Summed radiocarbon calibrations as a population proxy: a critical evaluation using a realistic simulation approach