Deforestation in the Amazon: Measuring forest loss in Rondônia, Brazil
1/8/10 Dr. Eric Compas (adapted from a lab by Dr. Joseph Kerski)
Amazon forests are famous for their massive extent and the high diversity of interesting and unique plants
and animals within them. More recently, climate change scientists have seen these forests as huge
reservoirs for storing carbon, buffering the planet from more extreme temperature changes. For decades,
though, these forests have served as Brazil’s frontier and an engine for the country’s economic
development. Thousands of acres have been cleared putting at risk the biodiversity within the forest and its
carbon storage capabilities.
For this project, you’ve been hired as a consultant by The Nature Conservancy, an international
conservation group, to help fix this problem. You’ll be flying to Brazil to travel up the Amazon later this year
to the Brazilian state of Rondônia, and you want to prepare by learning as much as you can about the
problems there. In this lab, you’ll be using geographic information systems (GIS) and satellite imagery to
explore and better understand deforestation in Rondônia. You’ll try to answer the following:
1. What is the annual rate of deforestation in Rondônia?
2. What are the potential drivers of deforestation there?
3. What do I need to investigate further when I arrive?
Good luck on your investigations and on your eventual trip up the Amazon and the Madeira Rivers (don’t
forget your passport and keep an eye out for the Akuntsu tribe!).
Step 1. Download data from Brazil
We’ll start our investigation by downloading data that we need from the Brazil Institute for Geography and
Statistics (the Instituto Brasileiro de Geografia e Estatistica or IGBE – don’t forget that they speak
Portuguese!). Head over to their web site at http://www.ibge.gov.br/english and click on “States” in the left
menu (open this site in Internet Explorer, other web browsers won’t work with this site). Find Rondônia and
explore information about it. Where is it? How many people live there? Is it bigger than Wisconsin? Feel
free to use Google Translator if you need to decipher some of the Portuguese.
Alright, return to the main IGBE
website and click on “Interactive
Maps” on the left menu and then
again on “Interactive Maps” on the
new page that appears (you’re a
GIS professional, right?). In the
next window, choose “Mapa de
Vegeta??o” and then click “OK.”
You’ll now be looking at the
following (may have to wait a
Let’s zoom into Rondônia. In the upper-left, select “Rondônia” under the “Zoom para UF:” menu. Next, click
on the download (“Extrair”) button. You may have to click on the “Temporarily disable pop-up” bar at the
top if Internet Explorer and click on this button again. In the new window that opens, click again on
“Extrair,” then “Download,” and save your file to our project folder (on the board). Open Windows
Explorer, go to the project folder, and unzip this file by right-clicking and selecting “Extract All...”
Now, we have the vegetation data that we need. Luckily, you’ve received the rest of the data you need
from a colleague that’s working in the same area.
Step 2. Preliminary look at data & map
Now, let’s have a look. Open
the Quantum GIS software (a
free and open-source GIS
package available at
take a moment to look at the
This is the main window in
QGIS and where we’ll explore
our data and conduct
analysis. Along the top are
the menus and tools that
you’re used to in most
software. Below that, you’ll
find the legend, or Layers, on
the left containing the data
sets that you’ll load and a Map on the right in which this data is displayed. Hold your mouse over several of
the tools on the tool bar to see the “tool tip” for each tool. I’ll be referring to each tool by this name in the
rest of this guide.
First, let’s add some data. Click on the “Add Vector Data” button to add a vector data set (data that
can be shown as points, lines, or polygons) and then click on the “Browse” button. Go to your project folder
and add the following data sets: rodovia.shp (roads), 3.shp (states), hidro.shp (rivers), and capital.shp.
Double-click on each layer name in the Legend to open options for that layer. Click on the Symbology tab to
change colors of each dataset (ask if you need help with this). You may also want to add labels under the
Labels tab. Now we have our base map to situate ourselves and explore the vegetation in Rondônia.
Next, add the veg.shp data set that you downloaded. This is a representation of current major vegetation
types. Change the symbology for this layer by going to Symbology and then choosing “Unique Values” as
Legend Type. Under “Classification Field,” choose “Tipo” (means “type”) and then click “Classify.” This will
use names with a table within this data set to color each type of vegetation differently.
Look at the resulting map. Here, you may need to pull up Google Translator again to figure out with the
vegetation labels mean. Where’s the rain forest? Where are the areas that have been cut for agriculture?
What’s the dominant vegetation cover type in Rondônia? Do you see any geographic relationships that are
apparent in our map? Do roads or rivers have any relationship to areas that have been cut?
Before you leave on your trip, you’ll want to make a map of this data. Under File->Print Composer, you’ll
find the tools to make a map (with labels, legend, and scale bar). We’ll leave this for later, though.
Step 3. Measuring change
From our vegetation map, we’re only getting a coarse look at what may be happening in Rondônia. Let’s
look at a finer scale to see what’s going on and actually measure the rate of deforestation. Open the
“Satellite Change.qgs” file in the project folder.
Within this map, you have two Landsat satellite images, one from 1990 and another from 2000. These were
downloaded from NASA at https://zulu.ssc.nasa.gov/mrsid/mrsid.pl and are “false color” images from three
of the Landsat bands (visible green is shown as blue and near-infrared is shown as green, and mid-infrared
is shown as red). The resulting image clearly show a difference between natural forests (bright green) and
clear-cut areas (light green or red).
Notice which image is on top. Are you looking at 2000 or 1990? Switch the 2000 layer off and on (clicking
on the x next to the image name) and see if you notice any changes. Are newer clear cuts visible? Zoom in
and out of particular areas to see if you notice more detailed
changes between the two dates.
