Our application is a global change simulation that models the carbon cycle. This model operates on dozens of parameters to calculate a measure of biological activity called net primary productivity. Data used as inputs to this model span dramatically different periods of time. To enable our application to reconcile differences in time scales, we have developed an interval - based representation of data, In this paper, we describe this representation, define operations on time intervals, and discuss how these operations affect parameters associated with time intervals. Note: Abstract extracted from PDF file via OCR