17/4 - breeding and modeling Flashcards
What is a model?
A model is a human construct that facilitates the understanding of real world systems
Also a metaphor for highly (overly?) simplified scientific models of complex real life phenomena
What a model is NOT?
Model ≠ Mathematics ≠ Modelling software
Mathematics–the formalism for quantitative models.
- Vad berättar detta?
- how the model does what it is supposed to do
- a simple, clear of
epresentation antitative mode
Modelling software — the tool a tool to facilitate calculations many options:
- Vad berättar detta?
- a tool to facilitate calculations
- many options: generalized
spreadsheet; modelling
software, high level
programming language
Crop modelling –What for?
- Crop models for research
- compare teories
- fill knowlage gaps
- get research questions
- Prioritize experiments
- Quantify expected results, with complex interacting factors:
- improve genetics or management
- Predict the future:
- klimate change
- Synthesize knowledge across disciplines
2. Crop models for crop management - Describe the effects of complex factors
- utgå från tidigare fall to predict
- Definition of best management practices
BUT
Applications are limited by availability and quality of input
User needs to understand model structure to grasp limitations
Steps to modelling
1) Definition of model purpose and system boundaries
2) Conceptualization: a conceptual, verbal description of the interactions within the system and its behavior
3) Quantification: coupling of functions, rules, equations to describe quantitatively the interactions within the system and its behavior
4) Calibration: adjustment of the model parameters to improve the representation of the system by the model
5) Validation: checking the accuracy of the model’s representation of the real system
6) use
7) Comparison with collected data - back to 1
Comparison with collected data - hur ta fram en modell för tillväxt som resultat för c fixation
Goal: model plant biomass growth
Systekm boundaries:
a single plant
A verbal description
plant biomass increases as a result of carbon fixation; a larger biomass
results in more carbon fixation”
A graphical representation: Causal Loop Diagram
- plant biomass påverkar c fixation posetivt som i sin tur påverkar plant biomass
Model calibration and validation
Model calibration
get the parameter r (increase in plant biomass per unit existing biomass and unit time) to best
describe the observed data
Model validation:
compare the result with another dataset
Plant biomass growth -Logistic
- plant growth utifrån c fixation och available nutrients
A verbal description:
plant biomass increases as a result of carbon fixation; a larger biomass
results in more carbon fixation, but a very large biomass reduces available
resources and hence carbon fixation and hence biomass growth
graf: casual loop diagram
lägger till ytterligare en funktion
Exponential growth and Logistic growth
Exponential growth: One parameter only; works well early in the growing season
Logistic growth: Two parameters; needed for late in the growing season
Descriptive
vs.
Predictive
Descriptive: Simply describe the observations, within the context of the current experiment
Predictive: Extrapolate beyond the scope of the experiment and current results - fortsätter framåt utifrån biomassen man har i början
Empirical (or functional)
vs.
Mechanistic (or process-based)
Empirical : Relying on (statistical) description of the observations (e.g., regression models)
Mechanistic: Constructed around a mechanistic understanding of
the underlying processes; explicitly representing
processes and cause and effect relations; can be used
for predictions and extrapolations (up to a certain
point)
Note that all process-based models become empirical
at lower levels of organization
Vad visar emperical och Mechanistic
Empirical model: parameter r obtained
from curve-fitting to data
Mechanistic model: describes th
processes leading to plant growth an
embeds them into the parameter
Vad baseras Mass balance på
based on the principle of conservation of mass
vad baseras Energy balance på
based on the principle of conservation of energy
Vilka varibles finns i modellen med växten?
Carbon Water Nitrogen Biomass = mass balans
Available solar
radiation
Leaf temperature
= energy balance
vilka kontroll volumes är av intresse i växten?
Entire plant or single leaf
Soil volume explored by the roots
Vilka growing conditions kan påverka växten?
Temperature and developmental stage
Effects of water limitation
model 0
Nollhypotesen är hypotesen att det inte föreligger något fenomen som kräver en förklaring.
När är det fördelaktigt att ha en eller flera parametrar
en: början i växtens liv
fler: slutet av växtens liv
Rate of photosynthesis
for given CO2 concentration at the photosynthetic site, gross photosynthesis is determined by enzyme kinetics (“RuBisCO” -Ribulose-1,5-bisphosphate carboxylase/oxygenase) and light availability
Most common modelling approach: Farquhar’s model
For CO2fixation, choose the smallest between two potential rates of CO2 fixation
Limited by enzyme (RuBisCO) and light
Vad är fotosyntes och respiration - vad behövs och vad bildas
Photosynthesis:
CO2, H2O and light to produce O2and glucose
Respiration:
O2and glucose to produce energy that plant cells can use, CO2and H2O
Vad driver co2 supply?
