Intro

Starting from SDMtune version 0.2.0 the new SWD() object bundles together presence and absence/background data. This choice is necessary in order to enable new methods that use presence/absence data and to use spatial cross validation. We tried to reduce to the minimum the change in code that the users have to apply and make the transition to the new version as smooth as possible.

The user has two possibilities:

  • adapt the code following the articles and rerun the analysis;
  • convert objects created with the previous version into the new format.

Here we explain how to convert the objects into the new format. If you get an error like:

#> This object was created using SDMtune v <= 0.1.1 and is now deprecated.
#> Check the article "Deprecated objects" in the package website to see how to convert this object into the new format.

you are in the right place.

SWD object

The new SWD() object allows to store together the presence and the absence/background locations and has a new slot @pa that contains a vector with ones for presence and zeros for absence/background locations. There are two possible scenarios:

  • you need to create an SWD() object to train a model;
  • you need to create an SWD() object used as test datasets for presence-only methods.

SWD object to train a model

For the following code we assume that you have two old SWD() objects, one called presence_swd and the other absence_swd:

train <- old2NewSWD(presence_swd, absence_swd)

It is important that the first argument is the one with the presence locations and the second is the one with the absence/background locations.

SWD object for test

For the following code we assume that you have two old SWD() objects called test_swd and absence_swd. Is important that the test absence object is the same used to train the model:

test <- old2NewSWD(test_swd, absence_swd)

SDMmodel and SDMmodelCV objects

The main change in the new SDMmodel() object is that there is only one slot called @data that stores the new SWD() object, instead of the two slots @p and @a. For the following code we assume that you have an old SDMmodel() called my_model:

Same for the SDMmodelCV, the main change is that the slot @models now contains a list of models that have the new SDMmodel() structure.

SDMtune object

The only change for the new SDMtune() object is that the slot @models contains a list of models of class SDMmodel() or SDMmodelCV(). You can use the following code: