optimized_thickness.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### This notebook allows Glen's A to vary to optimize modeled ice thickness by minimizing the quadratic error between observed and modeled ice thickness (following height change corrections and in situ mass balance inversion)\n",
    "\n",
    "### named MONSTER by M. Zeuner (Mega Organized Numerical Script Transformation Executing Rapidly)\n",
    "\n",
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import xarray as xr\n",
    "import geopandas as gpd\n",
    "import matplotlib.pyplot as plt\n",
    "from oggm import utils, workflow, tasks, graphics\n",
    "\n",
    "import os\n",
    "import pandas as pd\n",
    "import salem\n",
    "import oggm\n",
    "from oggm.workflow import execute_entity_task, gis_prepro_tasks, climate_tasks\n",
    "\n",
    "import oggm.cfg as cfg # could that cause problems?\n",
    "import gdal\n",
    "from PIL import Image as imagery\n",
    "import datetime"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Select the glacier by setting num to:\n",
    "\n",
    "- 0 for Kokanee\n",
    "- 1 for Haig\n",
    "- 2 for Conrad\n",
    "- 3 for W. Washmawapta (no DEM correction available)\n",
    "- 4 for Illecillewaet (not supported in this version of the script due to shapefile reasons)\n",
    "- 5 for Nordic (not supported in this version of the script due to shapefile reasons)\n",
    "- 6 for Zillmer\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "You selected:  Conrad\n"
     ]
    }
   ],
   "source": [
    "# Here the num variable is set:\n",
    "num=2\n",
    "\n",
    "# A list with all the glacier names\n",
    "glacier = ['Kokanee', 'Haig', 'Conrad', 'Washmawapta', 'Illecillewaet', 'Nordic', 'Zillmer']\n",
    "\n",
    "# Just confirm that the correct one is selected for this run of the script\n",
    "print('You selected: ' , glacier[num])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### File names settings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "# list with the names in the Randolf Glacier invetory (according to the glacier names above)\n",
    "rgi_list = ['RGI60-02.00147', 'RGI60-02.01857', 'RGI60-02.02171', 'RGI60-02.03411', 'RGI60-02.03688',\n",
    "            'RGI60-02.04264', 'RGI60-02.07780']\n",
    "\n",
    "# list with the names Lidar data (according to the glacier names above)\n",
    "dem = ['kokanee/20160913_kokanee_dem1_clip_slave.tif', 'haig/20150912_haig_dem_master_clip.tif', 'conrad/20160912_conrad_dem1_clip_slave.tif', 'not_assigned', \n",
    "       'illecillewaet/170917_illecillewaet_dem1_clip_slave.tif', 'nordic/20170927_nordic_dem1_clip_slave.tif', 'zillmer/160914_zillmer_dem2.tif']\n",
    "\n",
    "#products_directory (where the graphs and some numbers are saved for later usage)\n",
    "products= '/home/pelto/oggm_runs/products_opt/' + glacier[num] + '/'\n",
    "\n",
    "sw = 201.  #smoothing window meters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A function to run the OGGM using mass-balance correction and (if wished) the Lidar-DEM \n",
    "\n",
    "This is also a function that runs the OGGM on the selected glacier and returns a thickness map for it. It uses the mass-balance correction.\n",
    "\n",
    "In difference to the function above, this version does not use the SRTM DEM to run it but the more precise Lidar DEM which yields results closer to reality.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "def new_mb_grad_corr_dem(num, dem, glacier, name, parameter, best, use_dem_corr):\n",
    "    \"\"\"\n",
    "    :param num: number of selected glacier\n",
    "    :param dem: list with dem file names\n",
    "    :param glacier: the list with the glacier names\n",
    "    :param name: name of the parameter to be modified\n",
    "    :param parameter: value of the parameter to be modified\n",
    "    :param best: if TRUE, an extra save with prefix 'best' will be created\n",
    "    :param use_dem_correction: if TRUE, a Lidar DEM correction will be used\n",
    "    :type num: int\n",
    "    :type dem: str list\n",
    "    :type glacier: str list\n",
    "    :type name: str \n",
    "    :type parameter: float\n",
    "    :type best: boolean\n",
    "    :type use_dem_correction: boolean\n",
    "    :returns: distributed thickness\n",
    "    :rtype: 2D float array\n",
    "    \"\"\"\n",
    "    \n",
    "    # Initialize OGGM and set up the run parameters --> Read the configuration file containing the run’s parameters.\n",
    "    cfg.initialize(logging_level='WORKFLOW')\n",
    "\n",
    "    #num determines which gdir and gradient to use, refer to list below to select desired site\n",
    "    grad = [6.84, 9.8, 7.11, 7.95, 10.84, 6.01, 7.92] # mm w.e. m-1\n",
    "#     grad = 6.43 #regional\n",
    "#     grad[num] = grad\n",
    "\n",
    "    # Get the RGI glaciers for the run.\n",
    "    rgi_list = ['RGI60-02.00147', 'RGI60-02.01857', 'RGI60-02.02171', 'RGI60-02.03411', 'RGI60-02.03688', 'RGI60-02.04264', 'RGI60-02.07780']\n",
    "    temporarily = [rgi_list[num]]\n",
    "    rgidf = utils.get_rgi_glacier_entities(temporarily)#[rgi_list[num]]) !!!!!!!!!!!!!!!!!!!!!!!!!!!!\n",
    "    \n",
    "    ############################################################# here the DEM correction happens #########################################################\n",
    "    \n",
    "    if(use_dem_corr==True):\n",
    "    \n",
    "        custom_dem_path = '/home/pelto/Desktop/lidar_cbt_analysis/' + dem[num]\n",
    "        print(custom_dem_path)\n",
    "\n",
    "        WORKING_DIR2 = utils.gettempdir(glacier[num] + '_lidar_dem') # Let's make a working directory for this DEM \n",
    "        utils.mkdir(WORKING_DIR2, reset=True)\n",
    "        cfg.initialize(logging_level='WORKFLOW')\n",
    "        cfg.PATHS['working_dir'] = WORKING_DIR2\n",
    "\n",
