CHAZ (Columbia Tropical Cyclone Hazard Model)

Table of Contents

I. Overview

II. Getting Started With CHAZ

       1) Getting the code

       2) System Dependencies

       3) Operating System

       4) Setting Up Python Environment

III. Basic CHAZ Structure

IV. CHAZ-pre

V. Checking Initial conditions

VI. Running CHAZ (Downscaling)

VII. Output of CHAZ

VIII. Disclaimer

IX. License

I. Overview

The Columbia HAZard model (CHAZ) is a statistical-dynamical downscaling model for estimating tropical cyclone hazard. The CHAZ model consists of three primary components for describing tropical cyclone activity from genesis to lysis; these components are genesis, track, and intensity modules. The genesis module uses Tropical Cyclone Genesis Index (TCGI, developed by Tippett et al, 2011 to estimate the seeding rate of the storm precursors. The seeds are then passed to a Beta-Advection Model (BAM) that moves the storm forward with giving environmental sterring flow. The intensity module then evolves storms beyond genesis using an Autoregressive Model that involves a deterministic and a stochastic forcing elements. The underlying science of the CHAZ model can be found at Lee et al. 2018).

Generally speaking, there are two primary steps for running CHAZ: preprocessing and downscaling. Preprocessing includes collecting global model data, calculating PI, TCGI, wind covariance matrix, putting the data into a standard format for downscaling calculation, etc. Preprocessing codes are model-dependent, meaning that you will need to modify them when switch global models. Codes for conducting downscaling are insensitive to what global model you are using.

II. Getting Started with CHAZ

1) Getting the code

CHAZvBeta folder at top of the page.

2) System dependencies

The CHAZ model was developed in python2 and has only been running on linux and Mac OSX machines. Here we provide our experience. However, this does not mean CHAZ can only be run on a linux system. We welcome users' contributions in terms of operating system. We also appreciate any comments, bug-discovery, and improvement of the CHAZ model.

3) Operating System

• Linux/ Mac OSX

4) Setting Up Python Environment

1) Install Anaconda or Python 2. The current code and instruction below is for py2. One can easily convert the code to py3 using '2to3'. 2) Setting up virtual environment

2a. If [Anaconda Python Distribution](https://www.anaconda.com/) is Installed, run the following commands : 

   conda create --name py27 python=2.7 anaconda

   conda activate py27

2b. If Python directly installed, run the following commands: 

   virtualenv -p /usr/bin/python2.7 venv

   source venv/bin/activate

3) pip install the following code:

   pip install numpy==1.15.4

   pip install pandas==0.24.1

   pip install scipy==1.2.0

   pip install netcdf4==1.4.2

   pip install matplotlib==2.2.3

   pip install xarray==0.11.3

   pip install statsmodels==0.9.0

   pip install cftime==1.0.3.4

   pip install netcdftime==1.0.0a2

   pip install dask==1.1.1

   pip install toolz==0.9.0

III. Basic CHAZ structure

In CHAZ, Namelist.py defines all global variables such as the source of the global model forcing, the ensemble members of that global model, the version of genesis module, numbers of track and intensity ensemble members, and whether you are doing CHAZ-pre or CHAZ downscaling. In theory, Namelist.py is the only file you need to modify. CHAZ.py reads in Namelist.py and is the script that calls for all the modules/subroutines. Below we describe input variables that need to be modified when conducting CHAZ-pre and CHAZ downscaling.

Namelist.py: python file containing global variables modified for CHAZ preprocessing and CHAZ downscaling

CHAZ.py: python file that controls preprocessing and downscaling

getMeanStd.py: used to calculate data in coefficient_meanstd.nc if user does not have file (must execute separately from CHAZ)

input: the directory containing default input data necessary for running CHAZ. These data are generated during the CHAZ development stage and are used in Lee et al. 2018. Check https://drive.google.com/file/d/1v_NyAzqGyfUNolY-FOAg9P0_itExPGEC/view?usp=sharing for input data.

        bt_*.nc: basin-wide best track data

        bt_global_predictors: CHAZ predictors for historical events (from bt_*.nc) from monthly ERA-Interim reanalysis data.

        landmask.nc: landmask data

        result_l.nc: regression parameters for near land cases

        result_w.nc: regression parameters for cases over open water

pre: the directory containing data output from preprocessing

        coefficient_meanstd.nc: containing the mean and standard deviations of the predictors for historical events

        Cov*YYYY.nc: monthly-daily wind covaraince

        YYYY*r1i1p1f1.nc: intensity model predictors

        A_*YYYYMM.nc: covariance matrix

src: contains all of the source code necessary to run CHAZ

        calA.py: calculate mean and covariance of synthetic wind

        module_riskModel.py: contain functions necessary to calculate genesis, track, and predictors

        caldetsto_track_lysis.py: return intensities based on calGenisi.py and module_GenBamPred.py output

        preprocess.py: preprocess global model to a standard 2x2 grid format

        module_GenBamPred.py: calculates tracks via Beta-Advection Model

        calWindCov.py: calculate historic wind mean and covariance in preprocessing

tools: contains all of the files with tools necessary to run the source code of CHAZ.

        module_stochastic.py: functions to find means and errors of predictors

        readbst.py: functions creating ibtracs and bst objects

        regression4.py: regression tools

        util.py: miscellaneous functions used in downscaling and preprocessing

IV. CHAZ-pre

Preprocessing involves populating '\pre' with r1i1p1_YYYY.nc, coefficientmeanstd.nc and A_YYYYMM.nc from calWindCov.py, calA.py, getMeanStd.py, and preprocess.py. See preprocessing example.

