Bandgap optimization of the tandem cells has been calculated by many scientists under the standard air mass 1.5 spectrum or the site-specific conditions using statistics of meteorological observations and spectrum data. Calculation of the optimized bandgap combination at the specific site required many annually-observed atmospheric parameters. It is not realistic, because most of the installation site does not have detailed annual spectrum information. This paper has two purposes. One is to quantitatively understand the impact of the variance of the atmospheric conditions to the bandgap optimization design. Another is to propose the practical method to take the spectrum information of the specific site by even in a single day's measurement. For evaluating the impact of the atmospheric conditions, both the best and the worst distributions were taken from the measurement data. With including other meteorological and geographical data variance, the dataset of the annual spectrum with different sun-height was generated and distribution of the calculated optimized bandgap combination was examined. The optimum bandgaps varied by aerosol depth, water precipitation, latitude, and seasonal sunshine duration change as it was expected. It was also found that the optimization of the bandgaps combination at the specific site, considering ever-changing spectrum, may be determined by the matching condition to the sun height at the culmination on the winter solstice, because the distributions of optimum bandgap combinations were much closed to those of annual integration, in spite of the fact that they were calculated by a single time (single date) of the measurement. Different from AM1.5 the sun height can be taken two times a day throughout a year. (C) 2017 Elsevier Ltd. All rights reserved.