Frequent monitoring of temperature (FMT) for over 1 year at two aquaculture sites in the western Baja California peninsula was analysed in terms of hourly, daily and monthly variability, and with this information, temperature-change indices were calculated. These data were contrasted against a long-term series from a global database (Extended Reconstruction of Sea Surface Temperature (ERSST)) to evaluate whether these could substitute for FMT. The compatibility of species requirements with the thermal conditions was evaluated by comparing the temperature frequency distributions from the two FMTs, with the optimum and lethal temperature information available on five bivalve species of aquacultural interest. We concluded that there was no correlation between ERSST and FMT because the former underestimates the amplitude of real temperature fluctuations and exhibits a different pattern of variation during the year. Therefore, FMT was needed for a correct selection of an aquaculture site for bivalves. The FMT indicated high temperature variability at both sites studied on different time scales, with the site located at lower latitude (Rancho Bueno) warmer and with a higher variability than Laguna Manuela. Contrasting these results with optimum and lethal temperature values of bivalve species, it was possible to find the ideal site, for temperature, for culturing the species, taking into account the variability associated with large-scale phenomena.