Lightweight update

Dependencies for zoo and plotly have been removed from the package. The removal of plotly in particular allows for a much less cluttered installation. A potential backward compatibility breaking change was also made in that now when an empty dataframe is created for MHW results (i.e. when no MHWs are detected) this now outputs a one row dataframe with the correctly named columns, but with a single NA value. This was done to accommodate changes in dplyr. Various other improvements and bug fixes were attended to. Please see the news file for a more complete list.

Important vignette updates and 2019 SST data

The old ERDDAP location for the NOAA OISST data was abandoned when v2.1 was released. The vignettes that use these data have now been updated to point to the new location. The packaged time series have also been updated to contain all v2.1 data from 1982 to 2019.

Bug fixes and vignette updates

The primary purpose of this version update is to account for changes to the data.table and dplyr packages. It addresses bugs that were introduced into the heatwaveR code due to changes outside of this package. This update also introduces some temporary changes to the vignettes about how to download and prepare NOAA OISST data.

Category improvements

In this release a few important bug fixes have been addressed, most of which were due to other important packages changing their internal structures. Up first the geom functions were tweaked to keep up with the advances made in the ggplot2 package. Some apparent memory issues caused between multicore operations and data.table were also smoothed out. The tidyverse package was removed from the dependencies, drastically reducing the install time of heatwaveR. Lastly, there were also a few minor problems in the category() function that have been addressed while also increasing its functionality. This now allows the function to be used for a more diverse range of implementations. This represents the most important upgrade to the new version.

Dependency free is the way to be

Several bug fixes and typo corrections exist in the background of this update while the main feature to celebrate is that the package no longer has any dependencies. It still imports packages of course. Some additional functionality was also added into the base functions with no threat to backwards compatibility. One may now specify how many decimal places the results of ts2clm() should be measured to. This is useful if one is finding climatologies for values with very high precision. It is also now possible to tell category() which season of the MHW it should tell you. The default is to provide the range of seasons, but now one may also choose the start, peak, or end season during the MHW.

heatwaveR: A central algorithm for the detection of heatwaves and cold-spells

This release features many bug fixes throughout. It also endeavored to bring the naming conventions for the arguments in the functions more in line with the Python version of the code. A large addition to this version has been the allowance of a second climatology/threshold in the calculation of heatwaves and cols-spells. One potential backwards compatibility breaking change has been that by default the var column is no longer output from ts2clm(). One must now provide the argument var = TRUE.

heatwaveR: A central algorithm for the detection of heatwaves and cold-spells

As the world continues to warm, we see not only a steady increase in mean temperatures, but an increase in the count and duration of extreme events, known as 'marine heatwaves' (MHWs). These events may decimate ecosystems and impact the health of fisheries. It is therefore necessary that a standard definition for these events be provided for researchers that allows for the comparison of events at a global scale. The first framework that allowed for the measurement and comparison of events globally was first outlined by Perkins et al. (2013) for atmospheric events. Based on this work, Hobday et al. (2016) then developed a definition for MHWs. A publication by Schlegel et al. (2017) then explored the concept of 'marine cold-spells' (MCSs).

The heatwaveR package was developed and released in order to provide one central repository for the definition and visualisation of atmospheric and marine heatwaves and cold-spells. It also contains the functionality to calculate and visualise the categories of events as outlined in Hobday et al. (2018). The heatwaveR package is a project-wide update to the RmarineHeatWaves package, which is itself a translation of the original Python code written by Eric C. J. Oliver. The heatwaveR package has brought the inputs and outputs of the R code more in line with the Python code while also introducing substantial speed improvements over the previous R version by deconstructing and modularising it. The slow portions of the code have now been implemented in C++. The modular nature of the code allows for the use of custom baselines and climatologies in the calculations of events. This means that as the techniques for the detection of events change and improve over time, this package will be able to grow with them.