Temperature is the most sensitive parameter in hydrocarbon generation. Thus, reconstruction of temperature history is essential when evaluating petroleum prospects.

No measurable parameter can be directly converted to paleotemperature. Maturation indices such as vitrinite reflectance, T_max from Rock-Eval pyrolysis, spore coloration, Thermal Alteration Index (TAI), or concentration of biological markers offer an indirect approach. All these indices are a function of the thermal history through rather complex kinetics, frequently influenced by the type of organic matter. Their significance and validity are reviewed. Besides the problems of identification (e.g., vitrinite) and interlaboratory calibration, it is important to simultaneously interpret kerogen type and maturation and to avoid difficult conversions from one index to another. Geodynamic models, where structural and thermal histories are connected, are another approach to temper ture reconstruction which could be calibrated against the present distribution of temperature and the present value of maturation indices.

Kinetics of kerogen decomposition controls the amount and composition of hydrocarbons generated. An empirical time-temperature index (TTI), originally introduced by Lopatin, does not allow such a quantitative evaluation. Due to several limitations (no provision for different types of kerogen and different rates of reactions, poor calibration on vitrinite reflectance), it is of limited interest unless one has no access to a desk-top computer. Kinetic models, based on a specific calibration made on actual source rock samples, can simulate the evolution of all types of organic matter and can provide a quantitative evaluation of oil and gas generated. Examples from the Jurassic source rocks of the Paris basin, Monterey Formation of California, Green River shales of Utah, Paleozoic source eds of the Algerian Sahara, and Miocene rocks of the Mahakam Delta, Indonesia, illustrate various aspects of the discussion. Geological/geochemical models are the most efficient way to integrate geological, seismic, and geochemical data, and they should greatly help to reduce the risk in exploration.