GIS for Environmental Management

Using NetCDF data within ArcGIS

Objective:

This practical introduces the importing and use of Network Common Data
Form (netCDF) data within ArcGIS. The NetCDF format is commonly used in
climatology and oceanography as a means of handling spatio-temporal data
sets, particularly those in raster format.

Data:

To access the data for this exercise, go to the University of East Anglia
Climate Research Unit’s web site at:
[Link]

We will download a simple data set here that contains global average monthly
temperatures for the period 1961-1990, measured in degrees centigrade. The
data are labelled as ‘absolute’ in the file. There are thus 12 time periods and
the data are for grid squares that are 5 degrees of latitude by 5 degrees of
longitude.

To access the data, under ‘data for downloading’ on this page, scroll down to
the ‘absolute’ data set and choose the NetCDF format. Save the file to your
computer:

Practical instructions

1
 GIS for Environmental Management

Start up ArcGIS Pro with the map template. On the map menu ribbon, click
on add data, then choose multidimensional raster layer:

In ArcGIS Pro, formats such as NetCDF are known as multi-dimensional

rasters. They differ from multi-band rasters (i.e. rasters where multiple
values can be stored in each pixel) since the format is extensible can handle
more than one additional dimension. For example, a single file could be used
to store temperature and humidity (two variables) in raster format (x and y)
for different days (time) at different heights in the atmosphere (z – making
four dimensions in total). A multi-band raster would be able to handle one
dimension (e.g. time or height), but not all dimensions simultaneously.

Choose the NetCDF file that you just downloaded (with its .nc extension) and
you should be able to select tem as the variable to work with:

2
 GIS for Environmental Management

Note that this layer has mean monthly temperature in degrees centigrade,
multipled by 100, so for example 3000 = 30 degrees centigrade. You should
see January’s temperatures displayed by default. If you highlight the layer in
the left-hand table of contents panel, you should be able to click on the multi-
dimensional menu.

Because the layer consists of multiple rasters depicting temperature, we can

animate the layer via this button

3
 GIS for Environmental Management

When you press this button, you should see ArcGIS Pro cycle through the 12
months of the year, displaying temperatures for each one.

We can also use options on this ribbon to analyse our NetCDF file. However,
to do this, first you will need to activate the Spatial Analyst and / or Image
Analyst extensions. To do this, head for the project menu, then licensing and
press the configure licencing options button, checking both of these
extensions.

Next, select your NetCDF layer once more, then head for the multidimensional
menu again. If you choose Summary Statistics, you should be able to select
Range:

Running this option should calculate annual temperature ranges (maximum –

minimum monthly temperature), with continental and polar areas showing the
greatest temperature ranges. This will be added to your display as a new
raster layer.

Some other brief notes on functionality on this toolbar are as follows:

• Via the data management button on the ribbon, there are also options
to Make multidimensional raster layer and subset the NetCDF file.
Subset enables you to save just some of the time slices in our file to a
new multidimensional raster (e.g. just October-December’s
temperature). Make multi-dimensional raster layer will save our
NetCDF file in ESRI’s own multi-dimensional raster format, enabling its
subsequent use in various NetCDF tools. Try running this option now
with the default settings and saving your NetCDF file in ESRI format.

4
 GIS for Environmental Management

• To the right of summary statistics, there are also options for calculating
anomalies (e.g. difference between mean monthly temperature for
each month) and trends from our multi-dimensional raster.
• The temporal profile button will not work with this particular NetCDF
file. This is because the file contains average temperatures, rather
than a time series of temperatures for particular months or days (e.g.
Jan 1980 to March 2020), so the ‘time’ dimension is a number (1 to 12)
rather than a date. If the time dimension was formatted as a date
rather than a number, it would be possible to right-click on our layer,
choose properties and then time, and indicate that it had a time
dimension. We could then use temporal profile to produce graphs of
temperature changes over time by digitising a region of interest or
point on top of our multi-dimensional raster.

Converting NetCDF to other formats

It is also possible to extract data from the NetCDF layer for other forms of
analysis. For example, if you head for the analysis menu and choose tools,
you should be able to search in the geoprocessing panel for ‘NetCDF’. We
can use the tool ‘Make NetCDF raster layer’ to extract one month’s data from
our multi-dimensional raster and convert it to an ordinary raster:

5
 GIS for Environmental Management

Select the file that you just downloaded. Note that ArcGIS will automatically
choose tem (temperature) as the variable to be contained in this data set,
lon (longitude) as the X dimension, and Lat (latitude) as the Y dimension.
Under ‘dimension values’, select time to have time (month of the year)
loaded as the third dimension for this data set. Note that in the grid at the
foot of this dialog box, you can optionally select a value between 1 and 12
(i.e. January through December) to indicate the month’s temperature data
that will be displayed by default in map format within ArcGIS. If you leave
this blank, by default January’s data will be displayed as the first in the series.
You can also enter in the name of an output raster layer to contain the output.
Press OK, and ArcGIS will generate a multi-dimensional raster based on the
HDF data.

• We can also use a feature known as ‘table views’ to extract particular

values from our space-time grid in tabular format. What the table view
provides is a ‘slice’ through our large array of temperature values in a
tabular format that could be downloaded to a spreadsheet, for example.
To do this:
o In the geoprocessing panel, search for ‘NetCDF’ and then select
‘Make NetCDF Table View’ and select your netCDF file that you
downloaded under ‘Input netCDF file’.

6
 GIS for Environmental Management

o Under ‘variables’, select ‘tem’ (temperature).

o Under ‘row dimensions’, select ‘lat’ (latitude) and ‘lon’
(longitude).
o Under ‘dimension values’, select ‘time’. In the grid beneath this
option, you can pick out particular time periods that you want to
look at in more detail.
o You can also choose a name for the ‘output table view’ that will
be generated.
o Click on OK and ArcGIS Pro will generate a new table view
which has data for temperature for your chosen period, with
each row representing the temperature values for a given pixel.
o Note that in order to see the output, head for the bottom left of
your screen to find your table view, you can right-click on it and
choose open to view the result.

Having created this table, we could use a tool like Table to Excel to send
these data out to an Excel file for analysis, should we wish.

Summary:

7
 GIS for Environmental Management

This practical has summarised some of the ways of manipulating data stored
in NetCDF format, a format for handling spatio-temporal data. The tools
enable us to view ‘slices’ of information from such files, either visually on-
screen in the form of maps or in tabular format, and to extract subsets of the
information within the file for further analysis. There are other tools available
for analysing files, such as mechanisms for animating the data for different
time periods.

In this case, the file is very small – the grid resolution is very coarse, there
are few time periods, and there is only a single attribute. The power of this
format and the tools for slicing through it comes where the file has many
attributes (e.g. abundance data for many different species, with each species
held as a separate attribute), and finer spatial and temporal resolution.