Credit seminar

on
Crop Modelling in Fruit Crops

Presented by:
Mr. Dayyala Ramesh
Ph. D Scholar
What is crop modelling..???
Introduction
Concept
Reasons for building crop growth models
Brief history of crop modelling
Types of models
Model Structure
Model working
Model development
Uses and Advantages
Case studies
Limitations
Conclusion
What is crop modelling…???

It is a simplified description

(often, a mathematical representation)

of a system to assist

calculations and predictions.

Introduction
 The land resources are finite- no. of people continues to grow rapidly.

 Production / productivity must be increased to meet growing demands.

 Research is needed to supply info to farmers, policy and decision

makers to protect NR base, accomplish sustainable agriculture in a

climate change scenario.

 In this direction the use of crop models in research is being encouraged.

A State-Level Hunger Index for India
 Modelling

Modelling is the use of mathematical, logical or experiential relationships to predict

the status of an organism or system in response to changing conditions.
(Radha krishna murthy, 2004 ).

PROCESS OF DEVELOPING EQUATIONS ON GROWTH & YIELD

INPUT DATA COMPUTER USEFUL INFORMATION

INPUT CSM OUTPUT

Various functional
Site, soil, Weather, equations giving effect of
Results about growth,
crop characteristics Soil, Weather,
yield,
information, Management,
management needed or
Plant Management data crop characteristics on
specific information
growth and yield
 Concept
INPUTS OUTPUTS
Daily weather data: Mango

- Rainfall
- RH
- Solar radiation CROP MODEL
- Min temperature For each day of the
Citrus
- Max temperature growing season, a set of
equations is solved.
The simulated crop
grows and develops.
Soil type Grape

Apple
Planting
date
 Genetics Weather

• Yield
• Soil type
• Pests
• Elevation
• Drainage
• Fertility

•Causes of Yield Variability

•Develop Prescriptions
•Risk Assessment
•Economics Crop Model frame work

Irmak et al., 2000

Reasons for building crop growth models

 To better understand the process involved in crop production.

 To use as a managing tool in agricultural system.

 To provide a holistic and quantitative approach to problem solving with in the

agricultural field.

 To proved an analysis of the risks involved in adopting a particular strategy.

 To provide answers quicker and more cheaply than traditional experiments.

 Cornelis T. De Wit.(1924–93)
Brief history of crop modelling Crop ecologist

First attempt to model photosynthetic rates of crop canopies -- de Wit in 1965.

Elementary Crop growth Simulator (ELCROS) by de Wit et al. in 1970. This model
included the static photosynthesis model and crop respiration was taken as a fixed fraction
per day of the biomass, plus an amount proportional to the growth rate.
A functional equilibrium between root and shoot growth -- Penning de Vries et al., 1974.
The introduction of micrometeorology in the models and quantification of canopy
resistance to gas exchanges to improve the simulation of transpiration and evolve into the
Basic Crop growth Simulator (BACROS) (de Wit and Goudriaan, 1978).
To help poor farmers in the tropics and sub tropics IBSNAT (International Benchmark
Sites Network for Agro-technology Transfer) began the development of a model in 1982.
Decision Support System for Agro- Technology Transfer (DSSAT) which is currently
being used as a research and teaching tool.
Before getting started ….
• Models are assisting tools, stakeholder participation is essential
• The use of models requires high degree of technical expertise
• The merits of each model and approach vary according to the
objective of the study and they may frequently be mutually
supportive
• Therefore, a mix of models and approaches is often the most
rewarding
Models are assisting tools: Stakeholders’ interactions are essential
Scientists

Policy Technical
makers and
Applied experts
 Types of Models
According to Norton and Mumford (1993) Models divided into

Mechanistic Models Computer based

Statistical Models Models
Simple regression,

Correlation

Muitiple reggression

Stepwise regression Analytical models

Simulation models
Markovchain

Fishers orthogonal
 Types of models

Empirical model
• Generally expressed as regression equations.
• With one or a few factors.
• Eg: Response of crop yield to fertilizer application.
Mechanistic model
• Describes the behaviour of the system.
• Ability to mimic physical, chemical or biological processes.
• Eg: cell division.
Dynamic models
• Time is included as a variable.
• Both dependent and independent variables are having values which remain constant
over a given period of time.

Static Model
• Time is not included as a variables.
Deterministic models
• Estimate the exact value of the yield or dependent variable.
• Eg : crop yield or Rainfall.

.
Statistical models
• Relationship between yield or yield components and weather parameters.
• Eg: Step down regressions, correlation.
Stochastic models
• These models define yield or state of dependent variable at a given rate.
• Eg: Predicting Tsunamis
Descriptive model
• Defines the behavior of a system in a simple manner with one or more equations.
• Eg: equation gives quickly weight of crop
 Crop Parameters
Weather
surface LAI Map

Phenology map

Crop Simulation
Model
Soil Map

Soil Classes
Sandy Loam
Clay
Sandy Clay Loam
Clay Loam
Loam
Sand
Water Wheat yield
Mango Yieldmap
Map

Crop Yield Map

Wheat
Mangoarea
area
Area Weighted Yield
Table 1. Number of references dealing with crop modelling in horticulture for fruit, greenhouse vegetable,
field vegetable and ornamental productions Agris database, FAO.

Fruits Greenhouse Fruit vegetables Ornamentals

vegetables
Period
1975-1980 3 4 1 4
1981-1985 18 18 5 6
1986-1990 77 49 32 18
1991-1995 32 19 23 30

Topic
Growth 55 69 34 17
Development 61 10 25 33
Water balance 6 10 5 6
Nutrient uptake 8 4 4 3
Economy 8 8 2 4
Others (genetics, pathology, quality of product) 5 2 4 2
 Yield

Source : (Phongsak Yuhun, 2010)

Models - Advantages

 It helps in farm management decision.

 Its provide the alternative solutions of a problem.

 Conducting field experiments on all combinations are impossible.

 Reduces the risk in crop production.

 Accuracy in production forecast.

Table 2. Depending on what components are given, models can be used to give
answers to different questions and to determine the missing component ‘?’.

Controllable input Out put Structure

Analysis Given ? Given
Management/control ? Given Given
design Given Given ?
 MODEL DEVELOPMENT

Model calibration
• Calibration is adjustment of the system parameters so that simulation results reach a
predetermined level, usually that of an observation.
Model validation
• The model validation stage involves the confirmation that the calibrated model
closely represents the real situation.
Sensitivity analysis
• As a form of behavioral analysis and part of model evaluation, carried out in
order to assess the influence of selected key (‘critical’) parameters on, most
important output variables.
 Methodological steps used in model building – Vansteenkiste, 1983
Structural Knowledge Measurement Purpose Knowledge
Knowledge

Model formulation

Plant experiment, data collection,

simulation, computation

Attempt to falsify the

model

Continue Yes
Model
attempts false

Model conditionally valid

Use model until falsified

within its limitations
Module Structure and working

Kenig and Jones (1997)

 Model working

Soil data Weather data

Model

Agronomic data Plant data

Simulation

Growth Development

Yield
Moussa Sanon, 2007
General diagram indicating that as models increase in complexity from the most simple
(e.g. single driving factor), the accuracy of predictions may increase, but the
understandability of all the interactions and regulation of the model decreases
 Crop model applications
1.Research tools
a. Research understanding
b. Integration of knowledge across disciplines
c. Improvement in experiment documentation and data organization
d. Genetic improvement
e. Yield analysis

2.Crop system management tools

a. Cultural and input management
b. Risks assessment and investment support
c. Site-specific farming

