Parameters, priors, and toggles

Lucas da Cunha Godoy

2025-06-26

The fit_drm function is designed to be flexible, allowing different model features to be enabled or disabled through a set of toggles. This vignette provides a comprehensive guide to the parameters used in the DRM, their meaning, the priors they are assigned, and the specific toggles required to activate them.

DRM model parameters and priors

The table below details each parameter in the DRM. The “Toggle” column indicates the condition needed to include a parameter in the model. If a parameter is not activated by its toggle, it is excluded from the model, and its prior is not used.

Parameter	Code	Meaning	Toggle	Priors	Default hyperparameters
$\beta_t$	beta_t	Reg. coef. for proba. of absence	N/A	$\mathrm{N}(m_t, s_t)$	$m_t = 0, s_t = 1$
$\beta_r$	beta_r	Reg. coef. for recruitment	N/A	$\mathrm{N}(m_t, s_t)$	$m_r = 0, s_r = 1$
$\phi$	phi	Dispersion parameter	N/A	$\mathrm{Gamma}(a_p, b_p)$	$a_p = 2, b_p = 1$
$\beta_s$	beta_s	Reg. coef. for survival	est_surv	$\mathrm{N}(m_s, s_s)$	$m_s = 0, s_s = 1$
$\xi$	xi	$() = + ()	rho_mu	$- \mathrm{LN}(m_{lxi}, s_{lxi})$	$m_{lxi} = 0, s_{lxi} = 1$
$\zeta$	zeta	Prob. of remaining in the same patch	movement	$\mathrm{Beta}(a_z, b_z)$	$a_z = .5, b_z = .5$
$\alpha$	alpha	AR(1) temporal correlation	ar_re $\neq$"none"	$\mathrm{Beta}(a_a, b_a)$	$a_a = .5, b_a = .5$
$_t	sigma_t	AR(1) conditional SD	ar_re $\neq$"none"	$\mathrm{LN}(m_{lst}, s_{lst})$	$m_{lst} = -2, s_{lst} = .25$
$z_t	z_t	AR(1) random effects	ar_re $\neq$"none"	$z_t \sim AR(1)$	N/A
$_i	sigma_i	Patch level iid random effect SD	iid_re $\neq$"none"	$\mathrm{LN}(m_{lsi}, s_{lsi})$	$m_{lsi} = -2, s_{lsi} = .25$
$z_i	z_i	Patch level iid random effects	iid_re $\neq$"none"	$z_i \sim \mathrm{N}(0, \sigma^2_i)$	N/A
$_s	sigma_s	Patch level ICAR random effect SD	sp_re $\neq$"none"	$\mathrm{LN}(m_{lss}, s_{lss})$	$m_{lss} = -2, s_{lss} = .25$
$z_s	z_s	Patch level ICAR random effects	sp_re $\neq$"none"	$z_s \sim \mathrm{N}(0, \sigma^2_i)$	N/A

SDM

TBD

References