Dear Bruno,

 

Thanks for your interest. I know very little about amphipod ecology, so I cannot comment on whether or not MGET’s connectivity tool would provide a realistic model for them. But I can tell you a bit about the tool and maybe that will help you decide.

 

The tool is essentially a Eulerian advection/diffusion algorithm with some biological mechanisms added on. You define source patches (groups of raster cells) where “larvae” are released and sink patches where they can settle. Normally these patches are the same but they are allowed to be different. You can configure how many are released from each patch. You specify a date for the larvae to be released and a duration for which they are allowed to drift (the PLD).

 

The simulator moves larvae  around according to ocean currents. There are several sources of currents to choose from. For the Med, you might choose the Aviso 1/8 degree currents, on the basis that Aviso has done a lot of work to validate the currents there (as I understand it). When larvae drift over a sink patch, they settle at a certain rate. The model is mass-balanced; settled larvae are removed from the water column. You must provide the rate at which they settle. Choosing a low settlement rate will allow more to drift over sink patches to downstream areas.

 

Mortality is optional and is applied after the advection/diffusion stage of the simulation is complete. The simulator stores enough information that mortality can be applied as a simple matrix operation without having to rerun the main algorithm, allowing you to test different mortality rates cheaply.

 

We published a sensitivity analysis [1] that examined the influence of different parameters. Not surprisingly, the quantity of larvae released, PLD, and mortality were all important. Thus it seems problematic to model connectivity without knowing these parameters. But you might be able to work around it like this:

 

1.    You probably don’t need to estimate connectivity in absolute terms, e.g. number of individuals per connection. You’re probably interested in relative connectivity between habitat patches. You probably have decent maps of habitat patches (e.g. seagrass beds). You could start by assuming that the quantity of larvae released per km of seagrass is the same.

 

2.    Then conduct the simulation over several different PLDs that might bound the range of what is realistic. This would be the first dimension of a sensitivity analysis.

 

3.    If there is seasonal or interannual variability, you will also need to conduct simulations over a range of dates. This would be the second dimension of the sensitivity analysis. Hopefully the combination of date and PLD would not produce too many simulations.

 

4.    Running those will take time. After they are done, you can relatively quickly test different mortality rates. This would be the third dimension of the sensitivity analysis, but it should be quick.

 

Because 2 and 3 will take a lot of time, you should definitely run the tool through to the end at least one time, produce some results, and make sure you understand them and that they represent the connectivity information you’re looking for.

 

There are also the questions of settlement rate and precompetency period. In my experience, these parameters are less important. Ideally you would add them as additional dimensions but your time may be limited. I would prioritize them lower. My colleague Eric Treml at University of Melbourne might be able to advise you on values you could choose.

 

Best,

Jason

 

[1] Treml EA, Roberts JJ, Chao Y, Halpin PN, Possingham HP, Riginos C (2012) Reproductive output and duration of the pelagic larval stage determine seascape-wide connectivity of marine populations. Integrative and Comparative Biology 52: 525-537

 

 

Dear all,

I’m writing this email for a possible clarification about the use of MGET tools. I’m currently work on the Posidonia oceanica/amphipods system and I’d like to relate specific patterns with the potential connectivity of a series of seagrass meadows all around the Mediterranean Sea. Differently from other, amphipod lack a real pelagic larval form, so that I miss some information about the PLD, as well as mortality and settlement.

My question is: it is possible to use the tool without considering these specific parameters? My idea was to use MGET by ignoring these parameters, but I really don’t know if it is possible from a computational point of view. Before start the simulation (which probably require days if not weeks), I’d like to know the feasability of using the tool this way.
Obviously, any other suggestion is welcome.

Thanks in advance for taking the time to give an answer.

 

Cheers.

 

Bruno Bellisario, PhD

brunobellisario.weebly.com

Department of Ecological nad Biological Sciences

University of Viterbo

Italy