Abstract. Transient greenhouse warming events in the Paleocene and Eocene were associated with the addition of isotopically light carbon to the exogenic atmosphere–ocean carbon system, leading to substantial environmental and biotic change. The magnitude of an accompanying carbon isotope excursion (CIE) can be used to constrain both the sources and amounts of carbon released during an event and also to correlate marine and terrestrial records with high precision. The Paleocene–Eocene Thermal Maximum (PETM) is well documented, but CIE records for the subsequent warming events are still rare, especially from the terrestrial realm.Here, we provide new paleosol carbonate CIE records for two of the smaller hyperthermal events, I1 and I2, as well as two additional records of Eocene Thermal Maximum 2 (ETM2) and H2 in the Bighorn Basin, Wyoming, USA. Stratigraphic comparison of this expanded, high-resolution terrestrial carbon isotope history to the deep-sea benthic foraminiferal isotope records from Ocean Drilling Program (ODP) sites 1262 and 1263, Walvis Ridge, in the southern Atlantic Ocean corroborates the idea that the Bighorn Basin fluvial sediments record global atmospheric change. The ∼ 34 m thicknesses of the eccentricity-driven hyperthermals in these archives corroborate precession forcing of the ∼ 7 m thick fluvial overbank–avulsion sedimentary cycles. Using bulk-oxide mean-annual-precipitation reconstructions, we find soil moisture contents during the four younger hyperthermals that are similar to or only slightly wetter than the background, in contrast with soil drying observed during the PETM using the same proxy, sediments, and plant fossils.The magnitude of the CIEs in soil carbonate for the four smaller, post-PETM events scale nearly linearly with the equivalent event magnitudes documented in marine records. In contrast, the magnitude of the PETM terrestrial CIE is at least 5 ‰ smaller than expected based on extrapolation of the scaling relationship established from the smaller events. We evaluate the potential for recently documented, nonlinear effects of pCO2 on plant photosynthetic C-isotope fractionation to explain this scaling discrepancy. We find that the PETM anomaly can be explained only if background pCO2 was at least 50 % lower during most of the post-PETM events than prior to the PETM. Although not inconsistent with other pCO2 proxy data for the time interval, this would require declining pCO2 across an interval of global warming. A more likely explanation of the PETM CIE anomaly in pedogenic carbonate is that other environmental or biogeochemical factors influencing the terrestrial CIE magnitudes were not similar in nature or proportional to event size across all of the hyperthermals. We suggest that contrasting regional hydroclimatic change between the PETM and subsequent events, in line with our soil proxy records, may have modulated the expression of the global CIEs in the Bighorn Basin soil carbonate records.