    "        cfg.PATHS['dem_file'] = custom_dem_path\n",
    "        rgidf['DEM_SOURCE'] = 'USER'\n",
    "\n",
    "\n",
    "        cfg.PARAMS['smooth_window'] = sw\n",
    "        #cfg.PARAMS['grid_dx_method'] = 'fixed'\n",
    "        #cfg.PARAMS['fixed_dx'] = dx\n",
    "        if glacier[num] == 'Nordic' or glacier[num] == 'Washmawapta':\n",
    "            cfg.PARAMS['use_intersects'] = False\n",
    "#         gdirs = workflow.init_glacier_regions(rgidf)\n",
    "        else:\n",
    "            cfg.PARAMS['use_intersects'] = True\n",
    "        cfg.PARAMS['use_rgi_area'] = False\n",
    "        \n",
    "    else:\n",
    "        if glacier[num] == 'Nordic' or glacier[num] == 'Washmawapta':\n",
    "            cfg.PARAMS['use_intersects'] = False\n",
    "        else:\n",
    "            cfg.PARAMS['use_intersects'] = True\n",
    "        cfg.PARAMS['border'] = 10\n",
    "        cfg.PATHS['working_dir'] = utils.get_temp_dir('test_thick' + glacier[num])\n",
    "        cfg.PARAMS['use_rgi_area'] = False\n",
    "        cfg.PARAMS[name] = parameter\n",
    "        cfg.PARAMS['smooth_window'] = sw\n",
    "        \n",
    "    ###########################################################################################################################\n",
    "    \n",
    "    cfg.PARAMS[name] = parameter\n",
    "    cfg.PARAMS['border'] = 10\n",
    "    \n",
    "    gdirs = workflow.init_glacier_regions(rgidf)\n",
    "    workflow.gis_prepro_tasks(gdirs)\n",
    "    workflow.climate_tasks(gdirs)\n",
    "    workflow.inversion_tasks(gdirs)\n",
    "    workflow.execute_entity_task(tasks.distribute_thickness_per_altitude, gdirs);\n",
    "\n",
    "    gdir = gdirs[0]\n",
    "    tasks.init_present_time_glacier(gdir) ##This updates the mode_flowlines file and creates a stand-alone numerical glacier ready to run.\n",
    "\n",
    "    # Print the standard MB gradient from OGGM default\n",
    "    from oggm.core.massbalance import ConstantMassBalance, MultipleFlowlineMassBalance\n",
    "    # gdir = gdirs[num]\n",
    "\n",
    "    mb_mod = MultipleFlowlineMassBalance(gdir, use_inversion_flowlines=True, mb_model_class=ConstantMassBalance)\n",
    "    z, w, mb_on_z = mb_mod.get_annual_mb_on_flowlines()\n",
    "    mb_on_z *=  cfg.PARAMS['ice_density'] * cfg.SEC_IN_YEAR\n",
    "\n",
    "    # grad_line = np.polyfit(mb_on_z,z,2)\n",
    "    from scipy import stats \n",
    "    slope, intercept, r_value, p_value, std_err = stats.linregress(mb_on_z, z)\n",
    "    print(\"slope: %f    intercept: %f\" % (slope, intercept), \"  R-squared: %f\" % r_value**2)\n",
    "\n",
    "    # See the result of the inversion\n",
    "    workflow.execute_entity_task(tasks.distribute_thickness_per_altitude, gdirs);\n",
    "\n",
    "    ds_default = xr.open_dataset(gdir.get_filepath('gridded_data')).load()\n",
    "    ds_default.close() # I have the strange feeling that xarray sometimes won't re-read overwitten files (and we will overwrite it later)\n",
    "\n",
    "    # Let's apply a linear MB instead\n",
    "    from oggm.core.climate import apparent_mb_from_linear_mb\n",
    "\n",
    "\n",
    "    workflow.execute_entity_task(tasks.apparent_mb_from_linear_mb, gdir, mb_gradient=grad[num]);\n",
    "\n",
    "    from oggm.core.massbalance import LinearMassBalance\n",
    "\n",
    "    params = gdir.read_pickle('linear_mb_params')\n",
    "    print(params)\n",
    "\n",
    "    l_mb_mod = LinearMassBalance(params['ela_h'], grad=params['grad'])\n",
    "    l_mb_on_z = l_mb_mod.get_annual_mb(z) * cfg.PARAMS['ice_density'] * cfg.SEC_IN_YEAR\n",
    "\n",
    "    slope, intercept, r_value, p_value, std_err = stats.linregress(mb_on_z, z)\n",
    "    print(\"slope default: %f    intercept: %f\" % (slope, intercept), \"  R-squared: %f\" % r_value**2)\n",
    "    slope, intercept, r_value, p_value, std_err = stats.linregress(l_mb_on_z, z)\n",
    "    print(\"slope: %f    intercept: %f\" % (slope, intercept), \"  R-squared: %f\" % r_value**2)\n",
    "    \n",
    "    workflow.inversion_tasks(gdirs)\n",
    "    workflow.execute_entity_task(tasks.distribute_thickness_per_altitude, gdirs);\n",
    "    tasks.init_present_time_glacier(gdir) ##trying to update the model flowline\n",
    "\n",
    "    ds_new = xr.open_dataset(gdir.get_filepath('gridded_data')).load()\n",
    "    ds_new.close() \n",
    "   \n",
    "\n",
    "    ########### gather more detailed information about the glacier ##############\n",
    "    df  = utils.compile_glacier_statistics(gdirs, inversion_only=True)\n",
    "    volume=df['inv_volume_km3']\n",
    "    \n",
    "    mfl=gdir.read_pickle('model_flowlines')\n",
    "    thickness= mfl[-1].surface_h - mfl[-1].bed_h\n",
    "    thickness=thickness[thickness>0].mean()\n",
    "    \n",
    "    ############## write it into txt file: #################\n",
    "    \n",
    "    if(best):\n",
    "        file= open(products+'best_dem_corr_run_of_'+glacier[num]+'_with_modified_' + name+'.txt', mode='w')\n",
    "    else:\n",
    "        file= open(products+'dem_corr_run_of_'+glacier[num]+'_with_modified_' + name+'.txt', mode='w')\n",
    "    \n",
    "    file.writelines(['Run of the DEM corrected model on ' + glacier[num]+' on ', str(datetime.datetime.now()), '\\n'])\n",
    "    file.write('parameter modified: '+name+'\\n')\n",
    "    file.write('parameter value: '+str(parameter)+'\\n')\n",
    "    file.write('glacier total volume: '+str(volume)+' km^3 \\n \\n \\n')\n",
    "    file.write('mean thickness value: '+str(thickness)+' m \\n')\n",
    "    file.close()\n",
    "    ##########################################################\n",
    "    \n",
    "    return ds_new.distributed_thickness"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Comparing DEM-corrected model with non-DEM-corrected model:\n",
    "\n",
