The first step needed is to bring the DEM for the Hadley area into QGIS:

1. In "files of type", choose the option "AIG GRASS and all other files". Note: In linux, this step will not be available, but you will still be able to open the file.

2. This will make a number of files appear in the window, but it is not clear which one of those files contains the actual raster values.

3. A trick we used to figure this out was to open a "File manager" program and navigate to the same "stpl_DEM" folder, displaying the details of the files.

4. The file w001001.adf, the one with the most information, contains the raster values, and the other files are accompanying files with heather/metadata information.

Note: Remember that the "projection on the fly" capability of QGIS only works for vectors and 
not rasters. A workaround is to provide all the rasters in the same projection as the project 
or data frame. You don't need to worry about the vectors, as QGIS is able to reproject them 
"on the fly".

Note: QGIS is not consistent in reading the projection information from the data when you first 
add it to the QGIS table of contents. As a general habit, the first thing you should ALWAYS 
do when you add any layer to QGIS is to define the correct projection for each vector and raster 
data set.

Reminder: To define the projection of raster and vector datasets, go to the layer properties, 
accesible when right-clicking the name of the layer.

This is how the raster layer should look!. In this color scheme, blue colors mean low elevation and red show high elevation areas:

In order to perform any analysis operation with the GRASS plugin, we need to operate on GRASS format data. That's why the next step is to convert the GRID format raster into a GRASS raster.

REMEMBER: To be able to work with GRASS tools, you  need to have an open mapset. Otherwise, 
the GRASS toolbar won't be available( see this section for a detailed explanation). 

IMPORTANT: during the import process, the GDAL and OGR import modules, by default, OVER-RIDE 
the original projection of the dataset to match it to the current mapset. In our case, the 
projection of the grid and the mapset are the same, so this is not a problem. If they were 
different, the newly created GRASS raster would have a wrongly defined projection. The correct 
way to import a raster with different projection to GRASS would be:
 A. Create a new location with the original raster's projection, 
 B. Import the raster to that location
 C. Reproject in the GRASS shell with the comand r.proj 
(GRASS shell will be explained later in this module)

Note: So?far, (due to bugs?) it doesn't always work well to click on the "view output" 
button after executing a module. In vector operations it often provides just the 
opposite to the intended function, and in this particular case, it doesn't add the 
raster properly displayed.

Displaying the DEM layer (after tuning the display to 50% transparency) over an "ASPECT" layer will enhance the appearance of the DEM to look like 3D terrain:

To get this display, you will need to follow these steps:

Slope is a measure of the steepness of the terrain. It is defined by a plane tangent to a topographic surface. Slope values will be useful for our analysis exercise, because they will reveal flat areas suitable for comercial development.

Our ultimate goal in this module is to extract the areas with slopes under 3 degrees. In order to do that we need to reclassify the slope map into two classes: "slope < 3 degrees" and "slope > 3 degrees". While we could find the flat areas with a simpler selection process, learning how to RECLASSIFY is important, as it is a very common raster processing tool.

c://grassdata/raster1/hadley

"sh: <name of the path> is a directory". 

r.reclass input=hadley_sl output=hadley_sl3   <enter>
0 thru 3 = 1 flat    <enter>
* = 0 hill      <enter>
end    <enter>

 Clipping is the action of truncating data or an image by removing all the display elements 
 outside a boundary. The area enclosed by this boundary may be referred to as a clipping 
 window. While other software packages may be able to use a polygon to clip a raster, 
 in QGIS/GRASS that is not possible. The next sections will explain how to do it instead. 

Note: In QGIS-GRASS, raster operations and functions are executed in the RASTER MAP CALCULATOR. 
For this wizard to work, there must be a MASK defined in the mapset.

We will first learn what is a mask and how to set it up:

Conceptually, a mask is an area that, when "placed" or "applied" to a raster map, will block out certain areas of a map from analysis. In QGIS-GRASS, it is a raster named "MASK" that lives in the mapset. While a mask is active or present, GRASS commands will only act on data falling inside the masked area.

While in some applications it may make sense to use the perimeter of the area of interest as the 
mask, in general we don't recommend it because any raster in the mapset will automatically be 
displayed masked or clipped to that perimeter. That may not be what we want: sometimes we need 
to see the context around the area of interest. Also, some analytical processes (like spatial 
interpolation, for example) will suffer edge effects if data around the borders are excluded 
from the analysis. In general, we recomend to consider applying a mask slightly larger than 
the area of interest, and clip the final products at the end.

Now, the mapset has a raster named "mask" with the function of masking all the other rasters in the mapset for displaying and analysis purposes, and the raster map calculator will work (you don't need to add the "mask" raster to the map though, it just needs to be in the mapset).

In QGIS-GRASS, the "Raster Map Calculator" is the environment to perform arithmetic operations and mathematical functions inside and among rasters.

Note: For the raster calculator to work, GRASS needs to have a "MASK" defined in the mapset, 
otherwise it won't perform any operations. See the previous section about defining a mask.

 "Behind the scenes" of the map calculator, the GRASS command  r.mapcalc is at work. 
 Instead of having to write more or less complicated equations, the map calculator makes it easy 
 to construct models or equations in an intuitive way without having to worry about syntax issues. 
 You just need to build a diagram connecting "datasets" to an "output" using "operations". 
 The calculator translates this diagram into a r.mapcalc equation.

Note: Once an item is selected, you can also move it inside the diagram window.

Note: It is very important to select the items before attempting to delete them. Otherwise, QGIS will crash.

After clipping the classified slope raster, we need to extract the flat areas into a polygon layer. To do this, go into GRASS tools, and select "Convert a raster to vector areas". This will take the values (1 and 0 in this case) and convert them to a polygon that covers every area that has that value. You can see from the diagram below which files to input and output (it'll be called "hadley_sl3_pl"):

Once this is complete, you want to add this to your map. Go to "Add GRASS Vector Layer" and fill in as shown below (the same process you've already done):

Now that the layer has been added, you'll want to make it distinctive so that you can see exactly where the flat areas are. Right-Click on the layer in the legend and go to properties, then click the symbology tab. Fill in as described below. What you are actually doing with these steps is assigning a set of visual characteristics to each value (in this case, 1 and 0). You need to set the symbology to understand this (by setting it to "unique value"), then define the classification field (by scrolling to "value"), and finally by clicking on the "1" (which is our denominator of flat areas) and changing the color to yellow.

Now, after all of that, we have a vector file that we can use to see the flat areas and define a new layer that contains only those areas. To start, you need to select features by attribute. This is in the GRASS tools section. Fill in the values below. As we said above, we need to know the areas marked by the value of "1", and we want to give this area a new name, which we'll call "hadley_flat":

Once you've done that, you can bring "hadley_flat" into the map by "adding a GRASS vector layer". If you've done everything right, and you turn off all the other layers, it should look like this:

Congratulations! You've created a layer that shows only the possible areas of development in the town of Hadley, based on slope.

Moving to other computer? Just copy your GRASSDATA folder and any folder where you have datasets 
that you are using in QGIS, as well as your QGIS project file. If you ?reserve the same file names 
and replicate the "paths", the QGIS project will open without any problems.