Mini-project 2 Background Information

Guidelines for Research Component 1
Uncertainty in biomass and carbon stocks in rural communities

The following questions and procedures form the basis for generating the research results that follow this section.

Questions

1. What is the size of the carbon stocks that Sr. João Melo’s family is managing? What is the relative value of this carbon?

The response to this question requires determination of the carbon stocks on a per hectare basis and then estimating the size of the family’s holdings. Brown et al. (1992) estimated the carbon stocks as being approximately 200 tons of carbon per hectare. Based on the exercise of measuring biomass on the campus, what is your estimate of the uncertainty of this carbon stock estimate? For publication purposes, please justify, using error propagation methods and prepare it as a mini-paper with a format of introduction, methods, results and discussion.

2. How much of their area is deforested? Does it exceed the 10 percent limit for the extractive reserve

The size of the family’s forested land is difficult to estimate. One conventional method is to take the number of rubber tapping trails and multiply by 100 ha. But this method will be difficult to justify, especially if there is the possibility for Sr. Melo to receive funds for conservation via Clean Development Mechanisms. In this case, you will need to develop a method that is reproducible and quantifiable. One approach would be to use GPS measurements at the limits of the area that the family considers to be their responsibility in the Chico Mendes Extractive Reserve. These measurements, if taken in UTM coordinates, can then be easily plotted on graph paper at the appropriate scale and the number of squares counted to estimate the area. The field exercise of 26 July of an imaginary farmer’s field in the Sea Aquarium parking lot was an example of such quantification.

We can use satellite imagery to facilitate this calculation and provide information as to the size of the land deforested by the family and its forested lands. This is the challenge that the residents of the Chico Mendes Extractive Reserve and other such areas face as they must use such information for managing their lands in a social environment that now includes direct participation of the Brazilian national government, responsible for the land. One publication that provides material on how residents can use satellite imagery for mapping is Alechandre et al. (1998).

To do such analysis using satellite imagery, we will use the GIS/remote sensing software Idrisi32 (for more information, see www.clarklabs.org). The following commands are to help the less experienced in using this software.

Open Start, then Programs, then Idrisi32, then Idrisi32.

Once in Idrisi32, then open File, then DataPaths to confirm that the main working folder is d:\exercise idrisi def fire jul00\ and that the resource directories are d:\exercise idrisi def fire jul00\ exercise 1\ and exercise 2\. The product of your analysis will be stored in the main working folder and the information needed is in the resource directories. Click OK.

Open Display, then Display Launcher. Check to see that the File type to be displayed is Raster Layer. Click on the box with the three dashes --- to see the options, then pick exercise 1, then click on por67 bands 345. This is a cut composite image of the 002/67 5 October 1999 image from a Landsat TM7. The band composition is blue- band 3, green-band 4, and red-band 5. [For the more adventuresome, do your own composite (see the help option for more information). Why does it appear different? Use Layer Properties for the Min/Max…settings to compare the two images. Can you make them appear the same?]

This image has green areas, tropical forest in this case, and tan/brown areas that are deforested. To make the image fill more of the screen, click on the expanding rectangle icon. The tan/brown areas are the clearings that rubber tappers make for crops and pasture. What isn’t apparent from this figure is its scale and location. Pass the mouse cursor over the image and a set of numbers appears at the bottom of the program. Down in the corner is a value, say 1:126241. This is the relative scale. The x and y values are the location of the cursor in UTM coordinates for the 19L zone. Move the cursor to the upper right corner of the image and note the UTM coordinates. Repeat in the lower left corner. The differences in the x and y coordinates give the dimensions of the image.
UTM coordinates provide a simple means of estimating size and location. Go to the composer menu on the right side and click on the Map Properties, then Map Grid. In the Increment X, type in 2000, likewise for Increment Y. Click OK. Now the image is covered by a 2km by 2km geo-referenced grid. The original house of Seu Joao’s family is at the coordinates x: 524,696 y: 8,805,087. Zoom in on where you think is the house, using the zoom window icon (by passing slowly over the icon you will be able to read the identifiers).

Return to the original size of the image, using the icon with the curved arrow within the rectangle. Everything requires some knowledge of context. Seu Joao’s family has its holdings within the Porongaba Seringal, a former rubber estate. Go to composer and Click on Add Layer, then Vector layer, followed by the ---button. A few error messages will appear related to unknown data type; don’t worry about them. Open exercise 1 and click on seringal porongaba aprox and then OK. On Symbol file, click on User-defined and the ---button. Then click on d:\idrisi\symbols\ and select uniblack, then Ok and Ok again. From this superpositioning you can visually estimate the area of this Seringal, a former rubber estate, by counting the squares, each of which cover 4 square kilometers. The decision rule for the partially filled squares is if the area of Seringal is more that half, count it fully, if less than half do not count it. How big is this seringal? Note that the boundaries are fuzzy and were created for illustration purposes. Think about how rubber tappers might define their seringal and how the federal agency responsible for the Reserve might define it differently. There are over 40 seringals in the Chico Mendes Extractive Reserve. Could this cause problems for managing the Reserve? If so, how?

Estimating the deforested area of Seu Joao’s family should ideally involve the participation of its members. For this exercise, we will imagine that such participation has occurred, and the following vector file: Deforested area Seu Joao aprox is the result of this activity. Click on Add Layer and place this vector file on the image, using uniblack as the symbol file.

To calculate area, we need a more accurate measurement then the approximation we did for the seringal. Idrisi is a little more cumbersome than other softwares and we need to convert from vector to raster format. Go to Data Entry and then Initial. Copy spatial parameters from another image. Call the output image Deforested area Seu Joao. Image to copy parameters from: por67 bands 345. Don’t forget to fill in a description of what you are doing in the output documentation. Initial value should be 0. Click on Ok. Display the raster file. What does it appear as?

Go to Reformat and Raster/Vector Conversion. Choose Polyras. Choose as vector polygon deforested area seu joao and for the Image file to be updated deforested area seu joao. Use the Cursor Inquiry Mode to find out the value of the deforested polygon. Then go to Analysis, Area and click on tabular output format, and input image: deforested area seu joao. Calculate area as hectares. How much is the deforested area? What are the sources of uncertainty in this estimate?

Seu Joao has lived at this colocacao for more than 30 years. Based on this information, do the estimates of deforestation rates in Brown et al. (1992) seem consistent with the area deforested in Seu Joao’s colocacao?

The forested area of Seu Joao’s family holding is more difficult to estimate. For the sake of this exercise we will assume that in walking with Seu Joao and his family the following GPS points show the extent of their holding. These points are located in the point vector file Extensao da colocacao seu joao.

3. How may these answers affect Sr. João’s family decisions on land use, e.g. agroforestry systems versus pasture?


Page last Updated: Tuesday, August 1, 2000 at 3:13 PM
Contact: Guillermo Podestá (gpodesta@rsmas.miami.edu),
Summer Institute Science Coordinator
Telephone:+1.305.361.4142