    "- first we run the 'new_mb_grad' function\n",
    "- then we run the 'new_mb_grad_corr_dem' function\n",
    "- and then we subtract the thickness maps and can see where the usage of Lidar data actually impacted the outcome of the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:40:42: oggm.cfg: Using configuration file: /home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/oggm/params.cfg\n",
      "2020-03-31 08:40:43: oggm.workflow: Execute entity task glacier_masks on 1 glaciers\n",
      "2020-03-31 08:40:43: oggm.workflow: Multiprocessing: using all available processors (N=8)\n",
      "2020-03-31 08:40:43: oggm.workflow: Execute entity task compute_centerlines on 1 glaciers\n",
      "2020-03-31 08:40:43: oggm.workflow: Execute entity task initialize_flowlines on 1 glaciers\n",
      "2020-03-31 08:40:43: oggm.workflow: Execute entity task compute_downstream_line on 1 glaciers\n",
      "2020-03-31 08:40:43: oggm.workflow: Execute entity task compute_downstream_bedshape on 1 glaciers\n",
      "2020-03-31 08:40:43: oggm.workflow: Execute entity task catchment_area on 1 glaciers\n",
      "2020-03-31 08:40:45: oggm.workflow: Execute entity task catchment_intersections on 1 glaciers\n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/geopandas/io/file.py:108: FionaDeprecationWarning: Use fiona.Env() instead.\n",
      "  with fiona.drivers():\n",
      "2020-03-31 08:40:45: oggm.workflow: Execute entity task catchment_width_geom on 1 glaciers\n",
      "2020-03-31 08:40:45: oggm.workflow: Execute entity task catchment_width_correction on 1 glaciers\n",
      "2020-03-31 08:40:45: oggm.workflow: Execute entity task process_cru_data on 1 glaciers\n",
      "2020-03-31 08:40:46: oggm.workflow: Execute entity task local_t_star on 1 glaciers\n",
      "2020-03-31 08:40:46: oggm.workflow: Execute entity task mu_star_calibration on 1 glaciers\n",
      "2020-03-31 08:40:46: oggm.workflow: Execute entity task prepare_for_inversion on 1 glaciers\n",
      "2020-03-31 08:40:46: oggm.workflow: Execute entity task mass_conservation_inversion on 1 glaciers\n",
      "2020-03-31 08:40:46: oggm.workflow: Execute entity task filter_inversion_output on 1 glaciers\n",
      "2020-03-31 08:40:46: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n",
      "2020-03-31 08:40:47: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "slope: 0.237854    intercept: 2621.998664   R-squared: 0.989877\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:40:47: oggm.workflow: Execute entity task apparent_mb_from_linear_mb on 1 glaciers\n",
      "2020-03-31 08:40:47: oggm.workflow: Execute entity task prepare_for_inversion on 1 glaciers\n",
      "2020-03-31 08:40:47: oggm.workflow: Execute entity task mass_conservation_inversion on 1 glaciers\n",
      "2020-03-31 08:40:47: oggm.workflow: Execute entity task filter_inversion_output on 1 glaciers\n",
      "2020-03-31 08:40:47: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'ela_h': 2545.680813745039, 'grad': 7.11}\n",
      "slope default: 0.237854    intercept: 2621.998664   R-squared: 0.989877\n",
      "slope: 0.140647    intercept: 2545.680814   R-squared: 1.000000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:40:48: oggm.workflow: Execute entity task glacier_statistics on 1 glaciers\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# run the model without DEM correction\n",
    "new_mb_gradient= new_mb_grad_corr_dem(num, dem, glacier, 'ice_density', 900, False, False)\n",
    "\n",
    "# plot it as a map\n",
    "new_mb_gradient.plot()\n",
    "plt.title(glacier[num]+' standard settings thickness map')\n",
    "# and save it\n",
    "plt.savefig('/home/pelto/Desktop/ice_thick/products'+'standard_map_'+ glacier[num]+'.png', dpi=100)\n",
    "\n",
    "# replace 'nan' by  0\n",
    "new_mb_gradient=new_mb_gradient.fillna(0)\n",
    "\n",
    "if glacier[num]=='Washmawapta':\n",
    "    # make a copy for comparison reasons later on\n",
    "    new_mb_gradient_corr_dem=new_mb_gradient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "# custom_dem_path = '/home/pelto/Desktop/lidar_cbt_analysis/' + glacier[num] + '/' + dem[num]\n",
    "# custom_dem_path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:40:48: oggm.cfg: Using configuration file: /home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/oggm/params.cfg\n",
      "2020-03-31 08:40:49: oggm.cfg: Using configuration file: /home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/oggm/params.cfg\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/pelto/Desktop/lidar_cbt_analysis/conrad/20160912_conrad_dem1_clip_slave.tif\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:40:54: oggm.workflow: Execute entity task define_glacier_region on 1 glaciers\n",
      "2020-03-31 08:40:54: oggm.workflow: Multiprocessing: using all available processors (N=8)\n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/geopandas/io/file.py:108: FionaDeprecationWarning: Use fiona.Env() instead.\n",
      "  with fiona.drivers():\n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/geopandas/io/file.py:108: FionaDeprecationWarning: Use fiona.Env() instead.\n",
      "  with fiona.drivers():\n",
      "2020-03-31 08:40:56: oggm.workflow: Execute entity task glacier_masks on 1 glaciers\n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/oggm/core/gis.py:505: RuntimeWarning: invalid value encountered in less_equal\n",
      "  topo[topo <= -999.] = np.NaN\n",
      "2020-03-31 08:40:56: oggm.workflow: Execute entity task compute_centerlines on 1 glaciers\n",
      "2020-03-31 08:40:56: oggm.workflow: Execute entity task initialize_flowlines on 1 glaciers\n",
      "2020-03-31 08:40:57: oggm.workflow: Execute entity task compute_downstream_line on 1 glaciers\n",
      "2020-03-31 08:40:57: oggm.workflow: Execute entity task compute_downstream_bedshape on 1 glaciers\n",
