Teaching block 4 - systems and modelling - (weeks 10 & 11) Flashcards
(27 cards)
what is a system give an example
a collection of inter related objects eg the food and agricultural system or a leaf
what is an object
essential unit of which observations can be made
system boundary
separates internal processes from in and outputs (dotted line)
state variable
interacting component that defines the systems structure (linked via fluxes)
fluxes
flows transferring things with in the system
forcing/drivers
external variables that drive the system
closed system
no exchange with surroundings (mass and energy of the system must remain constant over time)
open system
inputs and outputs are imbalanced, must be accommodated by a change in the mass/energy of the system
explain the system of a leaf
state variables = photosynthesis, respiration
fluxes = sugar transfer, solar radiation, Co2, o2
drivers = light intensity, co2, humidity level, night/day
what is a 1 box model
hugely simplified only contains one state variable a few fluxes and a system boundary
what is a 2 box model
less simplifies - contained a few state variables and fluxed between them
what time frames can systems occur over
hourly (precipitation event)
daily (day, night P cycle)
yearly (canopy growth)
why do feedback’s occur
when a process in a system is dependant on one or more state variable
they create complex non-linear relationships and can be positive or negative
what is a perturbed system
a system containing a feed back that has been “moved”
give 3 examples of positive feedback
- sun melting ice and reducing albedo
- photosynthesis and canopy growth (more P, more C so more growth)
- canopy cover and forest fire (forest fire reduces shading and dries canopy, canopy heats and becomes more flammable)
give an example of negative feedback
stomata and water loss - If water loss exceeds supply stomata close, reducing transpiration (water loss) .
how do feedback’s drive system complexity
response to perturbation may bot be reversible (if system pushed beyond point of bouncing back to initial state)
-
why must system errors be quantified and well communicated
- policy makers need to know how accurate it is for policy making
- public trust in modelling things such as climate change
what are some uncertainties with systems (3)
- real life system converted into maths equations (not all processes and interactions would be considered)
- uncertainty in data collection (incomplete data and small errors will intensity in model so unreliable predictions made
- model values of parameters may be inaccurate. small mistakes amplified
what does equifinality mean
multiple model designs can come out with the same result
what can models do (example)
simulate how environments will respond to change over time in response to environmental forcing.
what are the first steps to setting up a model
- set working directory
- load functions & libraries for model
- look at the data frame and plot graphs with the data to understand it more
- run model parameters eg GPP, NPP, NEE, and biomass
what happened if model parameters are not a good representation of the ecosystem?
the model will be unable to model ecosystem dynamics.
the model should be calibrated to overcome this issue.
how can a models performance be checked
- compare it to actual observations & data
- calculate residuals and analyse with Sharpio wilks test. look for a low p-value here. low p-value = dist of residuals is significantly different from a normal dist,
- make sure all data points are positive
- residuals should be normally distributed
