Extra example scripts

In the extras/ directory, there are a few bash example scripts that show how one might run several processes in parallel (the *-all.bash files). The scripts contains nested loops over the epochs, the temperature and precipitation scenarios. The wait command waits for the child processes (that is, all the backgrounded processes) to finish before it continues, and putting in one of the loops limits the number of simultaneous processes. At the end of the script, it prevents the bash script from exiting until all backgrounded processes have finished (this way, one could, for example, time the bash script for an estimate of the total duration: time percentiles-all.bash).

The percentiles-all.bash also uses a small extra module, kcs.change_perc.runall, which is convenient when calculating all scenarios from a steering table. The default kcs.change_perc functionality requires two runs, one for CMIP data and one for the model of interest, each with different options set, such as --period, --run-changes and --no-matching in our case. The runall module takes care of that, and deduce the relevant periods from the steering table. Variables are deduced from the input files. It assumes that the data for the model of interest is already concatenated. The script does split things up with a nested loop into epochs, variables and seasons, so that processes can be run simultaneously.

The resample-all.bash is very similar to the percentiles-all.bash script, in that it loops through scenarios and epochs, running processes in parallel. The resample-and-plot.bash script goes one step further, as it will take the outputs from the resampling processes, and creates figures for the CMIP distribution with the resampled runs overplotted, for variations of epoch, variable and season.