      "2020-03-31 08:40:57: oggm.workflow: Execute entity task catchment_area on 1 glaciers\n",
      "2020-03-31 08:40:58: oggm.workflow: Execute entity task catchment_intersections on 1 glaciers\n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/geopandas/io/file.py:108: FionaDeprecationWarning: Use fiona.Env() instead.\n",
      "  with fiona.drivers():\n",
      "2020-03-31 08:40:58: oggm.workflow: Execute entity task catchment_width_geom on 1 glaciers\n",
      "2020-03-31 08:40:58: oggm.workflow: Execute entity task catchment_width_correction on 1 glaciers\n",
      "2020-03-31 08:40:58: oggm.workflow: Execute entity task process_cru_data on 1 glaciers\n",
      "2020-03-31 08:40:59: oggm.workflow: Execute entity task local_t_star on 1 glaciers\n",
      "2020-03-31 08:40:59: oggm.workflow: Execute entity task mu_star_calibration on 1 glaciers\n",
      "2020-03-31 08:40:59: oggm.workflow: Execute entity task prepare_for_inversion on 1 glaciers\n",
      "2020-03-31 08:40:59: oggm.workflow: Execute entity task mass_conservation_inversion on 1 glaciers\n",
      "2020-03-31 08:40:59: oggm.workflow: Execute entity task filter_inversion_output on 1 glaciers\n",
      "2020-03-31 08:40:59: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n",
      "2020-03-31 08:41:00: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "slope: 0.237523    intercept: 2616.883086   R-squared: 0.991220\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:41:00: oggm.workflow: Execute entity task apparent_mb_from_linear_mb on 1 glaciers\n",
      "2020-03-31 08:41:00: oggm.workflow: Execute entity task prepare_for_inversion on 1 glaciers\n",
      "2020-03-31 08:41:00: oggm.workflow: Execute entity task mass_conservation_inversion on 1 glaciers\n",
      "2020-03-31 08:41:01: oggm.workflow: Execute entity task filter_inversion_output on 1 glaciers\n",
      "2020-03-31 08:41:01: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'ela_h': 2540.3178242641993, 'grad': 7.11}\n",
      "slope default: 0.237523    intercept: 2616.883086   R-squared: 0.991220\n",
      "slope: 0.140647    intercept: 2540.317824   R-squared: 1.000000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:41:01: oggm.workflow: Execute entity task glacier_statistics on 1 glaciers\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# run the model with DEM correction\n",
    "if glacier[num] == \"Washmawapta\":\n",
    "    new_mb_gradient_corr_dem= new_mb_grad_corr_dem(num, dem, glacier, 'ice_density', 900, False, False)\n",
    "else:\n",
    "    new_mb_gradient_corr_dem= new_mb_grad_corr_dem(num, dem, glacier, 'ice_density', 900, False, True)\n",
    "\n",
    "# plot it as a map\n",
    "new_mb_gradient_corr_dem.plot()\n",
    "plt.title(glacier[num]+' DEM corrected Thickness map')\n",
    "\n",
    "# and save it\n",
    "plt.savefig(products+'dem_corr_standard_map_'+ glacier[num]+'.png', dpi=100)\n",
    "\n",
    "# replace 'nan' by  0\n",
    "new_mb_gradient_corr_dem=new_mb_gradient_corr_dem.fillna(0)\n",
    "\n",
    "# make a copy for comparison reasons later on\n",
    "new_mb_gradient_corr_dem_copy=new_mb_gradient_corr_dem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "grad_diff = new_mb_gradient_corr_dem-new_mb_gradient\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.QuadMesh at 0x7f336e4118d0>"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# # the actual differencing and plotting\n",
    "# (new_mb_gradient_corr_dem-new_mb_gradient).plot()\n",
    "#\n",
    "# # label it, save it\n",
    "# plt.title(glacier[num]+' Difference DEM corrected thickness less standard DEM')\n",
    "# plt.savefig(products+'difference_dem_correction_'+ glacier[num]+'.png', dpi=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Parameter optimization\n",
    "\n",
    "We want to run the Model with different values for a certain parameter. Therefore we define a few functions that will iterate over a sequenze of paramter values and allow us to compare the outcome to the observation data by displaying their respective squared residuals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_with_parameter(num,dem, glacier, name, parameter):\n",
    "    \"\"\"\n",
    "    :param num: number of selected glacier\n",
    "    :param dem: list with dem file names\n",
    "    :param glacier: the list with the glacier names\n",
    "    :param name: name of the parameter to be modified\n",
    "    :param parameter: value of the parameter to be modified\n",
    "    :type num: int\n",
    "    :type dem: str list\n",
    "    :type glacier: str list\n",
    "    :type name: str \n",
    "    :type parameter: float\n",
    "    :returns: distributed thickness\n",
    "    :rtype: 2D float array\n",
    "    \"\"\"\n",
    "    \n",
    "    # just to be 100% sure, we are setting the most common parameters to their presettings in order to only vary the desired parameter\n",
    "    cfg.PARAMS['smooth_window'] = sw\n",
    "    cfg.PARAMS['ice_density'] = 900\n",
    "    cfg.PARAMS['trapezoid_lambdas'] = 0.2\n",
    "    cfg.PARAMS['mixed_min_shape'] = 0.001\n",
    "    cfg.PARAMS['downstream_min_shape'] = 0.0001\n",
    "    cfg.PARAMS['mu_star_halfperiod'] = 15\n",
    "    cfg.PARAMS['tstar_search_glacierwide'] = False\n",
    "    cfg.PARAMS['border'] = 10\n",
    "#     cfg.PARAMS['glen_a'] = 2.4e-24\n",
    "#     cfg.PARAMS['inversion_glen_a'] = 2.4e-24\n",
    "    cfg.PARAMS['default_parabolic_bedshape'] = 0.003\n",
    "    cfg.PARAMS['use_shape_factor_for_fluxbasedmodel'] = ''\n",
    "    \n",
    "    #change the parameter (not sure, whether it has any effect here...)\n",
    "    cfg.PARAMS[name] = parameter\n",
    "\n",
    "    # run the model (in this case with DEM correction)\n",
    "    outcome=new_mb_grad_corr_dem(num, dem, glacier, name, parameter, False, use_dem_correction).fillna(0)\n",
    "\n",
    "    #return the thickness distribution\n",
