Abstract. In this study, we have analyzed extreme daily precipitation during the pre-industrial period from 1501 BCE to 1849 CE in simulations from the Community Earth System Model version 1.2.2. A peak-over-threshold (POT) extreme value analysis is employed to examine characteristics of extreme precipitation and to identify connections of extreme precipitation with the external forcing and with internal modes of variability. The POT analysis shows that extreme precipitation with similar statistical characteristics, i.e., the probability density distributions, tends to spatially cluster. There are differences in the distribution of extreme precipitation between the Pacific and Atlantic sectors, and between the northern-high and southern-low latitudes. Extreme precipitation during the pre-industrial period is largely influenced by the modes of internal variability, such as El Niño-Southern Oscillation (ENSO), the Pacific North American and Pacific South American patterns among others. In general, the modes of variability exhibit significant connection to extreme precipitation in the vicinity to their regions of action. The exception is ENSO which shows more widespread influence on extreme precipitation across the Earth. Effects of the changes in the orbital parameters on extreme precipitation are rather weak compared to those of the modes of internal variability. Still, some regions in central Africa, southern Asia and the tropical Atlantic ocean present statistically significant associations with the changes in the orbital parameters, implying that in these regions, extreme precipitation has increased linearly during the 3351 year pre-industrial period. Tropical volcanic eruptions affect extreme precipitation more clearly on the short term up to a few years. Although, eruptions alter both the intensity and frequency of extreme precipitation, more apparent changes are found in the frequency of extreme precipitation. After eruptions, the return periods of extreme precipitation increase over the extratropical regions and the tropical Pacific, while a decrease is found in other regions. The post-eruption frequency changes of extreme precipitation are associated with ENSO, which itself is influenced by tropical eruptions. Overall, the results show that climate simulations are useful to complement the information on the pre-industrial extreme precipitation, as it can elucidate statistical characteristics and long-term connections of extreme events with natural variability.