Next, turn on the StudyArea layer by clicking the box to the left
of it. These are four arbitrarily chosen areas of equal size in
which we’re going to measure the change in rainforest
coverage between 1990 and 2000. You’ll be assigned a study
area (1, 2, or 3) and a year (1990 or 2000) to map, or digitize,
the rainforest area within the study area boundary for that
Here’s an example of where we’re headed:
Now, pay attention to each of the following steps – there are quite a few places to lose your way. First, let’s
create a new “shapefile” (this is a file format, like a Word document, for storing map layers) by choosing
from the menu bar Layer -> New Vector Layer… (vector means
something comprising points and lines).
Set the “Type” to Polygon (we’ll be mapping forest areas) and
add an “Attribute” called Forest of “Type” String and “Width”
of 50. Click on the small button to the right to add this
attribute. Click OK and save this file in your data folder using a
naming convention like “Area1Year1990.shp”.
Zoom into your area so that it fills your screen, click on the
name in the Layers area (it’ll have a gray background when this
layer is selected), and then click on the blue pencil (the “Toggle
Editing” tool) within the tool bar. This will allow us to
start adding and editing polygons in our new shapefile. Before
starting, make sure you’re zoomed into the right study area
and have displayed the correct Landsat image.
Next, click on the “Capture Polygon” tool to
start editing. Your cursor will turn into a cross-hair
when over the map. When you left-click, that’ll
start a new polygon and repeated left-clicking will
define the boundary of your polygon. When you’re
finished, right-click to complete your polygon. You’ll
be asked for a value for your “Forest” attribute –
you can leave this as “NULL” and click OK. You’ve
created your first forest polygon. Your goal is to
create polygons for all of the dark green rainforest
in your study area. Here’s a brief guide for us to be
consistent across the study areas:
1. If it’s red, pink, or white, don’t include it. It’s not rainforest.
2. If it’s a small isolated bit of green, don’t include it. If it’s still rainforest, it doesn’t function
biologically as such.
3. If there’s an island of pink within a green area, include it for the moment. We’ll show you how to
exclude it below.
4. Along the edge of your study area, digitize along the boundary line as best as you’re able.
5. If you make a small mistake, keep going. We can correct it later.
6. If you make a big mistake, ask for help.
7. While we want an accurate area, we don’t have unlimited time before your trip. Don’t be too
Next, click on the blue pencil again, and click “Save” to save your edits (you may want to do this several
times while editing).
Now, we’ll make a few refinements. Double-click your layer in the Layer window to bring up the layer’s
properties and click on the Symbology tab.
Change the color of your forest layer to white and set the Transparency to 50% and click OK. You’ll now be
able to see through your polygons to the Landsat image below.
Click on the blue pencil again to start editing again. Use the move, add, and delete vertex tools (
) to add and delete vertices along your polygon boundary if you need to adjust anything.
To make islands (where there’s a red area inside one of your forest polygons), click on the Ring Tool
and then click on points to define the island (and right-click to finish). Once you’re finished with your
editing, click on the blue pencil one last time to save your edits (don’t forget!).
Whew! That took a while. Now for the important step: how much of the study area is still rainforest? We
need to calculate the total area of the polygons we just created.
Go to the menu Tools -> Geometry Tools -> Export/Add geometry columns. In the window, make sure your
forest polygon layer is selected, e.g. Area1Year1990, and then click Browse to tell QGIS where to save the
new shapefile. Call it something like Area1Year1990_Area.shp. Then click OK to run the tool.
Now, go to Tools->Analysis Tools->Basic statistics. In the tool window, select your new area shapefile, e.g.
Area1Year1990_Area and under “Target field” select Area and click OK. You should now see the number
we’re looking for, the “Sum” value. What units do you think this value is in? Hint: the unit our data is in is
meters. Multiply your sum by 0.0001 to convert it to hectares and add it the table below.
Table 1. Deforestation measurements for three study areas, 1990 to 2000
Study area Area (ha) in 1990 Area (ha) in 2000 Change 1990-2000
• What concerns do you have about the pattern of deforestation? How valuable to you think the
remaining forest areas are as plant and animal habitat?
• At this rate of deforestation, how long would the Brazilian Amazon rainforest last?
Step 4. Exploring drivers
We’re probably out of time, but if not, let’s look at one last data set. Zoom back to where you can see your
full Landsat image and then click on the Add Vector Layer button to add a new file. Find “RondoniaPA.shp”
and add it to your map (you may need to change its color and make it partially transparent). “PA” stands for
“protected area” – these are the equivalent of national parks within Brazil. See any patterns? Are projected
areas keeping forests from being cut?
• Briefly research Brazilian beef exports. Any potential role in Amazonian deforestation? What about
other agricultural crops?
• Are protected areas “saving” the rainforest? How might they accelerate or drive deforestation?
• Are there any lessons from Wisconsin for the Amazon? What happened to our forests? What are
we doing with them now? (Quick look at the Menominee Forest in Google Earth)
1. Area of Rondônia: 237,576 sq km. Area of Wisconsin: 145,436 sq km. Rondônia is 1.6x the size of
Put on board
1. Folder for them to work in and save their files in.
Área Antropizada Anthropic (human) area
Floresta Estacional Semidecidual (Seasonal) semideciduous forest
Floresta Ombrófila Ombrophilous (rain) forest
Floresta Ombrófila Aberta Open (less dense) rain forest
Floresta Ombrófila Densa Dense rain forest
Vegetação com Influência Fluvial ou Lacustre Riparian vegetation
Ombrophilous = tolerant of wet conditions