I am using Land cover indices to compare their relationship with temperature for different places. Is it important to convert the digital number to at ground reflectance for deriving the indices as the effect of topography, atmosphere and others will be subtracted anyway by applying the indices????
As the radiance / reflectance values are not linearly related to the DN in all cases. For example Red and IR may both have values of .5 but when corrected they likely will not be the same number (correction is band specific). The question has been answered before well on Research Gate. Here is the link.
https://www.researchgate.net/post/Is_it_incorrect_to_calculate_NDVI_using_DN_Digital_Numbers_values_associated_with_a_satellite_image_instead_of_using_the_Surface_Reflectance_values
red and NIR may be affected similarly in low aerosol cases, but not in high aerosol cases. Unfortunately most atmospheric correction models only include a standard aerosol assumption. And finding good aerosol retrievals before about 2000 will be difficult due to the lack of multispectral satellites.
So you are correct that the atmospheric correction mainly affects the blue end of the spectrum and that ndvi is largely unaffected, but even though you may not be able to correct for aerosol all the way through the data set, it's good to be aware if there is a significant aerosol trend in your region of study.
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Atmosphere