Introduction of Human Body & Structure of cell.pptx
Exploring management options to increase pulse production by using simulation models: An application to dryland systems
1. Exploring management options to increase pulse
production by using simulation models :
An application to dryland systems
H. Marrou, M. E. Ghanem, A. Soltani,
T. R. Sinclair
2016 International Conference on PULSES
“For Health, Nutrition and Sustainable Agriculture in Drylands”
Marrakech, Morocco
18th - 20 th April 2016
2. Context :
Resource use efficiency to be increased
Improved field
management required
Efficiency
• Timing
• Fractionning
• Adequation to crop demand
For water and N application
• Low resource availability Increasing food demand
2
3. • Farmers decision for resource allocation to production
• Crop rotation and crop partition in the farm land
• Farming systems to be re-designed
3
Context :
Resource use efficiency to be increased
4. 4
How crop models can help?
• Give order of magnitude for yield gains from new
plant traits and management options
• Assess risk and variability
• Decompose efficiencies and identify levers of action
• …
5. The SSM Model
Process
driven algorithm
Daily time step
Climate (4)
Crop (48)
MEASURABLE
Soil (11)
Management (2)
Crop
Soil
Transparency and accessibility
of the model algorithm
5
Soltani and Sinclair, 2012
6. Example 1 : bean irrigation management
In the South of Europe, is it possible to reduce irrigation water use
without impacting the yield?
farmers with different
irrigation schedule
5 years
22 different cropping
situations
4 irrigation scenario:
a) actual farmers’ irrigations
b)Model triggered irrigations
at 3 levels of soil water
content
6
7. Simulated yield with farmers’ and model
triggered irrigations
Simulated cumulated irrigation (farmers’ and
model triggered irrigations)
Marrou et al., 2014
Example 1 : bean irrigation management
In the South of Europe, is it possible to reduce irrigation water use
without impacting the yield?
7
High moderate no Actual
stress stress stress
High moderate no Actual
stress stress stress
8. • Farmers’ irrigation timing doesn’t match plant needs
• Degrading of the results and translation into practical advise :
– First irrigation within 20 days after sowing
– Irrigation in the 20 days before flowering is not necessary
– Support grain filling with irrigation until harvest
Marrou et al., 2014
Example 1 : bean irrigation management
8
9. Model evaluation :
• Using long term experiments at Tel Hadya (ICARDA station)
Example 2: water resource allocation at farm level
On a cereal –legume far, with reduced acces to water and fertilizer, could it
be a good bet to irrigate legumes?
9
Virtual experiment : 6 Middle-East farms, growing wheat and chickpea
• Same area: 5ha
• Same water availability : 2500m3 (50 mm * 5ha)
• Different rotations and wheat/chickpea proportion on the farmland
• 2 possible irrigation strategies:
i) 100% of farm water goes to wheat
ii) 50mm is kept for chickpea and the rest is applied on wheat
30 year simulations (historic weather data)
10. 10
Results : - For all farms
1. Food production hardly affected by irrigation strategy
2. Irrigating legumes decreases N requirement to achieve yield permitted by
water
Example 2: water resource allocation at farm level
On a cereal –legume far, with reduced acces to water and fertilizer, could it
be a good bet to irrigate legumes?
- For farms with reduced access to fertilizer
1. Saving water for wheat allow higher yield permitted by water
2. …but N stress leads to poor water use efficiency by wheat
3. Irrigating legumes contributes to meet N demand of cereal after legume in a
rotation
11. Take home messages
1. Even if they are « wrong », models can give directions to improve
legume production
2. To allow so, models should be transparent, process oriented and non
calibrated
3. Solutions to improve resource use efficiency in legume systems are to
be looked for at different scales
4. Comprehensive assessment of resource availability is needed to really
improve efficiency
11
Notes de l'éditeur
Questions
Give order of magnitude for yield gains from new varieties or new practices at field and farm scale
Assess risk and variability
Decompose efficiencies and identify lever of action
…
From the 90’s utopias …
Modeling every process for more « realism »
Modeling the impact of fine crop tuning at large scale (long time scale, economics)
Calibrating and the quest for 0 imprecision
Get the « right » universal model
… To parcimonious and process oriented modeling
Quantify and accept uncertainty
Use process oriented, non calibrated models
Adapt the model formalisms to temporal and spatial size of the system and data availability