    "    return outcome"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Definitions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# deletes the rows and lines that do only carry zeros for the sake of getting the right resolution \n",
    "# in order to rasterize the observation data to the same resolution and getting perfectly fitting arrays\n",
    "def delete_nodata(data):\n",
    "    \"\"\"\n",
    "    :param data: distributed thickness with 'nan' or '0' in it\n",
    "    :type data: 2D floar array\n",
    "    :returns: distributed thickness\n",
    "    :rtype: 2D float array\n",
    "    \"\"\"\n",
    "    \n",
    "    # get a numpy array out of it\n",
    "    data=data.values\n",
    "\n",
    "    #get the spots were actual values are\n",
    "    coordinates=np.nonzero(data)\n",
    "    \n",
    "    # create a fresh array of the minimal shape to contain all non-zero values\n",
    "    result=np.zeros(((1+np.max(coordinates[0])-np.min(coordinates[0])) ,(1+np.max(coordinates[1])-np.min(coordinates[1]))))\n",
    "\n",
    "    # fill it with the values\n",
    "    for i in range(0, len(coordinates[0])):\n",
    "        result[coordinates[0][i]-np.min(coordinates[0])][coordinates[1][i]-np.min(coordinates[1])]=data[coordinates[0][i]][coordinates[1][i]]\n",
    "    \n",
    "    # flip it and rotate it by pi\n",
    "    return np.flip(np.rot90(result,2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "# a simple function to open the observation data and return it as an array\n",
    "def read_obs_data(path):\n",
    "    \"\"\"\n",
    "    :param path: distributed thickness with 'nan' or '0' in it\n",
    "    :type path: str\n",
    "    :returns: measurement points as an array\n",
    "    :rtype: 2D float array\n",
    "    \"\"\"\n",
    "    src_obs = gdal.Open(path)\n",
    "    return src_obs.ReadAsArray()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fantanstic, both arrays have the same shape of:  (106, 87)\n"
     ]
    }
   ],
   "source": [
    "# We open the observation data for the respective glacier and print out the shape and compare it with the shape of the OGGM:\n",
    "\n",
    "obs_data = read_obs_data('/home/pelto/oggm_runs/OGGM/MONSTER_project/MONSTER/obs_data_' + glacier[num] + '.tif')\n",
    "\n",
    "if obs_data.shape == delete_nodata(new_mb_gradient_corr_dem).shape:\n",
    "    print('Fantanstic, both arrays have the same shape of: ' ,obs_data.shape )\n",
    "else:\n",
    "    print('Oh, Oh! Something went wrong because the observation data has the shape ' ,obs_data.shape , ' and the OGGM out put has a shape of: ' ,delete_nodata(new_mb_gradient_corr_dem).shape )\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Layover-Control\n",
    "\n",
    "This is a checking tool for adjusting the orientation of the observation data with respect to the model shapefile (in order to make it fit, we had to flip and rotate it in the 'delete_nodata' function). The resolution is not great but is should give an optical feedback on the fit. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 91,
     "metadata": {
      "image/png": {
       "width": 400
      }
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# run the delete_nodata function on one of the recent outputs and save it in a temporay variable\n",
    "temp=delete_nodata(new_mb_gradient_corr_dem)\n",
    "\n",
    "# get its shape\n",
    "h,w = temp.shape\n",
    "\n",
    "# create a new array with zeros and of same size\n",
    "data = np.zeros((h, w, 3), dtype=np.uint8)\n",
    "\n",
    "# and get the data point coordinates\n",
    "nonzero=np.nonzero(temp)\n",
    "\n",
    "# run through the OGGM output and write into the array as RED color code\n",
    "for i in range(0,len(nonzero[0])):\n",
    "    data[nonzero[0][i]][nonzero[1][i]]= [temp[nonzero[0][i]][nonzero[1][i]], 0, 0]\n",
    "    \n",
    "# run through the observation data and write into the array as GREEN color code\n",
    "for i in range(0,len(np.nonzero(obs_data)[0])):\n",
    "    data[np.nonzero(obs_data)[0][i]][np.nonzero(obs_data)[1][i]]= [0, obs_data[np.nonzero(obs_data)[0][i]][np.nonzero(obs_data)[1][i]], 0]\n",
    "    \n",
    "# save and show\n",
    "img = imagery.fromarray(data, 'RGB')\n",
    "img.save(products+ 'overlay_test_'+ glacier[num]+ '.png')\n",
    "\n",
    "from IPython.display import Image\n",
    "Image(filename=products+'overlay_test_'+ glacier[num]+ '.png', width=400) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Quadratic Error\n",
    "\n",
    "\n",
    "We want a function that gives us the accumulated quadratic residuals of the observation data compared to the OGGM output. Because we may have very few points that are assigned a measurement but do not locate on the OGGM glacier, we have to do a little extra: We count those pixels, print their share on the entire quadratic errors and subtract their errors from the result in order not to confuse the optimization process later on. However, their share usually is below 1-2% - their impact therefore limited. This is only done for checking and consistency reasons"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "def quadratic_errors(obs_data, mesharray):\n",
    "    \"\"\"\n",
    "    :param obs_data: array with measured values on respective spots\n",
    "    :param mesharray: distributed thickness of the model output (should have same shape as obs_data)\n",
    "    :type obs_data: 2D float array\n",
    "    :type mesharray: 2D float array\n",
    "    :returns: quadrativc error (problems corrected)\n",
    "    :rtype: float\n",
    "    \"\"\"\n",
    "    \n",
    "    # counts number of \"problematic\" pixels\n",
    "    count=0\n",
    "    \n",
    "    # coordinates of the data points\n",
    "    coordinates=np.nonzero(obs_data)\n",
    "    \n",
    "    # accumulated 'problematic pixel' quadratic error\n",
