Winds.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Retrieving winds from the vorticity and divergence fields"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Deep learning is often used to extract features from the data. In geosciences its increasingly used in the parameterization of unresolved sub-grid dynamics. Few examples include the convection parameterization in general circulation model [(Gentine et al., 2018)](https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018GL078202), ocean data inference [(Bolton and Zanna, 2019)](https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018MS001472), radiative transfer calculations [(Pal et al., 2019)]( https://doi.org/10.1029/2018GL081646), rainfall prediction [(Miao et al., 2015)](https://doi.org/10.1175/JHM-D-14-0174.1), flux estimation [(Jung et al., 2011)](https://doi.org/10.1029/2010JG001566).\n",
    "The idea is to allow the machine learning algorithm to learn directly from the data without the need for explicit formulation. These models tend to be computationally efficient and often more accurate than the traditional methods. \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook we try to retrieve the wind fields from the vorticity and divergence fields. Though its not exactly a parameterization problem but the idea is similar. We use deep learning to learn a function which maps the vorticty and divergence to the wind fields, but without using the spherical transforms. \n",
    "2.5$^{\\circ}$ X 2.5$^{\\circ}$, 6-hourly ERA-Interim data from 1979-2018 (40 years) is used. Vorticity and divergence fields are used as input and zonal and meridional winds are the targets. All the fields are at 500 hpa level. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn\n",
    "import keras\n",
    "import xarray\n",
    "plt.ion()\n",
    "seaborn.set_style('whitegrid',rc={\"axes.edgecolor\":'black'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = xarray.open_dataset('Data/era_winds.nc')\n",
    "vrt  = data['vo']\n",
    "div  = data['d']\n",
    "u    = data['u']\n",
    "v    = data['v']\n",
    "lats = data['latitude']\n",
    "lons = data['longitude']\n",
    "data.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First 50,000 samples are used for training and the remaining 8440 samples for testing. Sanples are randomly shuffled in each of these sets. Cross-validation is used to evaluate different models.\n",
    "The inputs and outputs are normalized by its training  mean and variance, so that they have similar scales. This often helps in speeding up the training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Normalization\n",
    "def normData(var, idx):\n",
    "    varMean = np.mean(var[0:idx])\n",
    "    varStd  = np.std(var[0:idx])\n",
    "    varNorm = (var - varMean) / varStd\n",
    "    return varNorm, varMean.values, varStd.values\n",
    "    \n",
    "\n",
    "vrtNorm, vrtMean, vrtStd = normData(vrt, 50000)\n",
    "divNorm, divMean, divStd = normData(div, 50000)\n",
    "uNorm, uMean, uStd = normData(u, 50000)\n",
    "vNorm, vMean, vStd = normData(v, 50000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape: (50000, 73, 144, 2)\n",
      "X_test shape: (8440, 73, 144, 2)\n"
     ]
    }
   ],
   "source": [
    "# creating train and test sets \n",
    "np.random.seed(23)\n",
    "\n",
    "ntotal = len(vrt)\n",
    "ntrain = 50000\n",
    "ntest  = ntotal - ntrain \n",
    "\n",
    "trainIdx = np.random.permutation(ntrain)\n",
    "testIdx  = np.random.permutation(ntest) + ntrain \n",
    "\n",
    "X = np.stack((vrtNorm,divNorm),axis=3)\n",
    "X_train = X[trainIdx]\n",
    "X_test  = X[testIdx]\n",
    "u_train = uNorm[trainIdx]\n",
    "u_test  = uNorm[testIdx]\n",
    "v_train = vNorm[trainIdx]\n",
    "v_test  = vNorm[testIdx]\n",
    "\n",
    "print (\"X_train shape:\", X_train.shape)\n",
    "print (\"X_test shape:\", X_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model has 2 inputs (vorticity and divergence) and 2 outputs (zonal and meridional wind).\n",
    "Six layer deep, fully connected network is used. First 3 layers are shared between the two outputs. It also has another 3 layers for each of the outputs which is not shared. Each layer has 512 units wuth ReLU activation, except the last one. The last layer has 10512 units (with no activation) which is then reshaped to the output image of size 73 X 144. Adam optimization is used with mean squared loss. No regularization was used as the model was not overfitting. Model summary is printed below along with the block diagram.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/suhas/miniconda3/envs/ml/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "inputs (InputLayer)             (None, 73, 144, 2)   0                                            \n",
