Paleoaltimetry reconstructions of the Central Andes based on stable isotope data indicate variable timing and rates of uplift, reflecting the operation of multiple geodynamic mechanisms. There are a range of isotopic methods and proxies involved in these reconstructions, and it may be that biases inherent in some of these methods has led to misinterpretations. Here we explored the limitations of carbonate archives in reconstructing past elevations, including underestimation of paleoelevations by oxygen-isotope evaporative enrichment and carbonate formation in the subsurface. To conduct our test, we obtained stable isotope values of carbonates from the Western Cordillera in northern Chile at ~18.5–19.5°S, where previously published paleoaltimetry results from the same sections, based on the hydrogen stable isotope composition of volcanic glass, indicate that this region of the Andes has remained at high elevations (>4.4 km) since at least ~23 Ma. We present new oxygen and carbon isotopic, and clumped isotopic results, petrographic and sedimentologic observations, as well as major element abundance (ICP-MS and microprobe analyses) from lacustrine carbonates (marlstones) and cements in sandstones of the early Miocene Chucal Formation in the central Andes of northern Chile. The clumped isotope paleotemperature estimates range from 26 to 45 °C in lacustrine carbonates and 27 to 54 °C in cements. These results indicate that carbonates from this section do not reflect surface temperatures, suggesting clumped isotope resetting in a rock-buffered setting for marlstones and carbonate formation at depth in the presence of groundwater for cements in sandstones. In contrast, the oxygen and carbon isotopic values from carbonates appear to be primary and combined with Mg/Ca ratios and Na concentrations document the transition from lake deposition into a dominantly fluvial environment within the Chucal Formation, also indicating variable evaporative conditions during deposition. The paleoaltimetry estimates in carbonates significantly underestimate the results determined in volcanic glass, with paleoelevation results ranging between 1.4 km and 2.6 km in marlstones and from 2.6 and 3.6 km in cements.. The results presented in this study highlight the importance of using multiple proxies for paleoaltimetry reconstructions, especially in arid land settings where meteoric waters have been modified by evaporation.