NDVI images with Sentinel 2

[ASalas]

Hello:

I am working with Sentinel 2 images to generate NDVI maps, using QGIS. I create the NDVI image with the raster calculator using the formula NDVI=(NIR-R)/(NIR+R), but I wonder how do other factors such as cloud shadow or the previous treatment of the images. I also would like to know the difference of working with the 1C data or the 2A data to get NDVI values.

[pasfans01]

1. For the shadow, it will make the NDVI values close to zero, whatever the landcover covered by it

2. for level 1C data, the digital number/reflectance is processed to top of atmosphere  value, which means atmospheric effect is not yet corrected, and for level 2A, the digital number/reflectance has been corrected to surface/bottom of atmosphere reflectance (atmospheric effects has been corrected). Theoretically , the level 2A data will give more accurate NDVI values compared to level 1C, 

[KonStAntINOS2018]

@ASalas Also adding to pasfans01 answer 2., keep in mind that if you want to use NDVI for data assimilation puproses (e.g. compute vegetation biophysical parameters etc.) you will need the Surface Reflectance (Level 2A) data to produce the NDVI map, since 2A are more close to the true radiance values. For more information, you could read the sentinel 2 users guide that ESA offers. If you use the 2A data, in your case it would be of great help to first check the quality assesment data in order to see and assess the impact of cloud coverage in you scene.

Regards

[ASalas]

Thank you for your answers. I have carried out some tests with 1C and 2A images of the same day and 2A values are usually higher due to the reflectance correction. I wonder to know if the time of the day when the image is taken (or in other words: the amount of light or the height of the Sun) affects to the real values of the NDVI.