Methods Flashcards
System dynamics modelling
Causal Loop Diagrams
Shows connection between components of a system and their effect on each other using + and -. This creates feedback loops that can be reinforcing or balancing.
Bayesian Networks
Is a graphical representation of a system to represent variables and their conditional dependencies.
It is helpful to understand the cause of outcomes and outcomes that arise from certain conditions. They can also be used to determine the impact of interventions or counterfactual worlds.
PESTLE
Political, Economic, Social, Technological, Legal, Environmental
Way of categorising and brainstorming
Can be restrictive, does not allow for cross category risks
Influence Diagrams
connecting components of a system = value, chance and decision nodes
Systems Mapping
Rich Pictures
using pictures to represent components of a system and showing the connections between these parts of the system
Trend Detection
Statistical Regression
This involves estimating the relationship between variables from statistics. This can involving testing multiple candidate models before evaluating to find the best fit model.
Brainstorming
Nominal Group Technique
- Brainstorm individually
- Present with no discussion
- Discuss ideas
- Anonymously vote on ideas
Change factor downscaling
Involves scaling the data produced from historical events by some factor that is representative of future change. Can use quantile-quantile scaling to represent the non-linearity of change.
Pros: Very low computational power required, high climate realism
Limits: Does not add any additional regional detail, will inherit error from the GCM, does not include specific local features, will only change the man not the distribution or temporal characteristics
statistical downscaling
Used to downscale climate data to a local scale. Uses more computational power than change factor but less than dynamical. Assumes that current climate statistical relationships are valid under future climate conditions.
Pros: Can generate sub-daily info, no restricted to length of GCM run
Limits: cannot represent inter-variable or spatial dependencies, underestimate variability, can be limited by training and be unable to produce events that have not been historically observed, specialist knowledge required
dynamical downscaling
Use to downscale climate data to a local scale. Highest computational power and can only be run on a subset of GCMs.
Pros: provide better representation of topography and its interactions, has capacity to simulate processes the GCM cannot resolve, can represent interactions between change in large scale features and local features such as complex topography or urban environments.
Limits: cannot overcome systematic bias in GCMs,
Integrated assessment models
Collects social, economic, atmospheric and biospheric phenomena
