Intro to GRASS workshop

Introduction to GRASS GIS

This document was first written to accompany a workshop to be given at Carleton University on 17 October, 2011, in the Department of Geography and Environmental Studies' main GIS lab (Loeb A200). Anyone is welcome to adapt it for other purposes, or to help improve and keep this tutorial up to date (see CC license conditions at bottom left of page).

Getting Started

If GRASS is not already installed on a machine for you, getting the software is probably your first step. This workshop will use the OSGEO Live project's bootable test environment, version 5.0, to showcase up to date versions of all of the software without requiring any changes to the machine it is running on. The Live DVD should already be booted for you when you arrive at the workshop, and you are welcome to take the disc with you when you leave. The software running when you boot the disc is based on an Ubuntu Linux distribution. The disc also contains installers for GRASS (and other software) for Windows and Mac OS environments.

Other ways to get GRASS:

• Download binaries: winGRASS 6.4.2 snapshot full installer or LinuxGRASS 6.4.2 snapshot (64bit) installer or 6.4.1. installer
• Download source code: GRASS 6.4.2svn from SVN for your own compilation + Compile and Install instructions

In the OSGEO Live environment, launch GRASS by going to the Geospatial menu icon near the top left of the screen, then choose the Desktop GIS sub-menu, then choose GRASS GIS. This will launch a shell script terminal and a graphical interface to choose your GRASS database, location and mapset. These concepts tend to confuse new users, but once you understand the logic used here, file maintenance is taken care of for you to a large degree.

The GRASS project has a good introduction to this opening screen on its website, so to avoid reinventing the wheel, please click on this link to bring up the relevant web page. We'll go through this material step by step in the workshop. If you are using this tutorial outside of the interactive workshop, you should be able to piece together the material by reading the full web page online.

Although there are sample databases provided on the DVD, I think most people wanting to try out GRASS will soon want to get their own data into a GRASS database, so today we'll start with making a fresh database location from scratch. Confirm that the GIS Data Directory is set to /home/user/grassdata, then click on the Location wizard button in the Manage section of the window (identified by number 4 on the GRASS introductory web site).

In the resulting window, fill in Ottawa as the location, and something like City of Ottawa for the Location Title. Click Next.

In the next window, choose "Read projection and datum terms from a georeferenced data file" and click Next.

For the first steps of this tutorial, we are using data distributed by the City of Ottawa, under a license that allows free redistribution and use under the Terms of Use linked here. These files are already retrieved and made easily available for you in the lab (details provided in the workshop), but if you are performing this tutorial on your own, an archive of the data we are using is available here: [GRASS workshop data distribution].

In the window asking you to Select georeferenced file, press the Browse button and navigate to the directory where the data is stored (instructions provided in the lab if you are present for the workshop; wherever you extracted the data on your own machine if you are doing this independently). Choose the roadways.shp file. Click Next, then if everything looks right in the Summary screen, click Finish. You will be asked if you want to get the default region extents and resolution - answer Yes.

In the resulting screen, a default extent has been taken from the maximum extents of the roadways data. The resolution parameters do not apply to vector data, however, therefore they are filled with nonsense values. To make things easier later on when we introduce raster data, set both the N-S and E-W resolution to 20m, then press the Set region button.

GRASS uses the concept of a "Window" on your data to define the spatial extent and, for raster processing, the spatial resolution at which all processing is done. This window can be changed at any time, and any data layer can have a different spatial extent and resolution. The Window defines how input layers will be sampled (rasters are resampled at the current window's resolution when data layers have different resolutions). We have just set the default Window for this dataset, which provides an easy way to reset to starting conditions while we're doing GIS processing.

Now you'll be back at the initial GRASS welcome screen, and the new Ottawa location should show up and be selected in the left-most column, with the PERMANENT mapset selected under Accessible mapsets. All locations have at least one mapset, called PERMANENT. Other mapsets can be created to separate out maps belonging to different sub-projects, or to separate maps being created by multiple users sharing the same database. For now, we'll stick with the PERMANENT mapset.

Click the Start GRASS button. The terminal window that was already launched behind the welcome screen will display some introductory text and a GRASS shell prompt (for interactive command-line use), and the GRASS GUI will launch (on these and most modern installations of GRASS, this means the wxPython-based mapping interface).

Importing Vector Data

At this point, we have created an empty database - although you pointed at a shapefile to define its spatial attributes, there aren't any data loaded yet. We will import some shapefile from the City of Ottawa GIS data using the OGR-based vector import tool, v.in.ogr, which is accessed in the GUI by selecting the File menu in the GRASS GIS Layer Manager window, then choosing Import vector data | Common import formats.

In the resulting window, in the Source name section, click on the Browse button, navigate to where your City of Ottawa shapefiles are stored, select roadways.shp, and press OK. In the Import vector data window, press the Import button at the bottom. If all goes well, the shapefile will be imported into the GRASS database, and will show up as a layer in the mapping windows.

Now we'll import the rest of the City's vector data, but this time we'll use a shortcut. Bring up the Import vector data tool again (File|Import vector data|Common import formats), but this time under Source type, choose Directory. Now in the next section, navigate to the DIRECTORY where your files are stored (select the parent directory, then press Open). When you return to the Import vector data window, all of the OGR-readable data that it found in that directory should show up (in this case, all the shapefiles). Deselect the roadways layer (because we've already imported it), then press the Import button at the bottom of the window.

The data in those shapefiles are now all stored (separately) in the GRASS database, in GRASS's native data format. The attribute data is, by default, stored in a DBF-format file, just as is used with shapefiles. GRASS can optionally connect to external database servers to hold attribute data, and (with compatible databases) spatial data as well.

Controlling the map display

map controls

Importing raster data

use r.in.gdal to import a landsat scene

== Analysis example: buffering

buffer one of the vector layers

== Analysis example: interpolation

provide 3-d points; interpolate; visualize with regular map, maybe nviz

Other resources

include books, websites