In atredennick/sageAbundance: Fits population model for sagebrush based on remote sensing data.

When we simulate equilibrium sagebrush abundance (Figs. 4-5), we use point estimates of model parameters. Currently, we are only using the year random effects on the intercept for the temporally-explicit forecasts (Fig. 6). But maybe we should be using them in the equilibrium simulations too; the question is how to incorporate them. As a reminder, we fit the model with a temporal random effect on the intercept, such that

\begin{align} \beta_{0,t} \sim \text{normal}(\bar{\beta_{0}}, \sigma_{\beta_{0}}^2) \end{align}

So, when we use point estimates to simulate equilibrium abundance based on observed climate, do we use the mean of $\bar{\beta_{0}}$ for the intercept? Or, do we use the posterior mean of a randomly chosen $\beta_{0,t}$ at each time step for the intercept? Or, do we draw a random climate year ct, and then use the same year for the intercept, e.g., posterior mean of $\beta_{0,ct}$?

Currently, for Figs. 4-5, we are using the posterior mean of $\bar{\beta_{0}}$ as the intercept for all time steps of the simulation. My intuition tells me we want to use the posterior mean of a randomly chosen $\beta_{0,t}$ at each time step for the intercept. But, I want to get a consensus on this.

atredennick/sageAbundance documentation built on May 10, 2019, 2:11 p.m.