High-quality Local Climate Zones (LCZ) and Land Surface Temperature (LST) map pairs Dataset of Five Southeast Asian cities (SEA-LCZ-LST)

The Southeast Asian LCZ-LST dataset (SEA-LCZ-LST) including 27 high-quality georeferenced map pairs of five cities: Kuala Lumpur, Jakarta, Bangkok, Manila and Singapore made publicly available in 100m resolution. This dataset contains a total of 16 LCZ types and is split into two folders, LCZ and LST. Both the LCZ and LST folders contain subfolders for each city, all LCZ and LST raster files use the naming convention “CITY_YEAR”. The LCZ files contain LCZ types 1 to 16, along with cloud affected pixels, which are assigned the value “17”. The LST files contain continues temperature values, with the cloud affected pixels assigned the value “9999”. No “NoData” value was assigned to allow easy use for machine learning tasks. The details of the scenes used to construct this dataset including file names, year, chosen Landsat Scene ID, cloud cover and the LCZ mapping overall accuracy are presented in Table 5. A total of 27 LCZ-LST map pairs were collected: 4 for Kuala Lumpur, 4 for Singapore, 7 for Bangkok, 5 for Jakarta, and 7 for Manila.东南亚局地气候区-地表温度数据集（Southeast Asian LCZ-LST dataset，简称SEA-LCZ-LST）包含覆盖吉隆坡、雅加达、曼谷、马尼拉与新加坡5座城市的27组高质量地理配准影像对，以100米空间分辨率公开共享。本数据集共涵盖16种局地气候区（Local Climate Zone, LCZ）类型，分为LCZ与地表温度（Land Surface Temperature, LST）两个文件夹。LCZ与LST文件夹下均设有对应各城市的子文件夹，所有LCZ和LST栅格文件均采用"CITY_YEAR"的命名规范。LCZ文件包含1至16号局地气候区类型，同时包含受云影响的像元，此类像元被赋值为"17"。LST文件存储连续的温度数值，受云影响的像元被赋值为"9999"。未设置"NoData"（无数据）值，以方便机器学习任务的开展。本数据集所使用的影像场景详情（包括文件名、拍摄年份、所用Landsat场景ID、云量以及局地气候区制图总体精度）详见表5。本次共收集到27组LCZ-LST影像对：吉隆坡4组、新加坡4组、曼谷7组、雅加达5组以及马尼拉7组。