    "    problems=0\n",
    "    prob=0 #problems per grid cell summed\n",
    "    # total accumulated quadratic error\n",
    "    err=0\n",
    "    ME=0\n",
    "    MAE=0\n",
    "\n",
    "    # we go through all the measured points and accumulate the squared differences to the OGGM at that point\n",
    "    for i in range(0,len(coordinates[0])):\n",
    "        \n",
    "        # if however, the OGGM shows a thickness of zero there, this seems to be a \"problematic\" pixel \n",
    "        if mesharray[coordinates[0][i]][coordinates[1][i]]==0:\n",
    "            \n",
    "            # add to problems share\n",
    "            problems=problems+obs_data[coordinates[0][i]][coordinates[1][i]]**2\n",
    "            prob=(prob+obs_data[coordinates[0][i]][coordinates[1][i]])\n",
    "            \n",
    "            # increment problem number\n",
    "            count=count+1\n",
    "        \n",
    "        # add squared difference\n",
    "        err=err+(obs_data[coordinates[0][i]][coordinates[1][i]]-mesharray[coordinates[0][i]][coordinates[1][i]])**2\n",
    "        \n",
    "        ##add mean error and mean absolute error\n",
    "        ME_arr = obs_data[coordinates[0][i]][coordinates[1][i]]-mesharray[coordinates[0][i]][coordinates[1][i]]\n",
    "        ME=ME+(obs_data[coordinates[0][i]][coordinates[1][i]]-mesharray[coordinates[0][i]][coordinates[1][i]])\n",
    "        MAE=MAE+(np.abs((obs_data[coordinates[0][i]][coordinates[1][i]]-mesharray[coordinates[0][i]][coordinates[1][i]])))\n",
    "    \n",
    "    # print of details on \"problematic\" pixels\n",
    "    print('found ', count , ' problem(s) within a total number of data point of: ', len(coordinates[0]), ' problem share ', problems/err,)\n",
    "    \n",
    "    ME_final = (ME)/(np.count_nonzero(obs_data))\n",
    "    MAE_final = (MAE)/(np.count_nonzero(obs_data))\n",
    "    problems_per = prob / (np.count_nonzero(obs_data)) ##quite small or zero at all sites, but worth evaluating\n",
    "    \n",
    "    print('corrected quadratic error is: ', err-problems)\n",
    "    print('ME:', ME_final-problems_per)\n",
    "    print('MAE:', MAE_final-problems_per)\n",
    "    print('mear', ME_arr)\n",
    "    return (err-problems, ME_final-problems_per, MAE_final-problems_per, ME_arr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "# a function that runs the (DEM-corrected) OGGM and uses the quadratic_errors function to return the quadratic errors\n",
    "def master_analysis(name, parameter, obs_data, glacier, num):\n",
    "    \"\"\"\n",
    "    :param name: name of the parameter to be modified\n",
    "    :param parameter: value of the parameter to be modified\n",
    "    :param obs_data: measured data points\n",
    "    :param glacier: the list with the glacier names\n",
    "    :param num: number of selected glacier\n",
    "    :type name: str\n",
    "    :type parameter: float\n",
    "    :type obs_data: 2D float array\n",
    "    :type glacier: str list\n",
    "    :type num: int\n",
    "    :returns: quadratic errors\n",
    "    :rtype: float\n",
    "    \"\"\"\n",
    "    quadmesh= run_with_parameter(num, dem, glacier, name, parameter)\n",
    "    mesharray = delete_nodata(quadmesh)\n",
    "    \n",
    "    return quadratic_errors(obs_data, mesharray)\n",
    "\n",
    "# this function takes in the name of the parameter that is to be modified and a float list of the values for whose the OGGM is \n",
    "# to be run and the quadratic errors to be calculated. It returns an array with those values and their respective quadratic errors\n",
    "def optimize(name, values, obs_data, glacier, num):\n",
    "    \"\"\"\n",
    "    :param name: name of the parameter to be modified\n",
    "    :param values: values of the parameter to be modified\n",
    "    :param obs_data: measured data points\n",
    "    :param glacier: the list with the glacier names\n",
    "    :param num: number of selected glacier\n",
    "    :type name: str\n",
    "    :type values: float list\n",
    "    :type obs_data: 2D float array\n",
    "    :type glacier: str list\n",
    "    :type num: int\n",
    "    :returns: [values, quad_errors]\n",
    "    :rtype: float array\n",
    "    \"\"\"\n",
    "    \n",
    "    # here we store the original parameter value\n",
    "    temp= cfg.PARAMS[name]\n",
    "    \n",
    "    #an array to store the quadratic errors in\n",
    "    quad_errors=[]\n",
    "    for i in range(0,len(values)):\n",
    "        quad_errors.append(master_analysis(name, values[i], obs_data, glacier, num))\n",
    "        \n",
    "    # and here we reset it to the default value\n",
    "    cfg.PARAMS[name] = temp\n",
    "    \n",
    "    return [values, quad_errors]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Enter the parameter, interval and number of points that you want to get the quadratic error from \n",
    "\n",
    "The parameter intervall will be a linearly spaced sequence from min_value to max_value with number_of_steps steps.\n",
    "\n",
    "Also, choose, whether you prefer to run the iteration with the DEM correction. In the most cases it is strongly recommended to do so."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "parameter_name='inversion_glen_a'\n",
    "# best_value = 4.84e-24, 0.596e-24, 9.5e-24, 0.142e-24, 4.544e-24, 14.627e-24, 3.55e-24#3.432e-24\n",
    "best_value = 7.5e-24, 2.75e-24, 10.5e-24, 0.142e-24, 4.544e-24, 14.627e-24, 3.55e-24#3.432e-24 \n",
    "max_value=best_value[num]+1.0e-24  #0.1e-24\n",
    "min_value=best_value[num]-1.0e-24  #54e-24\n",
    "number_of_steps=20\n",
    "\n",
    "if glacier[num] == \"Washmawapta\":\n",
    "    use_dem_correction=False\n",
    "else:\n",
    "    use_dem_correction=True\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The actual iteration takes place here\n",
