Tim Wheeler, University of Reading
Combined crop-climate forecasting
Find spatial scale
of weather-crop
relationships
Crop modelling
at appropriate
spatial scale Hindcasts with
observed weather
data and reanalysis
The spatial scale of yield-weather
relationships: EOF analysis
First EOF of sub-divisional yield Yield anomaly for 1985
(23.4% of variance)
Challinor et. al. (2004)
Principle component time series
• Coherent associated large-scale First principal
patterns of seasonal rainfall and component of
groundnut yield in India rainfall
yield
and PC3 of
850hPa
circulation
r2(ppn,yield)=0.53
r2(circ,yield)=0.45Combined crop-climate forecasting
≈ scale of the
climate model
Find spatial scale
of weather-crop
relationships
Crop modelling
at appropriate
spatial scale Hindcasts with
observed weather
data and reanalysisCrop modelling methods circa 2000
• Empirical and semi-empirical methods
+ Low input data requirement
+ Can be valid over large areas
– May not be valid as climate, crop or
management change
• Process-based
+ Simulates nonlinearities and interactions
– Extensive calibration is often needed
– skill is highest at plot-levelHIt
General Large Area Model Y W
TE
for Annual Crops
Rt SRAD
D Tr
T
LAI
PPN
(GLAM) CYG RH
SWF ASW
Challinor et. al. (2004)
• Combines:
– the benefits of more empirical approaches (low input
data requirements, validity over large spatial scales)
with
– the benefits of the process-based approach (e.g. the
potential to capture intra-seasonal variability, and so
cope with changing climates)
• Yield Gap Parameter to account for the impact of differing
nutrient levels, pests, diseases, non-optimal management
etc.T
Esc dLv / dL
dL / dt
soil water EFV
stress factor
Lv is root length
density
EFV is extraction
front velocity
Plenty of soil layers LAI is leaf area
index
Convergence of results at n=25General Large-Area Model for annual crops
Results: groundnut yield in Gujarat
Challinor, A. J., T. R. Wheeler, J. M. Slingo, P. Q. Craufurd and D. I. F. Grimes
(2004). Design and optimisation of a large-area process-based model for annual
crops. Agricultural and Forest Meteorology, 124, (1-2) 99-120.The yield gap parameter
Potential crop yields
GLAM uses a time-
independent site-specific yield
water gap parameter
Yield potential
nutrients
GLAM can be used to simulate
yield potential, but:
pests & diseases
• Less useful
many others ... • Validation data harder to find
Farmer’s yieldsTo what extent are mean yields
determined by YGP?
• YGP values differ when calibrated on different input data; Hence it
contains an element of bias-correction.
• However, does this mean
Colours show YGP:
that the mean yields are Black 0.05-0.15
tuned to correct values by Green 0.45-0.55
varying YGP? Red 0.9-1.0
Symbols show accuracy in
simulation of mean yield:
Circles: within 5%
Squares: within 10%
Triangles: within 25%
+ : within 50%
X : within 100%Optimisation of global parameters
for groundnut across India
Thin – IITM Optimal values are within
Thick - CRU literature range (~ 1.3 - 4 Pa)
Optimal values are stable
over space and input
dataset provided CYG is
calibrated
CYG can correct for (some)
UP – solid
GJ – dashed data input bias
AP – dotted
X – IITM-calibrated UP CRU runModel response to rainfall and
radiation
Irrigated GLAM RUE =
0.71, 0.99, 1.00 g/MJ
Yield (kg/ha)
Rainfed GLAM RUE =
0.83, 0.63, 0.40 g/MJ
Observed values e.g.
0.74 (Azim-Ali, 1998),
1.00 (Hammer et. al.
Rainfall (cm) 1995) g/MJ
UP, AP, GJGeneral Large-Area Model for annual crops
Results: all-India groundnut yield
900
850
800
750
Yield (kg ha -1)
700
650
600
550 Model results
500
Observed yield
450
(detrended to 1966 levels)
400
1965 1970 1975 1980 1985 1990
Year
Challinor, A. J., T. R. Wheeler, J. M. Slingo, P. Q. Craufurd and D. I. F. Grimes
(2004). Design and optimisation of a large-area process-based model for annual
crops. Agricultural and Forest Meteorology, 124, (1-2) 99-120.Combined crop-climate forecasting
≈ scale of the
climate model
GLAM-JULES
Find spatial scale Skill in the ensemble
of weather-crop mean and in the spread
relationships
Fully coupled crop- Seasonal
climate simulation forecasting
(ensembles /
satellite)
Crop modelling
at appropriate
spatial scale Hindcasts with
observed weather
data and reanalysis Climate
change
Perturbed parameter
simulations (QUMP-style)Conclusions Land surface and boundary-layer modelling based on a sound understanding of underlying physics and biology (and chemistry) enables • Prediction using robust process-based models • ‘Large area’ modelling, which is especially useful as computer power and resolution increase • Quantification of uncertainty due to both parameter values and model formulation • Study of interactions between processes
General Large Area Model for Annual
Crops (GLAM)
d(HI)/dt
Yield Biomass
transpiration
efficiency
Root system
Development Transpiration radiation
stage temperature
rainfall
Leaf canopy RH
water Soil water
CYG
stressYou can also read
