The integration of geostatistics and geologic data in the form of outcrop analogs and well core produce a detailed model of aquifer heterogeneity at the Schoolcraft study area, located in Kalamazoo County, Michigan. The sediments at this site were deposited as outwash distal to the Kalamazoo Moraine, and consist of primarily fine to medium sands with interbedded gravels and silts. Using geostatistics, the distributions of hydraulic conductivity (K) values in this aquifer were estimated. K data exists from 13 wells, six of which have been described for sedimentologic textures and features. These 13 wells occur within a 6x16-meter area. The K data used as conditioning for geostatistical simulations came from permeameter tests of core and outcrop analog samples. 5.1-cm diameter core obtained from the aquifer were cut into 15.2-cm segments. Core provides excellent information on vertical K distributions, but little about lateral variability. To obtain lateral K distributions, outcrop analog successions were analyzed for grain size and K variability. These successions were collected using 5.1-cm diameter by 15.2-cm long brass sleeves, inserted in lateral and vertical rows within certain facies. Based on this textural comparison and sedimentary structure comparisons, it was determined that the sand and gravel pits are a reasonable analog to the Schoolcraft site. The assumption of stationarity in geostatistical simulations is often difficult to honor, however, identifying sedimentologic units within each succession and confining the simulations to those units, stationarity is much more likely to be honored in each search neighborhood. Additionally, separating the data set into major sedimentologic units helps to preserve abrupt natural changes in K, which would be smoothed with traditional use of geostatistics. These major sedimentologic units were identified based on stratigraphic breaks found in well core data. Realizations of the aquifer’s heterogeneity were created by incorporating all of this information. We believe that this aquifer characterization approach can lead to more accurate groundwater flow and transport simulations.