V) Checking Initial conditions

Prior to conducting CHAZ downscaling, we should check the initial conditions those are nessary for CHAZ runs. They are TCGI_YYYY.mat, r1i1p1_YYYY.nc (from calPreprocess), A_YYYYMM.nc (from calA), coefficent_meanstd.nc (getMeanStd.py).

VI) Chaz Downscaling

1. Changing Global Variables in Namelist.py

2. Make sure all the inputs are in the ipath, opath, and obs_bt_path

3. Check that output is empty before running CHAZ

4. Simply type the following command:$ ./CHAZ.py

See downscaling example for reference.

VII) Output of CHAZ

The output of CHAZ gives a set of netCDF files with names [model name]_[year]_ens[ensemble number].nc where "model name" is the model used (also specified in Namelist.py), "year" is the year of the simulation, and "ensemble number" is the ensemble number represented with three digits.

The dimensions of each netCDF file are lifelength, stormID, and ensembleNum (the specific size of the dimensions can be found using the command $ ncinfo [filename], which returns the following:

<type 'netCDF4._netCDF4.Dataset'>
root group (NETCDF4 data model, file format HDF5):
    dimensions(sizes): lifelength(125), stormID(146), ensembleNum(40)
    variables(dimensions): float64 lifelength(lifelength), float64 stormID(stormID), 
    float64 ensembleNum(ensembleNum), float64 latitude(lifelength,stormID), 
    float64 Mwspd(ensembleNum,lifelength,stormID), float64 time(lifelength,stormID), 
    float64 longitude(lifelength,stormID), int64 year(stormID)
    groups:

"lifelength" is the amount length of time of the storm's life. "stormID" represents how many storms have been run at this year. "ensembleNum" is the number of members in the intensity ensemble.

One possible use of the output data is to create the below track plot:

The below snippet of code creates the above plot:

import matplotlib.pyplot as plt
import netCDF4 as nc
import numpy as np
import cartopy as cart
import cartopy.crs as ccrs
import cartopy.feature as cfeature
import matplotlib.ticker as mticker
from cartopy.mpl.gridliner import LONGITUDE_FORMATTER, LATITUDE_FORMATTER

years = ['2000','2001','2002']

fig = plt.figure(figsize=(8,4))
ax = fig.add_subplot(111)
ax = plt.axes(projection=ccrs.PlateCarree())
ax.coastlines(zorder=100)
gl = ax.gridlines(crs=ccrs.PlateCarree(), draw_labels=True,
        linewidth=0.5, color='gray', alpha=0.5, linestyle='--')
gl.xlabels_top = False
gl.ylabels_right = False
gl.xlocator = mticker.FixedLocator(range(-180,180,60))
gl.ylocator = mticker.FixedLocator(range(-90, 90,15))
gl.xformatter = LONGITUDE_FORMATTER
gl.yformatter = LATITUDE_FORMATTER
ax.add_feature(cart.feature.BORDERS)
ax.set_extent([-180,180,-50,50])

iN = 0
for iy in range(len(years)):
        filename = 'ERA5_'+years[iy]+'_ens000.nc'
        data = nc.Dataset(filename)
        lon = data['longitude'][:]
        lat = data['latitude'][:]
        mwspd = data['Mwspd'][:]
        plt.scatter(lon.ravel(),lat.ravel(),s=1,c=mwspd[iN,:,:].ravel())

plt.colorbar(label='intensity')
plt.xlabel('longitude (degrees)')
plt.ylabel('latitude (degrees)')
plt.title('CHAZ- ERA5 2000-2002')
plt.savefig('CHAZ_storm_tracks_ERA5.png')                                                    

Disclaimer

The output is not bias-corrected. Some bias correction may be necessary to estimate hurricane activity at the regional level.

License

MIT License

Copyright (c) [2020] [The Columbia HAZard model (CHAZ) team]; The CHAZ team consists of Drs. Suzana J. Camargo, Chia-Ying Lee, Michael K. Tippett, and Adam H. Sobel from Columbia University.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Publications

C.-Y. Lee, M.K. Tippett, A.H. Sobel, and S.J. Camargo, 2018. An environmentally forced tropical cyclone hazard model. Journal of Advances in Modeling Earth Systems, 10, 233-241, doi: 10.1002/2017MS001186.

A.H. Sobel, C.-Y. Lee, S.J. Camargo, K.T. Mandli, K.A. Emanuel, P. Mukhopadhyay, and M. Mahakur, 2019. Tropical cyclone hazard to Mumbai in the recent historical climate. Monthly Weather Review, 147, 2355-2366, doi: 10.1175/MWR-D-18-0419.1.

C.-Y. Lee, S.J. Camargo, A.H. Sobel, and M.K. Tippett, 2020. Statistical-dynamical downscaling projections of tropical cyclone activity in a warming climate: Two diverging genesis scenarios. Journal of Climate, 33, 4815-4834, doi: 10.1175/JCLI-D-19-0452.1.

P. Hassanzadeh, C.-Y. Lee, E. Nabizadeh, S.J. Camargo, D. Ma, and L. Yeung, 2020. Effects of climate change on the movement of future landfalling Texas tropical cyclones, Nature Communications, 11, 3319, doi: 10.1038/s41467/s41467-020-17130-7.