Karnieli, A, Kaufman, YJ, Remer, L, Wald, A (2001). AFRI - aerosol free vegetation index. REMOTE SENSING OF ENVIRONMENT, 77(1), 10-21.
Aircraft measurements using a field spectrometer over variety of ground surfaces in Israel reveals that under clear sky conditions the shortwave infrared (SWIR) spectral bands around 1.6 and 2.1 mum are highly correlated with the visible-blue, green, and red- spectral bands. Empirical linear relationships, such as rho (0.469) = 0.25(rho2.1); rho (0.555) = 0.33(rho2.1); rho (0.645) = 0.5(rho2.1); and rho (0.645) = 0.66(rho1.6), were found to be statistically significant and consistent with previous findings. Based on the above relationships, a modified vegetation index (VI) is proposed and named Aerosol Free Vegetation Index (AFRI). Two versions of this VI are formulated as: AFRI(1.6) = (rho (NIR) - 0 66 rho (1.6))/(rho (NIR) + 0.66 rho (1.6)) and AFRI(2.1) = (rho (NIR) - 0.5 rho (2.1))/(rho (NIR) + 0.5 rho (2.1)). It is shown that under clear sky conditions, the AFRIs (and especially AFRI(2.1)) closely resemble the Normalized Difference Vegetation Index (NDVI) and their values are almost identical. The advantage of the derived AFRIs, based on the ability of the SWIR bands is to penetrate the atmospheric column even when aerosols such as smoke or sulfates exist. Consequently, these indices have a major application in assessing vegetation in the presence of smoke, anthropogenic pollution, or volcanic plumes. This was demonstrated by applying the AFRI for a biomass burned forest in Brazil. Limited success of these indices is expected in case of dust due to presence of larger particles that are of similar size to the wavelength and therefore not transparent at 2.1 mum. The AFRIs can be implemented to data from any sensor that has the SWIR bands. Currently the most commonly used of such instruments are the Landsat-Thematic Mapper (TM) and Advanced Thematic Mapper Plus (ETM+), Moderate Resolution Imaging Spectrometer (MODIS), Advanced Spaceborne Thermal Emission and Reflection (ASTER), and Japanese Earth Resources Satellite-Optical System (JERS-OPS). Although the AFRI(2.1) was demonstrated to perform better than the AFRI(1.6), the latter still can be used for the same application in conjunction with instruments that have onboard only the 1.6-mum band, such as Systeme Probatoire d'Observation del la Terre (SPOT4)-VEGETATION, Indian Remote Sensing (IRS-IC/D), and Resource-21. (C) 2001 Elsevier Science Inc. All rights reserved.