I want to create a variable "snow cover" with values presence/absence of snow for different weather stations (to work later with NDSI from Landsat images with a pixel size = 100 m). I got snow depth values (cm) from weather stations, but I'm wondering if there is a formula or a minimum depth I should consider as presence of snow. Reading the methods used to estimate snow depth, values are based on different measurements around the station, but it does not specify how big the area is.

can you please advice on this matter?

  • 1
    $\begingroup$ IMO it would be intuitive that your snow cover variable would be true when snow depth is nonzero. If the depth is concluded via multiple measurements in the given area then your presence of snow could be a fraction varying from 0 to 1. $\endgroup$
    – Communisty
    Commented Oct 11, 2017 at 10:19

2 Answers 2


A measureable amount of precipitation is considered to be at least 0.01 inch of water. Since on the average 10 inches of snow melt to 1 inch of water, this corresponds on the average to 0.1 inch of snow. Anything less is considered a trace (too little to measure).


Not an expert here - but my thought is that you have to assume that the snow cover is reasonably even around the weather stations, because you have no way of doing otherwise - so my advice would be to "calibrate" by comparing against the Landsat images at the locations of the weather stations, and determine what amount of snow cover generally corresponds to the pixel appearing snowed.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.