The emergence of remote sensing, scientific simulation, telescope scanning, and other survey technologies has dramatically enhanced our capabilities to collect spatio-temporal data. However, the explosive growth in data makes the management, analysis, and use of data difficult and expensive. In decision support applications with spatio-temporal data, it is important to study the temporal relationships of the parameters that influence the decision. Because multiple spatio-temporal data sets contain volumes of data, and often there is a delay between the occurrence of an event and its influence on the dependent variables, finding interesting patterns can be difficult.This paper presents a layered architecture for a distributed GDSS that uses temporal rule discovery to aid the decision-making process. Data mining algorithms are used to identify temporal relationships between multiple spatio-temporal data sets where time lags may exist between the related events. These algorithms allow the user to specify target events, to prune rules that are not of interest to the current decision-making problem. A geo-spatial decision support system for drought risk management is used to demonstrate the effectiveness of building knowledge discovery into a GDSS.