      "__________________________________________________________________________________________________\n",
      "flat (Flatten)                  (None, 21024)        0           inputs[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "dense1 (Dense)                  (None, 512)          10764800    flat[0][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "dense2 (Dense)                  (None, 512)          262656      dense1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "dense3 (Dense)                  (None, 512)          262656      dense2[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "denseU1 (Dense)                 (None, 512)          262656      dense3[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "denseV1 (Dense)                 (None, 512)          262656      dense3[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "denseU2 (Dense)                 (None, 512)          262656      denseU1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "denseV2 (Dense)                 (None, 512)          262656      denseV1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "denseU3 (Dense)                 (None, 10512)        5392656     denseU2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "denseV3 (Dense)                 (None, 10512)        5392656     denseV2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "U (Reshape)                     (None, 73, 144)      0           denseU3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "V (Reshape)                     (None, 73, 144)      0           denseV3[0][0]                    \n",
      "==================================================================================================\n",
      "Total params: 23,126,048\n",
      "Trainable params: 23,126,048\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense, Input, Reshape, Flatten\n",
    "from keras.optimizers import Adam\n",
    "\n",
    "# model = Sequential()\n",
    "inputs = Input(shape=(X.shape[1],X.shape[2],X.shape[3]), name='inputs')\n",
    "flat   = Flatten(name='flat')(inputs)\n",
    "\n",
    "\n",
    "dense1 = Dense(512, activation = 'relu', name='dense1')(flat)\n",
    "dense2 = Dense(512, activation = 'relu', name='dense2')(dense1)\n",
    "dense3 = Dense(512, activation = 'relu', name='dense3')(dense2)\n",
    "\n",
    "denseU1 = Dense(512, activation = 'relu', name='denseU1')(dense3)\n",
    "denseU2 = Dense(512, activation = 'relu', name='denseU2')(denseU1)\n",
    "denseU3 = Dense(10512, name='denseU3')(denseU2)\n",
    "outU    = Reshape((73, 144), name='U')(denseU3)\n",
    "\n",
    "denseV1 = Dense(512, activation = 'relu', name='denseV1')(dense3)\n",
    "denseV2 = Dense(512, activation = 'relu', name='denseV2')(denseV1)\n",
    "denseV3 = Dense(10512, name='denseV3')(denseV2)\n",
    "outV    = Reshape((73, 144), name='V')(denseV3)\n",
    "\n",
    "model = Model(inputs, [outU, outV])\n",
    "model.summary()\n",
    "\n",
    "model.compile(optimizer=Adam(),\n",
    "              loss='mse')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 6,
     "metadata": {
      "image/png": {
       "height": 419,
       "width": 157
      }
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from keras.utils import plot_model\n",
    "plot_model(model, to_file='model.png')\n",
    "from IPython.display import Image\n",
    "Image(retina=True, filename='model.png')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model is trained for 10 epochs with a batch size of 128 samples. 20% of training data is used for validation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/suhas/miniconda3/envs/ml/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Train on 40000 samples, validate on 10000 samples\n",
      "Epoch 1/10\n",
      "40000/40000 [==============================] - 103s 3ms/step - loss: 1.0657 - U_loss: 0.4182 - V_loss: 0.6476 - val_loss: 0.7645 - val_U_loss: 0.3042 - val_V_loss: 0.4604\n",
      "Epoch 2/10\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.6674 - U_loss: 0.2588 - V_loss: 0.4086 - val_loss: 0.6036 - val_U_loss: 0.2305 - val_V_loss: 0.3731\n",
      "Epoch 3/10\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.5559 - U_loss: 0.2138 - V_loss: 0.3421 - val_loss: 0.5316 - val_U_loss: 0.2067 - val_V_loss: 0.3250\n",
      "Epoch 4/10\n",
      "40000/40000 [==============================] - 102s 3ms/step - loss: 0.5096 - U_loss: 0.1955 - V_loss: 0.3141 - val_loss: 0.5008 - val_U_loss: 0.1939 - val_V_loss: 0.3070\n",
      "Epoch 5/10\n",
      "40000/40000 [==============================] - 100s 2ms/step - loss: 0.4800 - U_loss: 0.1848 - V_loss: 0.2952 - val_loss: 0.4760 - val_U_loss: 0.1846 - val_V_loss: 0.2914\n",
