I'm pretty green when it comes to RS and I'm doing this school project that I need help with. The project is to measure albedo changes for the tongue of a glacier and I'm using R programming to do it. So I got my data from earth explorer and went for the surface reflectance product in order to save some steps. However, the values for these cells are higher than expected, around 2000 depending on the band. I say higher because when I use a weighted average formula for calculating albedo, which I found in some papers cited below, the outcome gives me cell values of around 1000, not the 0 to 1 I would expect for and albedo. The paper mentions its using a surface reflectance as well for this formula.

My question is, is this formula wrong? Albedo = 0.493Band2 + 0.203 Band + 0.150 Band5 + 0.154Band

Or does some other processing need to be done to my Earth Explorer Surface Reflectance Product. https://landsat.usgs.gov/landsat-surface-reflectance-high-level-data-products

Also I'm using Landsat 4-7 images.


Pimentel, Rafael, et al. "Comparison between Snow Albedo Obtained from Landsat TM, ETM+ Imagery and the SPOT VEGETATION Albedo Product in a Mediterranean Mountainous Site." Hydrology 3.1 (2016): 10.

Also, this is really similar but uses TOA instead of surface reflectance: http://yceo.yale.edu/how-convert-landsat-dns-albedo

1 Answer 1


The Landsat reflectance data you downloaded from USGS has been scaled using a scale factor of 0.0001. So multiplying the digital number by 0.0001 will give you a value between 0 and 1. The 2000 value you are talking about, i believe, is the saturate value (which should be 20000). See this document for more details: https://landsat.usgs.gov/sites/default/files/documents/ledaps_product_guide.pdf

Converting Landsat surface reflectance data to albedo is actually much more complicated than applying a single equation. The equation you listed is just a quick approximation, which should give you a general idea of how albedo changed. So if you are just using this for a simple comparison, I would say it’s good enough, as long as you are using surface reflectance (instead of TOA). If you are really interested in this and want to know more, you can take a look at this paper: http://www.sciencedirect.com/science/article/pii/S0034425711001477

  • Thanks a lot, that is very helpful. It makes sense now and I can get some plausible results. It does seem too simple, but I think for my purposes it will work. I will be watching for a trend through time, so their accuracy is not terribly important to me. I'll have a look at that paper tomorrow and maybe see which equation works best. Thanks again. Commented Mar 10, 2017 at 4:34

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.