Insights into the health impacts of synoptic-scale climatology in two Korean cities: A Retrospective Analysis of Morbidity

Important associations between the effects of weather and climate on human health outcomes are well-known. The human body responds concomitantly on varying scales of severity to a full atmospheric situation, rather than air temperature alone. Thus we apply a synoptic climatological approach, the “spatial synoptic classification” (SSC), to human health data in two large South Korean cities (Seoul: population 10.58 million; Busan: population 3.62 million).

The objective of this evaluation is to determine the impact of synoptic-scale weather on warm- and cold-season morbidity in Seoul and Busan. This is the largest synoptic evaluation of morbidity completed to date. Hospital patient visits were used as a proxy for human morbidity in both the warm and cold seasons (May–Jul; Nov–Feb, respectively). Hospitalization data was standardized to obtain anomalous morbidity to the mean patient number. The relative risk (RR) of hospital visits was examined for each SSC category using a distributed lag-model (days 0–3) based on a generalized linear model. Time confounders of year, day of week, and Julian day (third order spline) were adjusted for. All RRs were compared to the ‘control' type air mass of dry moderate (DM), which is relatively benign in character in both seasons. Day of week analysis displayed days to be roughly equal, excluding Sundays, which had approximately 80 percent less patients overall, and was therefore removed from the analysis. Wintertime presented the highest number of hospital patients. The dry polar (DP) air mass consistently displayed the greatest number of patients in winter, with moist polar (MP) weather also demonstrating high positive patient anomalies. The mean anomalous morbidity in the DP air mass ranged from 10–59% for Busan, and 4–44% for Seoul. Hospitalizations on dry and moist moderate days (DM and MM) were consistently below the mean. Further, the DP air mass demonstrated significant increases in cumulative risk of hospital visits (i.e., RR = 2.02 (95% CI 1.50–2.72) for Seoul; RR = 1.93 (95% CI 1.55–2.41) for Busan).

Summertime results differed greatly from that of winter for both cities, and also displayed considerably lower patient numbers. However, the hottest (and most oppressive) air masses of moist tropical plus (MT+) and dry topical (DT) displayed high positive anomalous patient numbers, with MT+ weather resulting in the greatest increases. Cumulative risks of hospital visits demonstrated a significant increase in the MT+ air mass in both Seoul (RR = 1.88 (95% CI 1.05–3.37) and Busan (RR = 2.78 (95% CI 2.17–3.57)). Positive patient numbers and increased RR in the MT, MT+, and MT++ air masses were more prevalent and higher in Busan as compared to Seoul, which most likely relates to the warmer and more humid climate. The summertime air masses of DP, MP, DM, and MM were generally below the mean number of patient visits.

The most extreme warm and cold season air masses displayed significant cumulative effects on patient numbers through a three-day lag. Our findings suggest that the number of hospitalizations is statistically related to variations in synoptic weather patterns, being significantly higher in cases of oppressive air masses. These results can add vital information to the heat-health warning systems already present in the two cities, which are based upon variations in mortality and not morbidity. Further, city-specific forecasting of such weather in Korea can provide more accurate projections of the numbers of patient visits, as well as the relative risk of morbidity due to weather.