ecoevosimulator 1.0 - Githubissues

sylvainschmitt / ecoevosimulator

Simulating eco-evolutionnary dynamics of virtual populations with topography and forest gap dynamics.

https://sylvainschmitt.github.io/ecoevosimulator

GNU General Public License v3.0

0 stars 1 forks source link

ecoevosimulator 1.0 #3

Open sylvainschmitt opened 3 years ago

sylvainschmitt commented 3 years ago

List of developments, bugs and fixes for ecoevosimulator 1.0

Major

[x] develop topography generators
[x] develop the forest gap dynamics generator
[x] tune the dispersion with a uniform kernel including autofecondation and mother and father dispersal
[x] define response variables (genotypic structure, environmental distance or mean viability or fitness, gene flow)

Minors

[x] abrupt tranistion in Paracou plots seems to matter for species divergence, maybe we should try to run TWI on generated topography
[x] cell size of Paracou is 3x3m, ideally the size should be of a typical reproducing adult tree in canopy

Bugs

Fixes

[x] spelling: ruggedness

mheuertz commented 3 years ago

Spelling: diamond-square
Create possibility to repeat a simulation on the same topography
Create a possibility to simulate a novel diamond-square topography with the same parameters otherwise

sylvainschmitt commented 3 years ago

Use a parameter for time-step length in years

mheuertz commented 3 years ago

As discussed this morning, use as time step an approximate number of years that would correspond to the number of years between recruitment and "full reproductive maturity". At this interval (for ex 30 years), the probability of mortality per grid cell is obtained from the "cumulated" per year probability of mortality, and dead individuals are replaced from the sapling pool available at the grid cell. Generations will be automatically overlapping since not all inds die at once; it will be approx 1 in 5 inds that has to be replaced (all inds contribute to sapling pool).

sylvainschmitt commented 3 years ago

use a null topo gradient to explore Dg_nci alone
uncertainty for wss score

sylvainschmitt commented 3 years ago

Niklas meeting

recruitment unclear => recruitment in canopy? => selection
reproduction => recruitment of seedlings ?
reproduction is gamete sampling
Explain simplicit vs reality between generators
what about link between NCI and TWI ?
background mortality unclear
beginning of chapter 2: inferring processes from real data
time-step = generation time (book: Matrix population models)
a_p => g_p
Recruitment image of euclidean distance
Precise initialization depending on biological meaning, e.g.:
- random genetic variation => local adaptation of standing genetic variation
- all individuals to the mean

sylvainschmitt commented 3 years ago

Bruno meeting

Interaction Htopo & Hnci => Make a grid Htopo:Hnci => sqrt(interaction)
Try to invert Rdispersion (1-Rdispersion) => Symmetric
Probably correctly handled within the lm function