Part_2_Ex_1.m

%% Launching the ensemble run
% First, we need to load the configuration file. Let's print the general settings 
% in the configuration to ensure we have it right.

clc, clear all;
load("Config_base.mat");
% config.run.todate = '2007-12-31';
disp(config.run)
%% 
% We also add the path where all the functions are located to Matlab.

addpath(genpath('./Codes/'));
%% 
% Finally, we will set a command to launch the debugger if an error occurs (this 
% should not happen).

 dbstop if error;
%% The ensemble run
% We will now launch the ensemble run. To keep it simple, in this example the 
% 8 model runs will here be run sequentially. In an real DA experiment, this process 
% would be parallelized in different threads in order to speed up the computation.

nE = config.ensembleOptions.nE; % Number of ensemble members.

% Perform model run.
for i=1:nE
%% 
% First, we need to replace the keyword %n for the number of the ensemble member 
% we are processing. Let us also print one of them to make sure the replacement 
% was done correctly.

    config_ens = config;
    % Assign proper directories to each ensemble member:
    config_ens.input.pars                   = strrep(config_ens.input.pars,'%n',num2str(i));
    config_ens.input.forcing.prcp           = strrep(config_ens.input.forcing.prcp,'%n',num2str(i));
    config_ens.states.directory             = strrep(config_ens.states.directory,'%n',num2str(i));
    disp(config_ens.states.directory)
%% 
% We are now ready to launch the model run for this ensemble member:

    % Run ensemble member.
    RUN_W3RA(config_ens);
end
%% Result extraction
% Now that all the ensembles have been run, we can compute the ensemble-averaged 
% results

F02_ensemble_extract_state_results_W3RA(config);
%% Result visualization
% Finally, let us visualize the results of the ensemble run. For that, we can 
% use an existing function.

F_visualizeEx1(config)
%% Reflection questions:
%% 
% # What can be the benefits and drawbacks of running a distributed model compared 
% to a lumped model?
%%