fit_agTrend_ssl_mgcv: Site-level fitting function

In dsjohnson/agTrendTMB: Interpolate Survey Counts For Site-Level Aggregating

Description

Fit separate discrete-normal distribution models to an Steller sea lion counts at a single site

Usage

fit_agTrend_ssl_mgcv(.x, obl.corr = FALSE, debug = FALSE)

Arguments

.x	Data frame for an individual site. Must have been passed through prep_for_fit first!
obl.corr	Logical. Should correction be made for oblique photos.
debug	Logical. Step into fitting function for interactive evaluation
model	Model for interpolation: 'const', 'lin', or 'gp' (see Details)

Details

There are 3 models which can be used for within site interpolation:

• const This interpolates using a simple mean and variance for the site

• lin A linear fit with time is used for interpolation

• gp i.e., Gaussian Process. This model used a time trend plus a random walk of order 2 (RW(2)) to interpolate counts. Further, the RW(2) process can be further restrained to minimize 'wiggliness' of the curve when necessary. This prevents overfitting of the curve.

Value

A list with the following elements:

• summary Parameter and abundance estimate information?

• raw_data Original data plus oblique photo correction estimates

• q_output quality control diagnostic data

dsjohnson/agTrendTMB documentation built on Jan. 1, 2021, 12:07 a.m.