Component based modelling
A modular structure of model components allows for an independent construction of processes at their specific spatial and temporal scales. Common interfaces allow for a standardised communication, and facilitate a more straightforward coupling of various model components into integrated models.
With model components operating at different spatial or temporal scales, a modeller needs to be able to bridge these discrepancies at component coupling. An option to align model components is to modify the internal processes. This approach, however, limits the concept of generic, reusable model components.
We propose the accumulator as interconnected building block aligning different spatial and temporal characteristics of coupled model components. The figure below shows the accumulator (A) concept in the coupling of two model components with shorter (C1) and larger (C2) time step. The accumulator obtains model component outputs for a specific interval, and performs an aggregation operation on this interval. The accumulator can be used, for example, to aggregate the daily values to a monthly mean value without changing internals of the model components.
We develop a modelling framework providing building blocks for the construction of spatio-temporal model components, and building blocks for their coupling. The framework uses the object-oriented features of the Python scripting language, and provides class templates as building blocks for the model construction.
To demonstrate the functionality of the framework, we develop a case study coupling three model components with different processes and time steps. The first component simulates the amount of biomass by logistic growth with a yearly time step, the script below shows the implementation for the biomass model component. The second component simulates random fire events and fire spread based on topography and available biomass with a daily time step. The third component simulates harvest on a yearly time step. The available biomass values are yearly updated in the fire and harvest model components. The areas affected by fires are aggregated to yearly total maps and used to update the biomass and harvest components.
The image below shows a set of output maps for the biomass model component. Lighter spots in the catchment result of biomass removal caused by fire in the previous year.
The following animation shows the Aguila visualisation tool used to display the resulting maps of the fire and the biomass model components. The top window shows the result maps of the fire model component, the bottom window the result maps of the biomass model component. The first part of the animation shows the development of a fire within one year, with a daily time step. The remainder of the animation shows a fast motion of the remaining years. The time series window shows the amount of biomass for the selected cell.
For further information please contact email@example.com