A geostatistical approach was used to study temporal structures in a time series of discharge and electrical conductivity (EC) in 15 karst springs from the Zagros mountain range, Iran. Two types of temporal behaviors, a periodic structure and nugget effect, plus one or two temporal structures, were identified and interpreted. These correspond to characteristics of karst systems, such as the catchment area, percent of conduit flow, and general degree of karst development. Springs were grouped into three categories based on their ranges (e.g., residence time) obtained by variogram analysis. The first group of springs include those that present the same temporal behaviour in variograms of discharge and EC. These springs are characterized by generally constant EC with increasing discharge suggesting the existence of a large underground reservoir. The second group of springs are those with varying temporal periodic behavior in variograms of discharge and EC. Positive correlation between discharge and EC values is the main characteristic of these springs and is interpreted to result from a piston-flow system in poorly developed karst aquifers. The third group of springs includes those that exhibit different temporal behaviors when compared with the periodic and non-periodic variograms. This group exhibits a negative correlation in scatterplots of discharge versus EC values suggesting a well-developed solution-conduit system that facilitates rapid response of the karst system to precipitation events. This study's results document the role of variogram analysis in delineating temporal structures of spring behaviors by means of time series of discharge and EC. Variogram analysis can be considered as a valuable tool for hydrogeological investigations in karstic terranes.