Example of Taylor diagram    (Click to enlarge) Taylor diagrams are usefull tools to help compare different statiscal datsets to one reference dataset.   See definition In my job context, it is used to help compare model output to observations or estimates. I program in Python a tool that computes and displays a Taylor diagram. Details of this program Sample gallery OCMIP results

This program is made of two following subprograms:

1. one computes comparison statistics needed for a Taylor diagram:
   
   
   • For each model, it computes the average, standard deviation, the correlation and RMS with data
   • • for a number of statistical components: total space time field, annual mean, zonal annual mean... (See complete list)
   • It also computes the same statistics for the median and the mean of all models
   • And also for an optionnal second data set
     
     
   • This program saves the results in a NetCDF file for use by the second subprogram or any other use.
   • It may also save the results in an ASCII plain text file. (See sample output)

   Here is a simple description and how to use it.
   You can also see the complete description of the program code.

2. the other plots a Taylor diagram, given the statistics in a NetCDF file.
   
   
3. • Actually, it plots a number of Taylor diagrams: one per statistical component
   • For each statistical component, all models are plot on the same diagram, along with the reference data set and an optionnal second data set
   • Each model is referenced by a one-or-two-letter code on the diagram (see example below)
   • The median and mean of all models and a second data set, if available, are also plot along with the models

   Here is a simple description and how to use it.
   You can also see the complete description of the program code.