In this study, the overall carbonation depth of existing bridges was investigated to establish practical guidelines for diagnosing structural carbonation. Carbonation depth was measured by obtaining cross-sections and core samples of girder-slabs from six existing concrete bridges scheduled for dismantling. The variation in carbonation depth based on the structures' location was analyzed. Moreover, various protocols were proposed for deriving the characteristic values of in-situ carbonation depth for concrete structures and each protocol was verified through probabilistic simulations. Carbonation depth exhibited inconsistency across the cross-section, varying significantly depending on the location, but a deeper carbonation was especially noticeable in locations subjected to bending or tensile stress by design. The method to select the largest value from the core data obtained are considered suitable to prevent both underestimation and overestimation.