Agreement in the vertical profiles of the temperature trends from radiosonde observation (HadAT) and four kinds of reanalysis dataset (ERA40, ERA-I, NCEP-DOE, and 20CR) are examined for the period of 1979-2000. There are noticeable spread among reanalysis and observation datasets in the temperature trend depending on region and vertical level. East Asia shows large discrepancy among datasets, while Europe shows relatively good agreement. Generally, biases in temperature trends are larger in the upper troposphere (above 300 hPa) than in the lower and middle troposphere. Comprehensive comparison of the long-term temperature trends among reanalyses is made for horizontal distributions with height, latitude-pressure cross-sectional distributions, zonally-averaged meridional distributions with height, and area-averaged vertical profiles in both DJF and JJA. Consequently, we find that the degree of agreement among reanalyses significantly varies with vertical level, region, and season. The highest discrepancy is found over southern high-latitudes and in the upper troposphere over southern tropics. In the tropical upper troposphere above 200 hPa, observation (HadAT) shows cooling trend increases with height, but three reanalyses show warming trends except NCEP-DOE reanalysis in which cooling trend is overestimated. In conclusion, discrepancies in the vertical profiles of long-term temperature trends among four kinds of reanalysis datasets are quite large, and then a scrupulous approach should be needed when reanalysis dataset is used for climate change study.

TCI(tourism climatic index) is a measure of the suitability of climate for outdoor sightseeing, which combines seven climate variables. Based on the TCI, we analyse the present climate resources for tourism in Gangwon-do and assess the recent changes. We use daily meteorological data from 11 stations in Gangwon-do. First, we compare mean annual cycles of the TCI for 5 stations(Gangneung, Sokcho, Wonju Chuncheon, and Daegwallyeong). This comparison reveals that range of the annual cycle is from minimum 35(for Daegwallyeong in January) to maximum 80(for Wonju in May and September). Daegwallyeong which is located in highland is characterized by summer season peak pattern while other regions have low TCI values in hot summer. In long-term trend of the TCI, Gangneung has increasing trends in February, April and December, whereas it has significant decreasing trends in summer and fall(June to October). In case of Daegwallyeong, increasing trends are found in February, November and December, and relatively steep declining trends in summer season. The overall decreasing trend in summer season is a common feature in Gangwon-do. Decreasing trends at Gangneung in August and September might be mostly explained by the increasing trend of rainfall amount in those months. Meanwhile, increasing trends of the TCI in winter season might be a positive impact of climate warming on the tourism sector.