STUDY OF APPLICABILITY OF MOBILE MEASUREMENTS BASED ON AIR TEMPERATURE MEASUREMENTS

 Urban environmental issues related to heat and aerial pollutants have become increasingly more serious. Atmospheric properties, such as air temperature, wind velocity, and pollutant concentrations, are typically measured at meteorological observation stations. However, due to the low spatial resolution of the station networks, we cannot observe local issues occurring on the scale of people's daily lives. Thus, mobile measurement is an effective method for addressing the insufficient measurement points in urban areas. However, the extent of Global Positioning System (GPS) measurement errors are unclear and results of mobile measurements are uncertain. Therefore, this paper discusses the applicability of mobile measurements to high-density urban areas. After investigating the range of GPS errors, we consider the following three aspects: (1) spatial distribution of air temperature at a high spatial resolution of 10 m; (2) the range of uncertainties in mobile measurement results; and (3) the impact of spatial resolution (10 m or 100 m) on the measured spatial distribution of air temperature.<br> In this study, we used bicycles as mobile measurement platforms to investigate the spatial distribution of air temperature. Two measurement sites in Tokyo were selected: Shinjuku, a commercial area (August 25^th to 28^th, 2015) and Shibuya, a residential area (January 20^th to 29^th and February 16^th to 19^th, 2016). We equipped bicycles with platinum resistance thermometers (a high responsivity rate of 2.2 s), a data logger, and a GPS logger to measure air temperature and spatial position during the sequential travel of the bicycles. Measurements in Shinjuku were conducted on cloudy days. To ensure that the measurement conditions were consistent, measurements in Shibuya were conducted from 16:00 to 18:00, when solar radiation, and thus its influence, was very low.<br> From the results, we draw the following conclusions. Firstly, GPS data collected in urban areas tend to include positioning errors of 14 m on average due to the reflective properties of building surfaces according to adapting our developing correction method. Secondly, spatial distribution of air temperature variation was almost ±0.65 °C (min to max). Thirdly, the average of uncertainty of air temperature variation was 0.03 °C. It was much smaller than the range of air temperature variation (min to max). Lastly, we can find a similar distribution of air temperature between a mesh size of 10 m and 100 m. However, a spatial resolution of less than 100 m was necessary in order to observe locations where the air temperature is locally higher than near street intersections. However, we have to consider GPS errors are 14 m on average, which is larger than a mesh size of 10 m. From the above, we obtained a characteristic spatial distribution of air temperature by correcting position data. Suppose that we reduce uncertainty to less than 0.1 °C in the case that the max value of standard deviation of air temperature at each cell is 0.4 °C, 16 samples at each cell must be corrected in one hour. Therefore, we found mobile measurement is an effective method for investigating environmental conditions in high-density urban areas for high spatial resolution.