Hello, Majid:

First of all, I assumed that you are asking for aerosol retrievals over land. The DT and MRT are two different ways to get surface reflectance in order to retrieval aerosols. First of all, DT aerosol retrievals is based on the fact that some relation existed between surface reflectance at a longer wavelength, such as shortwave-IR, where the effect from the aerosols, except for aerosol with a large particle size.  is usually minimum  except for,  and  surface reflectance at  visible bands. MRT is based on the assumption that for each pixel, surface does not change and the day with a minimum observed reflectance is a cleanest (both cloud and aerosol free) day over a certain time period.  Therefore,  by assuming a background aerosol amount, such as AOD of 0.02, 0.05, depending on the location, you can calculate the surface reflectance from the observed TOA (Top of Atmosphere) reflectance, then use the retrieved surface reflectance to derive the amount of aerosol for each day over this time period. Therefore are a lot of studies regarding to aerosol retrievals, I think you should read  this paper:

 http://meto.umd.edu/~zli/PDF_papers/Kokhanvsky%20et%20al.%20Atmos%20Res.pdf

However, MOD09 is a surface reflectance product already, which means it has been corrected for aerosol effect.

hope my answers are helpful.

Minimum Reflectance Technique (MRT) is used to generate a new image with minimum values for particular pixels from multiple images (e.g. seasonal data set or yearly dataset). Lets say you have five images having surface reflectance of 0.09 (value from image one), 0.04 (value from image two), 0.08 (value from image three), 0.1 (value from image four), and 0.07 (value from image five) values for specific pixel, and MRT will generate a new image containing 0.04 (minimum value) for that specific pixel. Similarly, other  pixels will have minimum values from five images, and pixels in the new image will have minimum values.

Hope this will help!

Dear Muhammad,

Thank you for your answer,So it is simple to run,I have seen this model is used in some of articles in case of retrieving a clean surface reflectance without any aerosols,What do you think about it?If we had some imeges for each season and run this model on them does the result shows a good estimation of surface reflectance?

Check this article on Dark Target

http://semiautomaticclassificationmanual.readthedocs.org/en/latest/Landsat_conversion.html

The Dark target is a family of image-based atmospheric corrections. Chavez (1996) explains that “the basic assumption is that within the image some pixels are in complete shadow and their radiances received at the satellite are due to atmospheric scattering (path radiance). This assumption is combined with the fact that very few targets on the Earth’s surface are absolute black, so an assumed one-percent minimum reflectance is better than zero percent”.  

MODIS operational Deep Blue (DB) aerosol retrieval algorithm uses surface reflectance estimated by the minimum reflectivity method. So, you can also try this method and see the results.  

Thank you so much,I'll try it,Is there any source that explains conditions and consideration if there is any for this technique?

Hi Majid,

There is no point to use MOD09 data. This is beacuse the MOD09 product is atmospherically corrected, i.e. the aerosol signal has been "removed". You should apply these two technics on at-sensor reflectance data.

Kind regards,

Panagiotis

Hi Panagiotis,

Thank you for your answer,So what would be deference between MOD09 and the result of MRT?In fact I need a good estimation of surface reflectance and I want to know which of these gives me a accurate result

Hello, Majid:

First of all, I assumed that you are asking for aerosol retrievals over land. The DT and MRT are two different ways to get surface reflectance in order to retrieval aerosols. First of all, DT aerosol retrievals is based on the fact that some relation existed between surface reflectance at a longer wavelength, such as shortwave-IR, where the effect from the aerosols, except for aerosol with a large particle size.  is usually minimum  except for,  and  surface reflectance at  visible bands. MRT is based on the assumption that for each pixel, surface does not change and the day with a minimum observed reflectance is a cleanest (both cloud and aerosol free) day over a certain time period.  Therefore,  by assuming a background aerosol amount, such as AOD of 0.02, 0.05, depending on the location, you can calculate the surface reflectance from the observed TOA (Top of Atmosphere) reflectance, then use the retrieved surface reflectance to derive the amount of aerosol for each day over this time period. Therefore are a lot of studies regarding to aerosol retrievals, I think you should read  this paper:

 http://meto.umd.edu/~zli/PDF_papers/Kokhanvsky%20et%20al.%20Atmos%20Res.pdf

However, MOD09 is a surface reflectance product already, which means it has been corrected for aerosol effect.

hope my answers are helpful.

Dear P. Ciren,

In fact I am going to calculate surface reflectance in case of generation some indexes to estimate pm10,because the study area is a polluted area ,minimum pm10 is always 40 and so I selected surface reflectance (M0d09) and I ran MRT on them.As you said MOD09 is corrected for aerosol effect and so MRT shows cleanest reflectance compare with MRT with TOA reflectance, Is this true?or I am wrong with sth?

Thank you for The article.

Dear Majid:

I believe, MOD09A1 is derived from a series images during of course of time period, (8 or 16 days)  to remove cloud  and heavy aerosol, which means the MRT was applied to derive MOD09A1. You should check the user guide or ATBD for MOD09 products.

http://modis-sr.ltdri.org/pages/documentation.html

I suggest you to use MODA1 product directly.

good luck!

pubu

Dear P. Ciren,

I have read in an article that in case of drive surface reflectance they suggested that we use at least a series of 30 days per season's,I'll read more about MOD09GA which I can find if MRT is applied on them or not,

thank you

Majid

Thanks to all. I am working with DB aerosol product. Could anybody explain me why Deep Blue AOD product at 550nm is convenient ?

Well, there are many differences between DT aerosol and DB aerosol retrieval, the fundamental difference is DB has aerosol retrieval over bright surface, such as over desert, DT will only give you aerosol retrievals over vegetated area. If the area you are studying is over bright surface, then using DB aerosol will give you more data. As for at 550nm, it is easy for you to compare with other aerosol products at the same wavelength.

hope my answer is helpful.

thanks

pubu

There is also two manuals about DT and DB algorithms which gives quite good information about advantages and weaknesses of these algorithms