Trends and correlations in recent air temperature and precipitation observations across Japan (1906–2005)

Abstract

Listen

Translate article

Air temperature and precipitation data observed and recorded by the Japan Meteorological Agency (JMA) from 1906 to 2005 were analyzed. Observed trends and correlations at 16 weather stations in urban and rural areas over Japan from Nemuro in the north (43.3oN) to Ishigaki in the south (24.3 N) were identified and quantified. Average annual values were emphasized in a previous analysis (Higashino and Stefan 2014). Herein monthly values and extreme events at the daily and hourly timescale were extracted and interpreted. As shown in 2014, observed air temperature rises are much larger in urban than in rural areas and are statistically significant at all sites. Air temperature rises in Japan in the period 1906–2005 are strongest in spring (March, April, and May) and weakest in midsummer (July and August). Monthly air temperature rises have accelerated over time by a factor of about 3 from the period (1906–2005) to (1981–2005), except for a short period in August. In the period (1906–2005), daily minimum air temperature rises have occurred at all sites. Daily and yearly minimum air temperature trends are more than twice as strong as daily and yearly maximum temperature trends. The numbers of annual hot days and hot nights have also increased at most of the study sites in the period (1906–2005), but only the number of hot nights and the daily minimum air temperature are strongly correlated with annual average air temperature increases. The trends in the annual numbers of frost days and hot nights are well correlated with the annual average air temperature trend, one negative, the other positive, as to be expected. Trends in mean annual precipitation are very small, near zero for (1906–2005), mostly negative for 1956–2005, and mostly positive for 1981–2005. There is no correlation with trends in air temperature. Daily maximum precipitation for (1906–2005) and (1956–2005) shows only small and often non-statistically significant changes at the 16 study sites. In (1981–2005), trends have increased significantly but fluctuate significantly between positive and negative trends.