This study investigated the relationship between heat-related illnesses obtained from healthcare big data and daily maximum temperature observed in seven metropolitan cities in summer during 2013~2015. We found a statistically significant positive correlation (r = 0.4~0.6) between daily maximum temperature and number of the heat-related patients from Pearson's correlation analyses. A time lag effect was not observed. Relative Risk (RR) analysis using the Generalized Additive Model (GAM) showed that the RR of heat-related illness increased with increasing threshold temperature (maximum RR = 1.21). A comparison of the RRs of the seven cities, showed that the values were significantly different by geographical location of the city and had different variations for different threshold temperatures. The RRs for elderly people were clearly higher than those for the all-age group. Especially, a maximum value of 1.83 was calculated at the threshold temperature of 35℃ in Seoul. In addition, relatively higher RRs were found for inland cities (Seoul, Gwangju, Daegu, and Daejeon), which had a high frequency of heat waves. These results demonstrate the significant risk of heat-related illness associated with increasing daily maximum temperature and the difference in adaptation ability to heat wave for each city, which could help improve the heat wave advisory and warning system.

The interannual variability of summer temperature during June-August (JJA) in South Korea was associated with geopotential height averaged in the East Sea (Korea-Japan Index, KJI) and in the subtropical western North Pacific (Western North Pacific Subtropical High Index, WNPSHI). The KJI was coupled with a decaying El Niño one month in advance, while the WNPSHI was influenced by Sea Surface Temperature (SST) anomaly in the western North Pacific and a developing El Niño one to three months ahead. Additionally, the JJA temperature over South Korea was affected by SST anomaly in the western North Pacific in May. Based on these teleconnections, a multivariate regression model using the SST surrogates for the KJI and WNPSHI and an univariate model using an area-averaged May SST were developed to reconstruct the JJA temperature over South Korea. Both of the empirical models reproduced the JJA and monthly temperatures reasonably well. However, when the simulated SSTs from global climate models were used, the multivariate model outperformed the univariate model. Further, for JJA temperature prediction, the multivariate model with 6-month lead SST outstripped one-month lead prediction of global climate models. Therefore, the empirical-dynamical approach can pave a promising way for summer temperature prediction in South Korea.

Persistent Extreme Temperature Events (PETEs) are defined in two steps; first, to define extreme temperature events, the 80th and 20th percentiles of daily maximum and minimum temperature were chosen. Then individual PETE was defined as an event which lasted three or longer consecutive extreme temperature days. In this study, we examined characteristics and changes of PETEs in Republic of Korea (ROK) using 14 weather stations with a relatively long-term period of data, 1954-2016. In ROK, PETEs lasted four-five days on average and occurred two-three times a year. PETEs lasted longer in summer than in winter and in maximum temperature than in minimum temperature. PETEs which lasted greater than seven days account for a greater proportion in summer than in winter. However, intensities of PETEs were greater in winter because of a larger temperature fluctuation. In both summer and winter, durations and intensities of persistent extreme high temperature events increased while those of persistent extreme low temperature events decreased. Changes of PETEs were closely related with both global warming and diverse large-scale climate variabilities such as AO, NAO and Nino 3.4.

In order to investigate the effect of air temperature reduction on an urban neighborhood park, air temperature data from five inside locations (forest, pine tree, lawn, brick and pergola) depending on surface types and three outside locations (Suwon, Maetan and Kwonsun) depending on urban forms were collected during the summer 2016 and compared. The forest location had the lowest mean air temperature amongst all locations sampled, though the mean difference between this and the other four locations in the park was relatively small (0.2-0.5℃). In the daytime, the greatest mean difference between the forest location and the two locations exposed to direct beam solar radiation (brick and lawn) was 0.5-0.8℃ (Max. 1.6-2.1℃). In the nighttime, the mean difference between the forest location and the other four locations in the park was small, though differences between the forest location and locations with grass cover (pine tree and lawn) reached a maximum of 0.9-1.7℃. Comparing air temperature between sunny and shaded locations, the shaded locations showed a maximum of 1.5℃ lower temperature in the daytime and 0.7℃ higher in the nighttime. Comparing the air temperature of the forest location with those of the residential (Kwonsun) and apartment (Maetan) locations, the mean air temperature difference was 0.8-1.0℃, higher than those measured between the forest location and the other park locations. The temperatures measured in the forest location were mean 0.9-1.3℃ (Max. 2.0-3.9℃) lower in the daytime than for the residential and apartment locations and mean 0.4-1.0℃ (Max. 1.3-3.1℃) lower in the nighttime. During the hottest period of each month, the difference was greater than the mean monthly differences, with temperatures in the residential and apartment locations mean 1.0-1.6℃ higher than those measured in the forest location. The effect of air temperature reduction on sampling locations within the park and a relatively high thermal environment on the urban sampling locations was clearly evident in the daytime, and the shading effect of trees in the forest location must be most effective. In the nighttime, areas with a high sky view factor and surface types with high evapotranspiration potential (e.g. grass) showed the maximum air temperature reduction. In the urban areas outside the park, the low-rise building area, with a high sky view factor, showed high air temperature due to the effect of solar (shortwave) radiation during the daytime, while in the nighttime the area with high-rise buildings, and hence a low sky view factor, showed high air temperature due to the effect of terrestrial (longwave) radiation emitted by surrounding high-rise building surfaces. The effect of air temperature reduction on the park with a high thermal environment in the city was clearly evident in the daytime, and the shading effect of trees in the forest location must be most effective. In the nighttime, areas with high sky view factor and surface types (e.g., grass) with evapotranspiration effect showed maximum air temperature reduction. In the urban areas outside the park, the high sky view factor area (low-rise building area) showed high air temperature due to the effect of solar (shortwave) radiation during the daytime, but in the nighttime the low sky view factor area (high-rise building area) showed high air temperature due to the effect of terrestrial (longwave) radiation emitted surrounding high-rise building surfaces.

We studied the distribution of air temperature using the high density urban climate observation network data of Daegu. The observation system was established in February 2013. We used a total of 38 air temperature observation points (23 thermometers and 18 AWSs). From the distribution of monthly averaged air temperatures, air temperatures at the center of Daegu were higher than in the suburbs. The daily minimum air temperature was more than or equal to 25℃ and the daily maximum air temperature was more than or equal to 35℃ at the elementary school near the center of Daegu. Also, we compared the time elements, which are characterized by the diurnal variation of surface air temperature. The warming and cooling rates in rural areas were faster than in urban areas. This is mainly due to the difference in surface heat capacity. These results indicate the influence of urbanization on the formation of the daily minimum temperature in Daegu.

We analyzed diurnal variations in the surface air temperature using the high density urban climate observation network of Daegu in summer, 2013. We compared the time elements, which are characterized by the diurnal variation of surface air temperature. The warming and cooling rates in rural areas are faster than in urban areas. It is mainly due to the difference of surface heat capacity. In addition, local wind circulation also affects the discrepancy of thermal spatiotemporal distribution in Daegu. Namely, the valley and mountain breezes affect diurnal variation of horizontal distribution of air temperature. During daytimes, the air(valley breeze) flows up from urban located at lowlands to higher altitudes of rural areas. The temperature of valley breeze rises gradually as it flows from lowland to upland. Hence the difference of air temperature decreases between urban and rural areas. At nighttime, the mountains cool more rapidly than do low-lying areas, so the air(mountain breeze) becomes denser and sinks toward the valleys(lowlands). As the result, the air temperature becomes lower in rural areas than in urban areas.

최근 대중매체를 통해 합천의 여름철 고온화에 대한 보도가 잇따르고 있다. 합천의 여름철 기온 특성을 파악하기 위해 1973~2007년 여름철(6~8월)의 기온자료를 사용하였고, 합천과 인접한 우리나라 대표적 혹서지 중 하나인 대구를 비교대상으로 선정하였다. 두 지점의 일최고기온 일최저기온, 일최고기온 극값, 열대일, 열대야를 비교하였다. 상대적으로 합천은 일최고기온, 대구는 일최저기온이 높게 나타났다. 열대일 시작일, 열대일 일최고기온, 열대일 수는 2002년을 기점으로 다른 양상이 나타났다. 일최고기온의 순별 분포에서 합천의 최고값은 8월로 7월인 대구보다 늦게 나타났고, 열대일 수는 비슷한 경향이었다. 열대야는 열대일과 다른 경향이 나타나서 열대야 수와 순별 열대야 분포 모두 합천이 상대적으로 적게 나타났다. 하지만 지난 2년간 합천의 열대야가 상당히 증가하는 것은 특징적이다. 최근 5년간 합천의 최고기온이 대구에 비해 상대적으로 크게 증가하는 것으로 나타났다.