MODIS衛星データを用いたPM2.5大気汚染の検出(その1: 黄砂検出との違い)

Abstract

In recent years, PM2.5 air pollution is a social and transboundary environmental issue with the rapid economic growth in many countries. As PM2.5 is small and includes various ingredients (e.g., sulfur oxide, nitrogen oxide, vitriol, nitrate salt, soot, etc.), the detection of PM2.5 air pollutions by using satellite data is difficult compared with the detection of dust and sandstorms (DSS). DSS can be detected by using AVI method and YDI method. AVI (Aerosol Vapor Index) is defined as AVI=T12-T11, where T12 and T11 are the brightness temperatures at 12μm and 11μm wave lengths, respectively. For MODIS data, T12 and T11 correspond to band32 and band31, respectively. YDI (Yellow Dust Index) is defined as YDI=(band4-band3)/(band4+band3). AVI and YDI methods detect PM2.5 air pollutions only a little. In this paper, we examine various RGB composite color images for detecting PM2.5 air pollutions by using MODIS data, i.e., {R, G, B = band4, band3, T11}, {R, G, B = band10, band9, T11}, {R, G, B = band9, band8, T11}, {R, G, B = AVI, band7-band1, T11}, {R, G, B = AVI, band10-band9, T11}, etc. A good method for the detection of PM2.5 air pollution is {R, G, B = band10, band9, T11}. In this composite color image, PM2.5 air pollutions are represented by light purple or pink color. This proposed method is applied to the detection of PM2.5 air pollutions in the wide area from China and India to Japan by using MODIS data on 12 January 2013, and AVI method is applied to DSS detection in the same area. By comparing the AVI image with the image by the proposed method, PM2.5 air pollutions can be distinguished from DSS. The proposed method is simpler than the method by Nagatani et al. (2013), and is useful to grasp the distribution of PM2.5 air pollutions in the wide area.