It takes as an argument an object of class Resources, called
arguments which is a container for all arguments and datasets that this
variable might need for its computation. Since this method is called from the
dataset method compute_variables(), it is important to pass all
required objects into the resources argument of
compute_variables(). If the variable needs attributes of another
dataset, this dataset should be stored in resources under its dataset
name.
The implementation of the compute() method can assume that all required variables are already computed and accessible by the dataset method get_attribute(), if they are listed in the dependencies() method (see Section 7.4.2).
The compute() method returns a numpy array of the size of number of elements in the ``owner'' dataset.