birdtrends

The goal of birdtrends package is to provide a flexible toolkit to estimate population trends (based on annual indices and uncertainty or similar data inputs), predict future trends, and compare temporal trends over time.

Data Input Types

This package currently accommodates three types of input data:

1. Annual indices of relative abundance with CI estimates (i.e. index, upper and lower confidence interval)

2. Matrix of estimated index based on Bayesian modeled posterior draws (rows) for each year (columns) representing the full annual indices of relative abundance.

3. Matrix of posterior draws (rows) for each year (columns) based on the smoothed annual indices of relative abundance.

Estimating annual indices

Depending on the input data type, various methods are available to estimate the trend for a given time period. Note in the case of datatype 3 this already represents a modeled smooth generated from the original modeled relative population abundance.

For data input 1: we can fit a hierarchical Bayesian General Additive Model (HGAM), using annual indices and uncertainty values.

For data input 2: We can fit a GAM for each posterior draw

For data input 3: This data represents a smoothed output and can be used directly for trend assessment.

Estimate trend between given time points

We can use two time points to estimate a trend. The default method uses a geometric mean to estimate the average change in values over the time period. Alternatively we can assess the trend based on a linear regression between two points.

Predict trends into the future

Installation

You can install the development version of birdtrends from GitHub with:

# install.packages("devtools")
devtools::install_github("ninoxconsulting/birdtrends")

Example

This is a basic example which shows you how to solve a common problem:

#library(birdtrends)
## basic example code