The use of carbonate 'clumped isotope' thermometry as a geochemical technique to determine temperature of formation of a carbonate mineral is predicated on the assumption that the mineral has attained an internal thermodynamic equilibrium. If true, then the clumped isotope signature is dependent solely upon the temperature of formation of the mineral without the need to know the isotopic or elemental composition of coeval fluids. However, anomalous signatures can arise under disequilibrium conditions that can make the estimation of temperatures uncertain by several degrees Celsius. Here we use ab initio calculations to examine the potential disequilibrium mineral signatures that may arise from the incorporation of dissolved inorganic carbon (DIC) species (predominantly aqueous carbonate and bicarbonate ions) into growing crystals without full equilibration with the crystal lattice.We explore theoretically the nature of clumping in the individual DIC species and the composite DIC pool under varying pH, salinity, temperature, and isotopic composition, and speculate about their effects upon the resultant disequilibrium clumping of the precipitates. We also calculate equilibrium clumped signatures for the carbonate minerals calcite, aragonite, and witherite. Our models indicate that each DIC species has a distinct equilibrium clumped isotope signature such that, δ47(H2CO3)>δ47HCO3->δ47(equilibrium calcite)>δ47CO32-, and predict a difference between δ47HCO3-andδ47CO32->0.033‰ at 25°C, and that δ47 (aragonite)>δ47 (calcite)>δ47 (witherite). We define the calcite clumped crossover pH as the pH at which the composite δ47 (DIC pool)=δ47 (equilibrium calcite). If disequilibrium δ47 (calcite) is misinterpreted as equilibrium δ47 (calcite), it is possible to overestimate or underestimate the growth temperature by small but significant amounts. Increases in salinity lower the clumped crossover pH and may cause larger effects. Extreme effects of pH, salinity, and temperature, such as between cold freshwater lakes at high latitudes to hot hypersaline environments, are predicted to have sizeable effects on the clumped isotope composition of aqueous DIC pools.In order to determine the most reliable and efficient modeling methods to represent aqueous dissolved inorganic carbon (DIC) species and carbonate minerals, we performed convergence and sensitivity testing on several different levels of theory. We used 4 different techniques for modeling the hydration of DIC: gas phase, implicit solvation (PCM and SMD), explicit solvation (ion with 3 water molecules) and supermolecular clusters (ion plus 21 to 32 water molecules with geometries generated by molecular dynamics). For each solvation technique, we performed sensitivity testing by combining different levels of theory (up to 8 ab initio/hybrid methods, each with up to 5 different sizes of basis sets) to understand the limits of each technique. We looked at the degree of convergence with the most complex (and accurate) models in order to select the most reliable and efficient modeling methods. The B3LYP method combined with the 6-311++G(2d,2p) basis set with supermolecular clusters worked well. © 2013 Elsevier Ltd.