Python Plotting Basics¶
Let’s take a look at some basic plotting with Python
If you’d like to follow along, connect to VDI then in a terminal run:
module use /g/data3/hh5/public/modules module load conda/analysis3 jupyter notebook
This will start up a Jupyter Notebook on VDI, which is the same tool that this post is written in. Jupyter lets us work interactively with python and easily share code and plots.
# This is a python command print("Hello world!") # the output appears below
To start out, let’s get some sample data to work with. NCI has a data store with all of the Australian CMIP5 data. You can find this on NCI’s Data Catlogue
I like to use a library called Xarray to work with netCDF datasets. Let’s grab the monthly surface temperature from the ACCESS1-0 AMIP run and load it in
import xarray datapath = "http://dapds00.nci.org.au/thredds/dodsC/rr3/CMIP5/output1/CSIRO-BOM/ACCESS1-0/amip/mon/atmos/Amon/r1i1p1/latest/tas/tas_Amon_ACCESS1-0_amip_r1i1p1_197901-200812.nc" data = xarray.open_dataset(datapath)
I’m accessing the data remotely here using what’s called a THREDDS URL. You could just as easily use the
/g/data filepath if you’re running on VDI.
To see what’s in the file we can just print it. This shows the dimensions, variables and metadata in the file:
<xarray.Dataset> Dimensions: (bnds: 2, lat: 145, lon: 192, time: 360) Coordinates: * time (time) datetime64[ns] 1979-01-16T12:00:00 1979-02-15 ... * lat (lat) float64 -90.0 -88.75 -87.5 -86.25 -85.0 -83.75 -82.5 ... * lon (lon) float64 0.0 1.875 3.75 5.625 7.5 9.375 11.25 13.12 15.0 ... height float64 ... Dimensions without coordinates: bnds Data variables: time_bnds (time, bnds) float64 ... lat_bnds (lat, bnds) float64 ... lon_bnds (lon, bnds) float64 ... tas (time, lat, lon) float32 ... Attributes: institution: CSIRO (Commonwealth Scientific and Indus... institute_id: CSIRO-BOM experiment_id: amip source: ACCESS1-0 2011. Atmosphere: AGCM v1.0 (N... model_id: ACCESS1-0 forcing: GHG, Oz, SA, Sl, Vl, BC, OC, (GHG = CO2,... parent_experiment_id: N/A parent_experiment_rip: r1i1p1 branch_time: 0.0 contact: The ACCESS wiki: http://wiki.csiro.au/co... history: CMIP5 compliant file produced from raw A... references: See http://wiki.csiro.au/confluence/disp... initialization_method: 1 physics_version: 1 tracking_id: 7cfe11fc-5b1c-457d-812b-e95f45e7def4 version_number: v20120115 product: output experiment: AMIP frequency: mon creation_date: 2012-02-17T05:21:53Z Conventions: CF-1.4 project_id: CMIP5 table_id: Table Amon (01 February 2012) 01388cb450... title: ACCESS1-0 model output prepared for CMIP... parent_experiment: N/A modeling_realm: atmos realization: 1 cmor_version: 2.8.0 DODS_EXTRA.Unlimited_Dimension: time
Variables are accessed using dot syntax. Just like the full dataset we can print individual variables to see their metadata:
surface_temp = data.tas print(surface_temp)
<xarray.DataArray 'tas' (time: 360, lat: 145, lon: 192)> [10022400 values with dtype=float32] Coordinates: * time (time) datetime64[ns] 1979-01-16T12:00:00 1979-02-15 ... * lat (lat) float64 -90.0 -88.75 -87.5 -86.25 -85.0 -83.75 -82.5 ... * lon (lon) float64 0.0 1.875 3.75 5.625 7.5 9.375 11.25 13.12 15.0 ... height float64 ... Attributes: standard_name: air_temperature long_name: Near-Surface Air Temperature units: K cell_methods: time: mean cell_measures: area: areacella history: 2012-02-17T05:21:51Z altered by CMOR: Treated scalar d... associated_files: baseURL: http://cmip-pcmdi.llnl.gov/CMIP5/dataLocation...
Before we can make a plot we first need to reduce the 3D data to 2D. One of the nice things about Xarray is that you can select data by a coordinate value, rather than manualy working out an index:
surface_temp_slice = surface_temp.sel(time = '1984-03') print(surface_temp_slice.time)
<xarray.DataArray 'time' (time: 1)> array(['1984-03-16T12:00:00.000000000'], dtype='datetime64[ns]') Coordinates: * time (time) datetime64[ns] 1984-03-16T12:00:00 height float64 ... Attributes: bounds: time_bnds axis: T long_name: time standard_name: time
Now that we’ve selected our data we can get plotting
The notebook requires a special
%matplotlib inline command to show plots, which isn’t needed if you’re writing a script.
To make a quick plot we can just call
.plot() on a 2d variable:
%matplotlib inline import matplotlib.pyplot as plt surface_temp_slice.plot() plt.show()