    "\n",
    "%%capture supresses the output, which is quite annoying if the OGGM runs several dozen times. But if you want to watch it work, go for it and delete that line.\n",
    "\n",
    "Apart from that, we basically run the optimize function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "to_plot=optimize(parameter_name,np.linspace(min_value, max_value, number_of_steps), obs_data ,glacier, num)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we plot the quadratic errors as a function of the parameter values. We also print out the minimum."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(2117182.6765332455,\n",
       "  -2.2190624533347716,\n",
       "  36.98069467822325,\n",
       "  -2.4510955810546875),\n",
       " (2109734.8683832046,\n",
       "  -1.8746393791680198,\n",
       "  36.93067601815011,\n",
       "  -2.2041549682617188),\n",
       " (2102624.566840528, -1.5352352359919874, 36.88406077866415, -1.9608154296875),\n",
       " (2095838.9100245433,\n",
       "  -1.2007183149022964,\n",
       "  36.83991756254029,\n",
       "  -1.720977783203125),\n",
       " (2089365.8172289664,\n",
       "  -0.8709596083002183,\n",
       "  36.79713355453269,\n",
       "  -1.48455810546875),\n",
       " (2083193.7328446321,\n",
       "  -0.5458379046430866,\n",
       "  36.7562759246641,\n",
       "  -1.2514572143554688),\n",
       " (2077311.6155048637,\n",
       "  -0.2252345316618392,\n",
       "  36.71772100393054,\n",
       "  -1.0215911865234375),\n",
       " (2071708.8179012868,\n",
       "  0.09096478952944853,\n",
       "  36.68134386562607,\n",
       "  -0.794891357421875),\n",
       " (2066375.3959255726,\n",
       "  0.4028690046477086,\n",
       "  36.648783213883924,\n",
       "  -0.5712661743164062),\n",
       " (2061301.623965072, 0.7105840849644929, 36.61975255058807, -0.35064697265625),\n",
       " (2056478.293880534,\n",
       "  1.0142121986278052,\n",
       "  36.593955602460696,\n",
       "  -0.13295745849609375),\n",
       " (2051896.5895188665,\n",
       "  1.3138520370409328,\n",
       "  36.56971227312551,\n",
       "  0.081878662109375),\n",
       " (2047548.0809768452,\n",
       "  1.609598050997095,\n",
       "  36.547586758159895,\n",
       "  0.293914794921875),\n",
       " (2043424.7332834536,\n",
       "  1.9015419853543771,\n",
       "  36.52641462816776,\n",
       "  0.5032272338867188),\n",
       " (2039518.6558735422,\n",
       "  2.1897733512433986,\n",
       "  36.50682829421701,\n",
       "  0.7098770141601562),\n",
       " (2035822.5465431125,\n",
       "  2.474376708558462,\n",
       "  36.48873199768437,\n",
       "  0.9139328002929688),\n",
       " (2032329.3134875912,\n",
       "  2.7554354621368704,\n",
       "  36.47138888164631,\n",
       "  1.1154403686523438),\n",
       " (2029032.0509528574,\n",
       "  3.0330298155256843,\n",
       "  36.45503937619404,\n",
       "  1.3144607543945312),\n",
       " (2025924.245479132,\n",
       "  3.3072366830214714,\n",
       "  36.439910279894335,\n",
       "  1.5110626220703125),\n",
       " (2022999.63315514, 3.5781318863618723, 36.42661375906861, 1.7052841186523438)]"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# np.shape(to_plot)\n",
    "to_plot[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "OUT=pd.DataFrame({})\n",
    "for j in np.arange(0,len(to_plot[1])):\n",
    "    er=to_plot[1][j][0]\n",
    "    me=to_plot[1][j][1]\n",
    "    mae=to_plot[1][j][2]\n",
    "    ME_AR = to_plot[1][j][3]\n",
    "    OUT=OUT.append({'Error':er,'ME':me,'MAE':mae, 'mear':ME_AR},ignore_index=True)\n",
    "    \n",
    "ME_arr = to_plot[1][3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.7052841186523438"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ME_AR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-2.4510955810546875"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "OUT.mear[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minimal value is 36.42661375906861 at inversion_glen_a = 1.05e-23\n"
     ]
    }
   ],
   "source": [
    "from matplotlib import rcParams\n",
    "rcParams['axes.titlepad'] = 30\n",
    "plt.plot(to_plot[0],OUT.MAE,linestyle='--', marker='o', color='b')\n",
    "plt.xlabel('value of '+parameter_name)\n",
    "plt.ylabel('pointwise accumulated quadratic discrepancies [m²]')\n",
    "plt.title('Quadratic errors of depth in dependence of ' +parameter_name)\n",
    " \n",
    "# save it\n",
    "plt.savefig(products+'quadratic_errors_'+parameter_name+'_from_'+str(min_value)+'_to_'+str(max_value)+'_'+ glacier[num]+'.png', dpi=100)\n",
    "\n",
    "plt.show()\n",
    "\n",
    "#small minimum function\n",
    "x=0\n",
    "for i in range(0,len(OUT.Error)):\n",
    "    if OUT.Error[i]==np.min(OUT.Error):\n",
    "        x=i\n",
    "\n",
    "# print out\n",
    "print('Minimal value is '+str(np.min(OUT.MAE))+ ' at ' + parameter_name+' = ' +str(np.round(to_plot[0][x],decimals=28)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Deep analysis of best parameter value\n",
    "\n",
    "Enter the value of the given Paramter, that yields a minimum in quadratic errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.05e-23"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# take the minimum from plot above\n",
    "best_value=to_plot[0][x]\n",
    "best_value\n",
    "#regional value\n",
    "# best_value = 5.08e-24 \n",
    "\n",
    "# select it manually from list\n",
    "# best_value = 6.740e-24, 0.774e-24, 8.095e-24, 0.142e-24, 4.050e-24, 13.0e-24, 3.810e-24  \n",
    "# best_value = 4.84e-24, 0.596e-24, 7.381e-24, 0.142e-24, 4.544e-24, 14.627e-24, 3.432e-24  \n",
    "# best_value = best_value[num]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we run the model with the best parameter value to create seperate saves and a deep analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:45:15: oggm.cfg: Using configuration file: /home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/oggm/params.cfg\n",
