Main.m

% Main script for running the surface - subsurface model
% Here you can define which year you want to compute, define your
% param{kk}eters, plot output variables, evaluate model performance....
%
% Author: Baptiste Vandecrux (bava@byg.dtu.dk)
% ========================================================================
clearvars
close all
clc
addpath(genpath('.\lib'))
addpath(genpath('Input'),genpath('Output'))

% All param{kk}eters are defined in a csv files in Input folder, however it is
% possible to overwrite the default value by defining them again in the
% "param{kk}" struct hereunder.

station_list =   {'KAN_M'};
% station_list = PROMICE_dir; %
RCM_list = {'RACMO'};

param = cell(size(RCM_list));
for kk = 1:length(RCM_list)
    
    % %%%%%%%%%%%%%%%%%%%%%
    % High resolution grid (comment if not needed)
    NumLayer = 100;
    param{kk}.z_max = 50;
    param{kk}.dz_ice = param{kk}.z_max/NumLayer;
    param{kk}.verbose = 1;
    param{kk}.lim_new_lay = param{kk}.z_max/NumLayer/10;

    % param{kk}.ConductionModel = 0;      % if 1 does CONDUCTION ONLY
    % In the conduction model, the temperature profile is reseted every night
    % at 2am local time usingg thermistor string readings

    % Heterogeneous precolation from Marchenko et al. (2017)
    % this considers only redistribution of the water from the first layer into
    % the rest of the subsurface
    % An alternative is to go through the whole column and check whether piping
    % can occur at any depth
    param{kk}.hetero_percol = 0; % 1 = whole scheme on; 0 = standard percolation
    param{kk}.hetero_percol_p = 1; % binomial probability for a piping event to be initiated
                                % In Marchenko et al. (2017) this happens at
                                % every time step (probability 1)
    param{kk}.hetero_percol_frac = 1; % fraction of the available water that can be
                                % In Marchenko et al. (2017) all the available
                                % water goes into redistribution (frac = 1)
    param{kk}.hetero_percol_dist = 5; % characteristic distance until which
                                % preferential percolation operates
                                % When using uniform probability distribution 
                                % for the redistribution Marchenko et al. (2017)
                                % recommends between 4.5 and 6 m as cut-off value

    % result of a tuning of densification schemes
    %    param{kk}.a_dens = 30.25;
    %    param{kk}.b_dens = 0.7;

    param{kk}.year    =  0;
    % by defining param{kk}.year, the model will be run only for that melt year
    % (i.e. 1st. april to 1st april) however you still need to make sure that 
    % "rows"  is set so that the appropriate values will be read in the weather 
    % data. To run the model only from 1 to rows, just leave param{kk}.year=0.

    param{kk}.track_density = 0;
    param{kk}.avoid_runoff = 0; 
    param{kk}.retmip = 0;

    % PROMICE_dir = dir('../AWS_Processing/Output/PROMICE/');
    % PROMICE_dir = {PROMICE_dir.name};
    % PROMICE_dir(1:2) = [];
    % for i = 1:length(PROMICE_dir)
    %     PROMICE_dir{i} = PROMICE_dir{i}(1:(end-6));
    % end
    param{kk}.vis = 'off';
    %%%%%%%%%%%%%%%%%%%%%%%%%%%5
    model_version = RCM_list{kk};
        
for i =1:length(station_list)
    param{kk}.station =  station_list{i}; 

    switch param{kk}.station
        case 'KAN_M'
            param{kk}.InputAWSFile = 'Input/Weather data/data_KAN_M_combined_hour.txt';
    % ======== other stations =================

        otherwise
            disp('Missing data file for the requested station.');
    end

% When you add sites
% 1) in Main.m: define path of InputAWSFile
% 2) in  .m: Define path for density profile
% 3) Make sure all information is reported in Input/Constants/InfoStation.csv file

%%  ========= Run model with name tag of your choice =======================
 [RunName, c] = HHsubsurf(param{kk});

end
end