ABSTRACT
In the data base of the World Resources Institute, there are many variables that deal with forestry and soils: from deforestation data expressed in percentages and thousands of hectares for closed forests and total forests to soil degradation data based on 22 different indicators. Roughly, the idea is to fit pieces of cubic curves between a finite set of sample data points: one cubic is fit between two adjacent sample points and another cubic is fit between another pair of adjacent sample points. Because the curve is fit between points, and never extended beyond any sample point, it is a fit that is bounded and is useful only for interpolation--not for extrapolation. If the shade plants chosen had only a quick effect, in which case the curve would rise and then drop sharply after the initial planting, the corresponding geometric dynamics of feedback would fly out of control.
