# Gridded Data Renderer¶

PAVICS uses ncWMS2 to render netCDF data on a map canvas. ncWMS2 use the OGC’s WMS standard to serve images rendered from a netCDF file. The WMS GetMap operation specifies a layer (/), styles, figure size and format, projection and color range and the server returns an image that can be displayed in a figure or a map canvas.

Note that with the current configuration, the path for ncWMS layers starts with outputs and not birdhouse as in the THREDDS server. Note also that the image we obtain below has the North down.

In [1]:

from owslib.wms import WebMapService
from IPython.display import Image

url = "https://pavics.ouranos.ca/twitcher/ows/proxy/ncWMS2/wms"
wms = WebMapService(url, version='1.3.0')
resp = wms.getmap(
format='image/png',
colorscalerange='{},{}'.format(250, 350),
size=[256,256],
srs='CRS:84',
bbox=(150, 30, 250, 80),
time='2006-02',
transparent=True)

Out[1]:


One issue with the figure above is the colorscale range and the colormap, which do not provide a lot of contrast. In PAVICS, the data crawler while parsing the netCDF metadata also computes the min and max, so that the frontend knows for a given file the data range and can include it in its WMS request. Here, we’ll do the same using the OPeNDAP protocol.

In [2]:

import xarray as xr
dap = 'https://pavics.ouranos.ca/twitcher/ows/proxy/thredds/dodsC/'
ncfile = 'birdhouse/testdata/flyingpigeon/cmip5/tasmax_Amon_MPI-ESM-MR_rcp45_r1i1p1_200601-200612.nc'
ds = xr.open_dataset(dap+ncfile)
subtas = ds.tasmax.sel(time=slice('2006-02-01', '2006-03-01'), lon=slice(188,330), lat=slice(6, 70))
mn, mx = subtas.min().values.tolist(), subtas.max().values.tolist()
mn, mx

Out[2]:

(241.01828002929688, 308.3053894042969)


Now we’ll simply pass those min/max to getmap with the colorscalerange parameter, and change the palette in the same go using the styles parameter.

In [3]:

resp = wms.getmap(
styles=['default/div-Spectral'],
format='image/png',
colorscalerange='{},{}'.format(mn, mx),
size=[256,256],
srs='CRS:84',
bbox=(150, 30, 250, 80),
time='2006-02',
transparent=True)

Out[3]: