Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Do you really want to delete this prezi?
Neither you, nor the coeditors you shared it with will be able to recover it again.
Make your likes visible on Facebook?
You can change this under Settings & Account at any time.
DISPLACE - Introduction to Model Architecture & Parameterization
Transcript of DISPLACE - Introduction to Model Architecture & Parameterization
Look at individual vessels
of the bigger picture
Do you need
Introduction to model architecture
A Dynamic Individual-Vessel based Modelling for Fishing Planning and Spatial Displacement
a modelling approach to anticipate interlinked effects on fish and fisheries of alternative management and spatial plans
like a flight simulator...
but still possible to look at spatial
...then take a
Images used under license from Shutterstock.com and colourbox.com
Mistakes or uncertain parameterization is likely to create crashes...which are still hard to detect (WIP)
Hands in the c++ code?
...and individual decision-making
...and individual footprint
Raw input data
post processing routines
Logbooks coupled to VMS data (at best)
Spatial distribution of stocks
Stock biological features (N at size, Wt at size, growth transition matrix, etc.)
Individual vessel building incl. specific fishing grounds and harbours
Scenario crafting including TAC mngt, fishing closure, (fishing credits), etc. and decision trees possibly indexed on external time series
Geographical arena crafting (grid of cells/nodes at sea e.g. of 4 by 4 km and harbour nodes)
Fish prices on harbour nodes
Stock-, vessel-, and metier- specific spatial catch rate
Metier building for size-based selectivity accounting for gear type
Other non-explicit vessels depleting on stocks
https://github.com/frabas/DISPLACE_GUI (Not Yet Public)
https://github.com/frabas/DISPLACE_R_outputs (Not Yet Public)
...for Windows and Linux
For users, parameterizing a new app:
For users, testing the software with an illustrative dataset:
For (c++) developers, looking at core simulator and GUI:
e.g. lgbooks, VMS, ICES, fishbase,
surveys like IBTS , GIS layers
...e.g. where and when to go fishing, when return to port, to which port?
...and account for diversity
...emerging outcomes (local knowledge to overall statistics)
A graphical user interface to help parameterize, run several simulations, compare and explore the outcomes, store in sql databases and replay simulations, etc.
...or all outputs in simple text files and use R post-processing routines to produce table & figures......
...alteratively to the ui, a large number of simulations/replicates can be generated using shell scripts in command line mode and making benefit of High Performance Computing facilities
...economic and ecological performance
of alternative spatial plans
...from the analysis of fishermen´s behaviour
...to understand the intertwined dynamics
...doesn´t stick to the law of the large numbers...
instead it is recognized that the dynamic of the system can be strongly non-linear therefore sometime leading to some bifurcations from divergent individual trajectories magnified by different initial conditions..So the final statistics/risk probabilities should reflect large uncertainty interval when e.g. two different modes are produced!
In our case ABM is to go beyond an *average* vessel approach by saying that all the uncertainties at the individual scales do not compensate to statistically leading to the median outcome....
Law of large numbers does not apply e.g. in the case of the Simpson paradox justifying that ABM + decision trees approach to capture non linearity is a way forward to avoid bias from high level of aggregation models...
from Tamás Székely Jr. and Kevin Burrage 2014
..a simulator is needed because likely no analytical solution when user react to the controls