3. Policy analysis tools

a. Best management practices
b. Yield forecasting
c. Introduction of a new crop
d. Global climate change and crop production
 Some crop models reported in recent literature
SUCROS Crop models
Software Details
ORYZA1 Rice, water
SLAM II Forage harvesting operation
SPICE Whole plant water flow SIMRIW Rice, water
REALSOY Soyabean SIMCOY Corn
CERES-Rice Rice, water
MODVEX Model development and validation system
GRAZPLAN Pasture, water, lamb
IRRIGATE Irrigation scheduling model
EPIC Erosion Productivity Impact Calculator
COTTAM Cotton
CERES Series of crop simulation models
APSIM Modelling framework for a range of crops
GWM General weed model in row crops Framework of crop simulation models including
DSSAT
MPTGro Acacia spp.and Leucaena Spp. modules of CERES, CROPGRO and CROPSIM
GOSSYM- PERFECT
Cotton QCANE Sugarcane, potential conditions
COMAX
CropSyst Wheat & other crops AUSCANE Sugarcane, potential & water stress conds., erosion
SIMCOM Crop (CERES crop modules) & economics
LUPINMOD Lupin CANEGRO Sugarcane, potential & water stress conds
TUBERPRO Potato & disease APSIM-
Sugarcane, potential growth, water and nitrogen stress
SIMPOTATO Potato Sugarcane
WOFOST Wheat & maize, Water and nutrient NTKenaf Kenaf, potential growth, water stress
WAVE Water and agrochemicals CLIMEX Oriental fruit fly (OFF)
ARIMA Mango, Banana, Guava
 FEATURES OF FRUIT TREE MODELS

 For modeling purposes, the fruit tree is a very complex organism.

 It is probably not by coincidence that sugar beet was among the first crop plants to be modeled. Sugar beet

has just two major organs; a rosette type crown of leaves (source) and a storage root (sink) and is therefore

relatively easy to model.

 Other annual crop plants (e.g. soybean, cotton) are more complex, having leaves, shoots, roots and

reproductive structures, but the entire lifespan is a few months.

 Fruit trees are still more complex, however, having perennial trunk, branches and scaffold roots of several

ages, in addition to a very large number of current year’s vegetative and reproductive organs.
Fig 1. Crop potential, attainable, actual yield controlling factors
CLIMEX simulated predictions of Oriental fruit fly,
Bactrocera dorsalis (Hendel) (Diptera:Tephritidae)
geographical distribution under climate change
situations in India

(Sridhar. V et al., 2014)

 The oriental fruit fly is a destructive polyphagous pest on a range of wild and
cultivated fruits. Fruit fly is poikilothermal pest.
 In mango it causes up to 31% fruit loss in India.
 Quarantine importance due to restricted distribution. Losses up to Rs. 26,902 millions.
 Spread by infested fruits, wide host range, high fecundity, food adaptability of the
larvae, short life cycle, rapid dispersal (can fly 50–100 km) and possible influences of
climate change.
 Table 3. CLIMEX parameter values for Oriental fruit fly modelling
Parameter Mnemonic Value
Limiting low temperature DV0 11.8˚C
Lower optimal temperature DV1 22˚C
Upper optimal temperature DV2 28˚C
Limiting high temperature DV3 35˚C
Limiting low soil moisture SM0 0.1
Lower optimal soil moisture SM1 0.5
Upper optimal soil moisture SM2 1
Limiting high soil moisture SM3 2
Minimum degree-day cold stress threshold DTCS 10˚C-day
Degree-day cold stress DHCS rate –0.00025 week–1
Heat stress temperature threshold TTHS 35˚C
Heat stress temperature rate THHS 0.0002 week–1
Dry stress threshold SMDS 0.1
Dry stress rate HDS –0.0001 week–1
Wet stress threshold SMWS 2
Wet stress rate HWS 0.009 week–1
Degree-days per generation PDD 358
Figure 2. Global known distribution of Oriental fruit fly (OFF)
Figure 3. Current global distribution of OFF as modelled using CLIMEX. Ecoclimatic Index (EI) values:
Unsuitable (0–0.99); Marginal(1–7.35); Suitable (7.35–13.7); Optimal (13.7–20); Highly suitable (20 +).
Figure 4. EI-based climate suitability for OFF under the reference climate (1961–1990 average) projected
using CLIMEX for India.
 a b

c d

Figure 5. EI-based climate suitability of Bactrocera dorsalis under future climate scenarios as projected
by CLIMEX: a, 2030; b, 2050; c, 2070; d, 2090.
Forecasting Major Fruit Crop Production in Bangladesh using Box-Jenkins ARIMA Model

