The Wonderful World of IDRISI!

Please meet IDRISI, a "highly sophisticated [computer] system... [that] is very easy to use." Whenever any computer program claims to be "easy to use" red flags should imediately pop up. But so far IDRISI has been a very been a good friend and hasn't thrown too many curve balls my way, but our friendship is quite new and I am only beginning to understand his personality, likes and dislikes. I have waited a long time to meet IDRISI - there was a lot of preparation that led up to our introduction, but we had a good first meeting when we were introduced by my IMAZON mentor, Amintas.

IDRISI is a program specializing in land change modeling, time series analysis, multi-criteria and multi-objective decision support, uncertainty and risk analysis, simulation modeling, surface interpolation, and statistical characterization. So IDRISI has a lot of tricks up his sleeve, but our friendship will be based on multi-criteria decision support analyis.

Multicriteria decision analysis (MCDA) aims to support policy/decision makers who face many difficult decisions that can sometimes generate conflicting results. MCDA highlights these conflicts and uses a transparent method to derive a new solution.

So... here is an example of how IDRISI can help decision makers analyze policy problems. Consider a new residential development is slated to be built in Westborough, New York. Even though it sounds like it would be a great idea to build more houses in the community, there are concerns that have been brought to the at the town meeting.

1. "I want my kids to be able to grow up in a community close to the center of town where the schools are located" (Mrs. Long, concerned mother and wife).
2. "There are wetlands and streams in Westborough and building next to them or on top of them will severly impact the ecosystems and the endangered species of salamander that live in the area" (Matt Green, an environmental activist).
3. "I cannot build on steep slopes. So just because there are sensitive ecosytems in the area does not mean that you can push my development into the hills. Wetlands are great to build on because they are flat" (Joe Smith, Smith's Construction Company).
4. "This development is vital to the town of Westborough and will bring revenue to the town in the form of increased sales and property taxes" (Ann Apple, member of the Town Council).

So with these concerns at hand, IDRISI can be put to work. The variables used to help make this decision easier will be: slope, distance from roads, distance from town center, distance from streams and wetlands.

First, we must understand the behavior of the variables: are the relationship linear, sigmoidal, etc? Below are the variable slope and distance from town center are analyzed. Slope displays a sigmoidal relationship (low slopes/ "flat land" are much more preferable to steep land, yet there are some areas that are inbetween that are not totally undesirable). On the other hand, the desirability of the distance to town center can be best understood as a linear relationship (the farther the distance to the town center the more undesirable the building location).

The next step is to generate factor images for each variable (distance from roads -shown, distance to town center, slope of land -shown, wetlands/rivers). To better use the variables each variable must be normalized (percents, distances, degrees, etc. need to be evaluated equally using the same units). Each variable is transformed to a scale from 0 to 255, as shown below.

After, each variable has been normalized the data can be evaluated and given different weights depending on their relative importance. The final map for Westborough shows areas of high desirability as deep reds.

So this example is just a preview of what my analysis will be like for IMAZON. But for my work at Imazon I will be analyzing the best locations for CDM reforestation and afforestation projects in the county of Paragominas.