      "2020-03-31 08:45:17: oggm.workflow: Execute entity task glacier_masks on 1 glaciers\n",
      "2020-03-31 08:45:17: oggm.workflow: Multiprocessing: using all available processors (N=8)\n",
      "2020-03-31 08:45:17: oggm.workflow: Execute entity task compute_centerlines on 1 glaciers\n",
      "2020-03-31 08:45:17: oggm.workflow: Execute entity task initialize_flowlines on 1 glaciers\n",
      "2020-03-31 08:45:17: oggm.workflow: Execute entity task compute_downstream_line on 1 glaciers\n",
      "2020-03-31 08:45:17: oggm.workflow: Execute entity task compute_downstream_bedshape on 1 glaciers\n",
      "2020-03-31 08:45:17: oggm.workflow: Execute entity task catchment_area on 1 glaciers\n",
      "2020-03-31 08:45:18: oggm.workflow: Execute entity task catchment_intersections on 1 glaciers\n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/geopandas/io/file.py:108: FionaDeprecationWarning: Use fiona.Env() instead.\n",
      "  with fiona.drivers():\n",
      "2020-03-31 08:45:18: oggm.workflow: Execute entity task catchment_width_geom on 1 glaciers\n",
      "2020-03-31 08:45:19: oggm.workflow: Execute entity task catchment_width_correction on 1 glaciers\n",
      "2020-03-31 08:45:19: oggm.workflow: Execute entity task process_cru_data on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task local_t_star on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task mu_star_calibration on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task prepare_for_inversion on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task mass_conservation_inversion on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task filter_inversion_output on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n",
      "2020-03-31 08:45:20: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "slope: 0.237854    intercept: 2621.998664   R-squared: 0.989877\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:45:21: oggm.workflow: Execute entity task apparent_mb_from_linear_mb on 1 glaciers\n",
      "2020-03-31 08:45:21: oggm.workflow: Execute entity task prepare_for_inversion on 1 glaciers\n",
      "2020-03-31 08:45:21: oggm.workflow: Execute entity task mass_conservation_inversion on 1 glaciers\n",
      "2020-03-31 08:45:21: oggm.workflow: Execute entity task filter_inversion_output on 1 glaciers\n",
      "2020-03-31 08:45:21: oggm.workflow: Execute entity task distribute_thickness_per_altitude on 1 glaciers\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'ela_h': 2545.680813745039, 'grad': 7.11}\n",
      "slope default: 0.237854    intercept: 2621.998664   R-squared: 0.989877\n",
      "slope: 0.140647    intercept: 2545.680814   R-squared: 1.000000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-03-31 08:45:22: oggm.workflow: Execute entity task glacier_statistics on 1 glaciers\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "best_model= new_mb_grad_corr_dem(num, dem, glacier, parameter_name, best_value, True, False)\n",
    "best_model.plot()\n",
    "plt.title(glacier[num]+' best approximation thickness map ' + parameter_name + ' = ' +str(best_value))\n",
    "plt.savefig(products+'dem_corr_best_map_'+ glacier[num]+'.png', dpi=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<xarray.DataArray 'distributed_thickness' ()>\n",
      "array(83.4092, dtype=float32)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "83.40920249469214"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy.ma as ma\n",
    "print(best_model.mean())\n",
    "best_model=best_model.fillna(0)\n",
    "thickness = ma.masked_less(best_model,1)\n",
    "thickness.mean()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # plot and save\n",
    "# (best_model-new_mb_gradient_corr_dem_copy).plot(cmap='RdBu', vmin=-100, vmax=100)\n",
    "# plt.title(glacier[num]+' Difference '+ parameter_name+ ' corrected less uncorrected thickness map (both DEM corrected)')\n",
    "# plt.savefig(products+'difference_'+ parameter_name + '_corrected_vs_uncorrected_'+ glacier[num]+'.png', dpi=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fantanstic, both arrays have the same shape of:  (106, 87)\n",
      "BM: 138.12878050387485\n",
      "Obs: 138.5816\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/ipykernel_launcher.py:8: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  \n",
      "/home/pelto/anaconda2/envs/oggm_env/lib/python3.6/site-packages/ipykernel_launcher.py:8: RuntimeWarning: invalid value encountered in true_divide\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "#need fill na for shapes to match, then delete no data\n",
    "\n",
    "if obs_data.shape == delete_nodata(best_model).shape:\n",
    "    print('Fantanstic, both arrays have the same shape of: ' ,obs_data.shape )\n",
    "else:\n",
    "    print('Oh, Oh! Something went wrong because the observation data has the shape ' ,obs_data.shape , ' and the OGGM out put has a shape of: ' ,delete_nodata(best_model).shape )\n",
    "\n",
    "BM = delete_nodata(best_model) * obs_data / obs_data\n",
    "print('BM:', np.nanmean(BM))\n",
    "obs_data[obs_data == 0] = np.nan\n",
    "print('Obs:', np.nanmean(obs_data))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Export thickness as pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import pickle \n",
    "# # outfile= open(products + glacier[num] +'_optimized_gprlocs.pkl', 'wb')\n",
    "# outfile= open(products + glacier[num] +'_regional_gprlocs.pkl', 'wb')\n",
    "# pickle.dump(BM, outfile)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "# best_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "# best_model.to_netcdf(products + glacier[num] + '_optimized.nc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}