Graphical Comparison between Original and Forecasted Series (Mohammed A. M., 2014)
Testing a Flowering Expert System for the ‘Decision Information System for Citrus’

( Albrigo, L. G., et al.,2006)

Computer screen print of the flowering model for the 2001-02 flower induction season showing accumulated
induction hours at the time the first wave of flower bud growth was predicted to start. Note the shallow
slope and low number of accumulated hours (644).
Computer screen print of the flowering model for the 2002-03 flower induction season showing accumulated induction
hours at the time the first wave of flower bud growth was predicted to start. Note the steep slope and high number of
accumulated hours (841).
Computer screen print of the flowering model for the 2003-04 flower induction season. Accumulated induction hours and
initiation of growth of two waves of flower buds with predicted bloom dates are shown as of 15 January. Note the
intermediate slope of accumulated induction hours compared to the two previous years. The number of accumulated hours
at initiation of the two waves was 778 and 974 hours, respectively.
SIMBA: a comprehensive model for agro-ecological assessment and prototyping of banana-based
cropping systems.

Tixier, P., et.al., (2004)

SIMBA MODEL: A . An intensive monoculture of banana. B. A banana after sugarcane
( Weekly and accumulated profit margine (in euros) vs. time (weekly)
SIMBA MODEL: A . An intensive monoculture of banana. B. A banana after sugarcane
( Weekly erosion and pesticides water pollution risk indicator note)
SIMBA MODEL: A . An intensive monoculture of banana. B. A banana after sugarcane
( Weekly nematode (Radopholus similis) stock per square meter)
PEST-MAN : A forecasting system model orchard pests

Figure. 6 : Schematic diagram of PEST-MAN model

(Morgan, D. and Solomon, M. G., 1996)
Table 4. Comparison of observed and predicted Julian day peak flight activity of first generation moth pests.

Adoxophyes orana Cydia pominella

Year
Observed predicted Observed predicted
1974 161 163 154 158
1975 174 174 167 164
1976 165 161 165 155
1977 165 162 150 146
Modelling assimilate transfer and fruit growth in peach

Model of a girdled stem bearing a fruit and a leafy shoot

(Bruchou. C and Genard .M., 1996)
Figure. 7. Simulation of fruit dry weight for three treatments using the parameters estimated on
the “15 leaves per fruit “ treatment. For a given treatment, a number represents a fruit the the
lines are the simulations.
 Predicting Pineapple Harvest Date in Different Environments, Using a Computer
Simulation Model

Fig. 8. Observed vs. simulated days from forcing to harvest. Fig. 9. Observed vs. simulated harvest date for Smooth
Smooth Cayenne pineapple grown in Australia Cayenne pineapple.DOY= Day of year.

(Malezieux, E., et al., 1994)

Predicting Pineapple Harvest Date in Different Environments, Using a Computer
Simulation Model

Fig. 10. Simulated and observed days from forcing to harvest

(Malezieux, E., et al., 1994)
 SiBaToKa Model: For assisting Mycosphaerella leaf spot diseases and control
Aims of SiBaToka model :
Better understanding of disease epidemics define the best efficient control methods
(eg. quantitative resistance, cultural methods)
SiBaToka model :
Mechanistic simulation model at plant scale based on infectious cycle
SiBaToka characteristics:
Stochastic
Discrete (daily step)
Modelling unit = 1 mm² (smallest lesion size)
Modelling duration= 1 crop cycle (360 d)

SiBaToka structure : 2 modules

Banana growth MLSD epidemics

SiBaToKa structure
 Applications of SiBaToKa

To evaluate the efficacy of resistance of new hybrids

Selection assisted by model (SAM)
The efficacy of control methods at plant scale
Models – Limitations
Input data: models require large amount of input data, which may not available with the user.
Skilled man power
Knowledge of computers and computer languages
Limited awareness and acceptance towards modelling
Multidisciplinary knowledge
 Simulation results have errors
A model is a tool for improving critical thought, not a substitute for it
Model formulate hypothesis and improve efficiency of field experiments, but they do not eliminate the
need for continued experimentation
Models develop for specific region cannot be used as such in other region. Proper parameterization and
calibration is needed before using a model.
CONCLUSION