Results from a field experiment in southern Lake Michigan are used to quantify the cross-shelf nearshore variability in Great Lakes temperatures during the stratified season. The experiment was conducted along the Indiana coast of southern Lake Michigan, with temperature and velocity moorings arranged in a cross-shelf transect that extended to approximately 20 km from shore (40 m depth). The field site is noteworthy because of its location at the end of a major axis of an elliptical Great Lake, the relatively mild bathymetric slope, and local shoreline orientation that is perpendicular relative to the dominant summer winds. Measurements demonstrate that the location of the thermocline-bottom intersection is highly variable, causing a wide zone of extreme thermal variability in the nearshore region with time scales of variability ranging from hours to months. Near-inertial internal Poincaré waves are shown to cause large thermocline excursions but primarily only during periods of elevated activity. Several full upwelling events were observed, but in general, they were brief, lasting only 1–2 days, and had very limited spatial extent (2.5 km or less). Nonetheless, the offshore extent of the upwelling front was shown to be reasonably estimated with a simple estimate of the cross-shelf transport caused by alongshore wind events. A persistent feature that determined the zone of elevated thermal variability (the thermocline-shelf intersection point) was the strongly tilted thermocline, which resulted in the thermocline being located very close to shore. No evidence was found to support the hypothesis that internal Kelvin waves affect thermal variability at the study location.