CO2supply driven by stomatal aperture, i.e., stomatal conductance to CO2(and water vapor)
CO2 concentration at the photosynthetic site= a fraction of atmospheric CO2concentration
Leaf level –C uptake
The result of demand and supply
där demand och suply möts är the working point
Vad är cue?
CUE: Carbon Use Efficiency, i.e.,
net primary productivity per unit C assimilated by photosynthesis
CUE≈0.4-0.6 because of -construction costs -cell/ion concentrations/gradient -active N uptake -protein resynthesis
Roots, stems, leaves or seeds?
Allocation patterns
Trade offs implicit in allocation patterns:
- Large allocation to leaves, low allocation to roots: high photosynthetic capacity, but likely low water and nutrient availability
- Large allocation to root, stems and leaves, low allocation to seeds: low reproductive ability
EFFECTS OF GROWING CONDITIONS
temp and h2o limitations
Degree-day model:
Based on the concept of cumulated degree days +
thresholds corresponding to different stages
Photothermalunits:
Based on the concept of cumulated degree days + simple dependence on light availability + thresholds corresponding to different stages
Modelling varieties
Change parameters:
- Metabolic parameters
- Allocation patterns
- Rooting depth
- Length of developmental stage
- Response to stressors
vad skapar fotosyntesen för demand?
Photosynthesis creates a demand for CO2.
Such demand of CO2is met by CO2uptake from the atmosphere through stomata, at the cost of water transpiration.
The demand of water is met through water uptake from the soil, taken up by the roots and transported through the plant.
external limitations are accounted for by a two step approach
1)
definition of a ‘potential’ (in the absence of limitation)
e.g., yield, or at a lower level, assimilation rate under well-watered or high fertility conditions
2) introduction of a (often empirical) correction
Är alla modeller ekvationer?
Kan formulera modellen med equations men innan dess måste du ha the essense - din understanding av vad som händer
Hur vet man vilken nivå man ska jobba på?
Karta över sv eller en över världen - beror på vad man vill ha ut
Måste ha rätt level of complexity för att fungera I praktike
What questions can models answer?
Depens on the model
Predicted values based on data
Siplifye representation of realiety
Dercribe data
Why and how are simulations models useful in crop research and strategic management?
Predict what can happen in future
Make amodell of last year and apply it next year - not predictive but describing
What is the difference between mechanistic (or process-based) and empirical models?
And when is ‘best’ to use one or the other kind?
Empetics - kolla bara på sttestik - om sloen lyser mycket så växer det mer
Mecanical - care abourt the underlying mekanism - it is the fotysyntes that makes the plant grow
Mekanisk - vill förklara why - ändra någon parameter
Emperical - när använder samma system
Predict: behöver ej vara över tid utan även under andra förutsättningar tex har en modell för lite ljus och ska ha en med mer ljus
How can the role of modelling uncertainties be assessed?
P value - accurensy
Calibration
Validation to chek it
Var sätter jag in errors?
Då baserad på prediction är den bara så stark som the prediction is - hur mycker vet vi om tex future? Väldigt lite
Ha olika teorier bakom men kommer fram till samma resultat - ej så känslig för condition
Hur vet jag om den r känslig för olika conditions - running different senarios
Modellen kan var känslig/okänslig för de assomptions som tas
Twerk parametrarr och se om resultatet blir det samma
- sensetivity analysis
Tar ej med specefika frågor från seminariet
nope
2 aspects att alltid ha in mind
Energy and mass is conserved - bildas aldrig nytt
Formuleras som mass balance and energy balance
Mass balance - stuff dyker bara inte upp, det kommer någonstanns ifrån
Vad beror limits på?
på skala och faktorer
Varför är det viktigt att veta hur modellen ser ut?
you need to know whats in the
model, so you know whats going
to happen, its not magic
Hur spelar kunskapen man har in i skapandet av modellen?
limited knowledge = limited quality on the model
the model cant generate knowledge for you
DAS
days after sowing
CDD (C)
= measurment of maturity
