In IMOS-AnimalTracking/remora: Rapid Extraction of Marine Observations for Roving Animals

## working when developing vignette
require(tidyverse, quietly = TRUE)
require(remora, quietly = TRUE)

---

The typical entry point to the remora package will be to apply quality control (QC) processes to acoustic telemetry detections data downloaded via the IMOS Australian Animal Acoustic Telemetry Database. The remora package uses the QC process described in Hoenner et al. (2018), with necessary modifications to accommodate the IMOS database's data and metadata formats. The QC workflow has been made as simple as possible, given the constraint of manually downloading the data and metadata prior to interaction with remora. This vignette describes the QC workflow.

Required data

Typically, users will download four .csv files via the IMOS Australian Animal Acoustic Telemetry Database:

1. Transmitter detections data, accessible via the detections page

2. Transmitter deployment metadata, accessible via the transmitters page

3. Receiver deployment metadata, accessible via the receiver deployments page

4. Animal measurements data, downloaded along with either detections or transmitter deployment metadata

Users will tell remora where to find these files by creating a files list as follows:

files <- list(det = "path_to/IMOS_detections.csv",
              rmeta = "path_to/IMOS_receiver_deployment_metadata.csv",
              tmeta = "path_to/IMOS_transmitter_deployment_metadata.csv",
              meas = "path_to/IMOS_animal_measurements.csv")

The detections data, receiver and transmitter metadata files are required to enable the full QC process, however because of some redundancy in variables among the detections data and metadata files, a partial QC can be conducted when one or both metadata files are missing. In these latter cases, the path for a missing metadata file is set to rmeta = NULL and/or tmeta = NULL; a warning will be issued when the QC is run. Similarly, if no animal measurements data are present the path is set to meas = NULL. The presence or absence of this latter file does not affect the QC process.

What if my tags are detected across many reciever array projects?

In situations where tagged animals may be detected across several receiver projects, we recommend downloading the entire IMOS receiver metadata file via the Web App. This file is relatively compact and the single download will save users time by not having to manually determine and download receiver metadata from many individual projects.

Usage of the runQC() function

The primary function for conducting the QC is the runQC() function. Let's start with an example tagging project included in the remora package by first creating a files list:

files <- list(det = system.file(file.path("test_data","IMOS_detections.csv"), package = "remora"),
              rmeta = system.file(file.path("test_data","IMOS_receiver_deployment_metadata.csv"),
                    package = "remora"),
              tmeta = system.file(file.path("test_data","IMOS_transmitter_deployment_metadata.csv"),
                    package = "remora"),
              meas = system.file(file.path("test_data","IMOS_animal_measurements.csv"),
                    package = "remora"))

We invoke the QC using runQC():

tag_qc <- runQC(files, .parallel = TRUE, .progress = FALSE)

tag_qc <- runQC(files, .parallel = TRUE, .progress = FALSE)
save(tag_qc, file = "R/tag_qc.rda")

where the argument .parallel = TRUE runs the QC in parallel across the available number of processor cores - 2. This can be modified using the .ncores argument. If .parallel = FALSE then the QC is run sequentially, which can take considerable time but may be more efficient for projects with a small number of transmitters. The .progress = FALSE argument turns off the QC progress indicator, it is set to TRUE (turned on) by default but set to FALSE here to keep the vignette tidy.

As part of the QC process, a logfile is generated by runQC to record common issues found in the data and/or metadata during the QC process. This logfile is saved to the working directory as QC_logfile.txt. Entries in the logfile provide some indication of a problem that the user may wish to explore and follow up as appropriate. An empty logfile indicates the QC process detected no common data issues, but this is not a guarantee that your data are problem-free! Simple issues, such as detection or receiver deployment latitudes in the northern hemisphere, are resolved automatically and corrected in the QC'd data output.

Accessing the QC'd data and metadata

The QC output object tag_qc is a nested tibble with each row corresponding to an individual animal using the transmitter_id, tag_id and tag_deployment_id as a unique identifier:

load("R/tag_qc.rda")
tag_qc

The QC list variable contains the QC'd detections data and metadata for each individual deployment. The utility function grabQC() provides a simple method to extract manageable segments of the QC output for subsequent review and analysis. For example, we can grab only the basic detection data with the QC flags and then filter the detections to retain only those flagged as 'valid' or 'likely valid':

dQC <- grabQC(tag_qc, what = "dQC", flag = c("valid","likely valid"))
dQC

or, to return all detection regardless of QC flag:

dQC <- grabQC(tag_qc, what = "dQC")
dQC

The dQC object can form the basis for subsequent annotation with environmental data (see vignette('extractEnv'); and vignette('extractMoor')) and analysis.

The metadata and animal measurements associated with QC'd detections can also be extracted in a compact form (without duplication of records):

tag_meta <- grabQC(tag_qc, what = "tag_meta")
rec_meta <- grabQC(tag_qc, what = "rec_meta")
meas <- grabQC(tag_qc, what = "meas")

If a user wishes to work with a simple data.frame containing all the QC output then the nested tibble can be converted to a flat file using tidyr::unnest() and dplyr::ungroup() as follows:

require(tidyverse)
qc <- tag_qc %>%
  unnest(cols = QC) %>%
  ungroup()

qc

Visualising the QC'd detections

The QC'd detections can be visualised as a map using the plotQC() function. Prior to remora version 0.8-0, plotQC generated a static map of detections categorised by QC flag. In version 0.8-0, plotQC is now interactive and renders by default (path = NULL) to a viewer window in RStudio. Other options to the path argument include path = "wb" which renders in the default web browser, if a valid file path is provided then the interactive map is saved as a self-contained .html file. Interactivity includes the ability to zoom in and out to reveal more or less detail on the spatial distribution of detections, and change which QC'd detections are displayed, map background layers, and colour scheme.

plotQC(tag_qc)

---

Vignette version 0.0.5 (09 Nov 2023)

IMOS-AnimalTracking/remora documentation built on Jan. 29, 2025, 4:38 p.m.