Stepwise setup of PVA analyses

[ChloeRN]

Following up from our discussion today, we will continue expanding on the PVA for fishing cat. Currently, it employs the simplest model for population dynamics over 2-year time steps, which uses estimates of population growth rate (lambda) and does not include any age structure:

N[t+2] = N[t] * lambda[t]

Uncertainty in parameter estimates is not completely included yet.

We will expand the model through the following steps:

• [ ] Implement uncertainty on all parameters in simplest model
• [ ] Change to "Vital rates, no age structure" model, keep 2 year interval
• [ ] Reduce time interval to 1 year
• [ ] Implement "Vital rates, with age structure" model, 1 year interval only

These steps above are the bare minimum we need to do to meet the objectives. The additional steps below are optional/bonus:

• [ ] Optional: Collate data on female/kitten observations
• [ ] Optional: Partition recruitment rate into litter size (LS), breeding probability (pB), and early survival (S0)

In the comments below, I summarise additional thoughts, equations, and instructions for the different tasks that we discussed during the meeting.

[ChloeRN]

TASK 1: Implement uncertainty on all parameters in simplest model

In your current code, you are using a fixed number for starting population size, and a fixed number for the standard deviation of population growth rate. Instead, we want to use the distribution we obtained from the SCR model estimates. (This principle also applied to all the other implementations of the PVA; i.e. once we use vital rates, we have to specify the distributions for those accordingly).

[ChloeRN]

TASK 2: Change to "Vital rates, no age structure" model, keep 2 year interval

N[t+2] = N[t] S[t] + N[t] f[t] = N[t] * (S[t] + f[t])

This is at least in theory how it should work. However, this would imply that lambda = S + f, which is NOT the case according to @TaeKP's estimates. This means that one or several of the parameters estimated by the SCR model are not what we think they are. @TaeKP please revisit the documentation and implementation of your SCR model to check definitions of parameters and how they relate to lambda.

[TaeKP]

Thank you very much for your summary. These are really helpful to me, so I will proceed step by step and keep you posted on my progress.