      "Epoch 6/10\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.4648 - U_loss: 0.1785 - V_loss: 0.2863 - val_loss: 0.4644 - val_U_loss: 0.1809 - val_V_loss: 0.2835\n",
      "Epoch 7/10\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.4547 - U_loss: 0.1745 - V_loss: 0.2802 - val_loss: 0.4487 - val_U_loss: 0.1722 - val_V_loss: 0.2765\n",
      "Epoch 8/10\n",
      "40000/40000 [==============================] - 99s 2ms/step - loss: 0.4373 - U_loss: 0.1682 - V_loss: 0.2691 - val_loss: 0.4436 - val_U_loss: 0.1702 - val_V_loss: 0.2734\n",
      "Epoch 9/10\n",
      "40000/40000 [==============================] - 101s 3ms/step - loss: 0.4339 - U_loss: 0.1663 - V_loss: 0.2676 - val_loss: 0.4285 - val_U_loss: 0.1650 - val_V_loss: 0.2634\n",
      "Epoch 10/10\n",
      "40000/40000 [==============================] - 100s 3ms/step - loss: 0.4289 - U_loss: 0.1642 - V_loss: 0.2647 - val_loss: 0.4293 - val_U_loss: 0.1649 - val_V_loss: 0.2644\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, [u_train, v_train], epochs=10, batch_size=128, validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 10)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(1,len(loss)+1)\n",
    "\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, label='Training loss')\n",
    "plt.plot(epochs, val_loss, label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.xlim(1,len(loss))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training and validation losses are approaching a steady value around epoch 10. Now we will predict the test output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "up,vp = model.predict(X_test) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Rescale the normalized outputs to its original values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "u_pred = up * uStd + uMean\n",
    "v_pred = vp * vStd + vMean\n",
    "u_true = u_test * uStd + uMean\n",
    "v_true = v_test * vStd + vMean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMS Error of u wind (in m/s) : 4.897034\n",
      "RMS Error of v wind (in m/s) : 4.71643\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "u_error = mean_squared_error(np.ravel(u_true), np.ravel(u_pred))\n",
    "v_error = mean_squared_error(np.ravel(v_true), np.ravel(v_pred))\n",
    "\n",
    "print (\"RMS Error of u wind (in m/s) :\", np.sqrt(u_error))\n",
    "print (\"RMS Error of v wind (in m/s) :\", np.sqrt(v_error))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comparison of predicted and actual outputs for one time step in the test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'v actual')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 8 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "seaborn.set_style('whitegrid',rc={\"axes.edgecolor\":'black'})\n",
    "fig,ax = plt.subplots(2,2,figsize=(12,10),squeeze=False)\n",
    "\n",
    "def plotFig(var,i,j):\n",
    "    plotLevels = np.linspace(-np.amax(var), np.amax(var),11)\n",
    "    contours = ax[i,j].contourf(lons,lats,var, levels=plotLevels, cmap='coolwarm', extend='both')\n",
    "    cb = plt.colorbar(contours,ax=ax[i,j], orientation='horizontal',format='%.0d',shrink=0.8,aspect=30,pad=0.14)\n",
    "    ax[i,j].grid(alpha=0.7,color='k',linestyle='dotted',dashes=[1,5 ],linewidth=1,zorder=10)\n",
    "    ax[i,j].set_xticks(np.linspace(0,360,7))\n",
    "    ax[i,j].set_yticks(np.linspace(-90,90,7))\n",
    "\n",
    "idx = 1000\n",
    "plotFig(u_pred[idx],0,0)\n",
    "plotFig(u_true[idx],0,1)\n",
    "plotFig(v_pred[idx],1,0)\n",
    "plotFig(v_true[idx],1,1)\n",
    "\n",
    "ax[0,0].set_ylabel('Latitude',size='12', fontname = 'Dejavu Sans')\n",
    "ax[1,0].set_ylabel('Latitude',size='12', fontname = 'Dejavu Sans')  \n",
    "ax[1,0].set_xlabel('Longitude',size='12', fontname = 'Dejavu Sans')\n",
    "ax[1,1].set_xlabel('Longitude',size='12', fontname = 'Dejavu Sans')\n",
    "\n",
    "ax[0,0].set_title('u predicted',size='14', fontname = 'Dejavu Sans')\n",
    "ax[0,1].set_title('u actual',size='14', fontname = 'Dejavu Sans')\n",
    "ax[1,0].set_title('v predicted',size='14', fontname = 'Dejavu Sans')\n",
    "ax[1,1].set_title('v actual',size='14', fontname = 'Dejavu Sans')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predicted outputs look more or less similar to the actual ones. Its capturing most of the large scale structures though its more smoothened. The predicted values also underestimate the actual magnitudes. Perhaps a higher capacity network would better capture these subtleties. But overall the performace seems to be satisfactory."
   ]
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}