Development of an urban landcover classification scheme suitable for representing climatic conditions in a densely built-up Asian megacity

The objective of this paper is to acquire a better insight into a landcover classification scheme which is suitable for representing the climatic properties of Asian megacities. Seoul, Republic of Korea, was used as a testbed. First, landcover classification was performed using Landsat 7 ETM+ using eight landcover types suitable for representing the climatic conditions of Seoul. As a next step, 10 land surface and thermal parameters were estimated for individual landcover classes. Finally, the suitability of the climatologically classified urban landcover scheme was analyzed in two ways. Firstly, the spatial distribution of land surface temperature (LST) was retrieved using the thermal band of Landsat 7 ETM+ and the retrieved LST values over eight landcover classes were statistically analyzed. Secondly, 24-h temperature changes based on both eight landcover classes and four classes were simulated using the meso-scale model MetPhoMod. These were compared with measurement data from 24 automatic weather stations in Seoul. This study showed that the eight landcover classification scheme is suitable for representing the climatic properties of Seoul, which is one of the Asian megacities. Moreover, the eight classes represent the thermal conditions better than the four conventional classes during nighttime, which implies that this eight landcover classification scheme can be useful to analyze the nocturnal urban heat island effect.