Generally, no big difference.

It depends on which satellite dataset you are going to use + what precision you need regarding the forest cover change monitoring. Normally, radiometric and atmospheric corrections can significantly increase the accuracy of your results, especially for complex terrain and in more large areas.

Also, please have a look at:

https://www.mdpi.com/2072-4292/13/18/3550

Please be welcome to assess our papers using NDVI without such calibrations (e.g. forest, Italy; vegetation Namibia) at countrywide level and grasslands locally at fine resolution.

Article Ground-based reflectance measurements for standing crop estimates.

Dear Rawaz Rostam ,

if you are studying the trend in the NDVI (greener vegetation, enhancement of a vegetated land, aridification,...) these corrections do not create "much" differences.

However, for gaining a better accuracy it is best to do corrections.

Depends on the final aim of the study.

Dear Abdelrazek Elnashar Vahid Nasiri Jehan-Antoine Vayssade Hein Van Gils Roshanak Tootoonchi thank you a lot to share your thoughts. All appreciated.

I think it is only appropriate if your using Collection 2 level 2 Landsat satellite images that has already been atmospherically corrected. But I will advise that you perform all the corrections on the level 1 imagery for an enhance results.