Hello,

you are welcome to use GEWEX Cloud Assessment Database:

http://climserv.ipsl.polytechnique.fr/gewexca/index-2.html

which covers all imaginable cloud properties on a month-to-month basis.

Thank you for your response. Databases are available in .nc format which I have never used so I have to work on it and I will work surely. Actually sir, I need data in image format using which I can train my adaptive system. In this direction if you have any thing to share kindly acknowledge me back.

you are welcome.

Actually, netCDF is a standard format, and lots of modern programming languages have a corresponding library.

Speaking about optical images - each pixel there contains an information "compressed" along the line of sight. On the other hand, the database I'm talking about provides height-stratified cloud parameters which you can project onto a 2D plane if you wish.

Training a system with 2D images may or may not give the desired results - the problem is that the cloud systems have 3D structure and are quite complex. Moreover, any system trained on 2D optical dataset will almost surely fail in the case of bright clouds over snow/ice areas. I would think twice before investing into it.

Exactly Artem sir. Reflectance of cloud and snow/ice is almost similar in most of the region of electromagnetic spectrum and hence difficult to differentiate both.

I am trying to extract information using Matlab from Databases link which you have sent. Hope I can crack something from those databases.

Thank you again.

Thank you Upal sir. 

I saw the link which you just sent. I found this link very useful to me. But the only challenge in front of me is to get the original data with same lat. long. as of in cloud database. I am trying to find it out the same. 

Dear Rachana,

Including above, you can use satellite product like CALIPSO (lidar) and CloudSat (radar).

CALIPSO having an issue to detect thin clouds, but CloudSat is good for that and additionally, you can go through DADAR product, which is a combination of both lidar and radar data from Calipso and Cloudsat with modified algorithm.

Hope this can help you

Best Regards

Piyush Patel

Hello Piyushkumar

Thanks for response.

Currently I am not focussing on Radar data as I feel different algorithms are required to work on those data.  

Dear Rachana

If i can understand yours current focus is to differentiate between clouds. as you are using the IR and VIS data both the techniques are indirect where BT values and cloud top temperatures are used to define the cloud type the more negative temperature means convective clouds or it can be cirrus.

If you just want a spatial pattern so that you can differentiate them i think the best data is supplied in the below mentioned link. Here you can choose between the data sets with varying spatial and temporal resolution and the data can be downloaded in Geotiff format also.

https://giovanni.gsfc.nasa.gov/giovanni