Aircraft contrails can persist for hours and spread into contrail cirrus, thereby contributing to aviation's climate impact. Reducing this impact through aircraft rerouting requires skillful predictions of contrail formation, persistence, and impact. Evaluating these predictions is difficult because discrepancies with observations can arise from uncertain meteorological inputs, observational errors, and the limitations of contrail models themselves. This work locates aircraft carrying in-situ instruments in high-resolution satellite imagery to observe contrail formation and early evolution. By combining these observations with meteorological measurements from the aircraft, the resulting dataset is used to evaluate contrail models while reducing uncertainty associated with meteorological inputs and observational constraints.