Hi Suman,

Your question is not quite clear to me. Would you like to have plots for multi-dimensional pdf's and histograms ? For one parameter like rainfall, humidity, or wind speed, you might use several pdf functions in R. See: http://statmethods.net/advgraphs/probability.html.

R is freely available, but you have to invest some time learning it. Once you have mastered this software you will never quit. Maybe I can help with an example of your own data. Just send them as csv to me and I'll give it a try.

regards,

Marcel

Hi Marcel

Thank you very much for your attention and comment. Problem is that the data is not csv format but in netcdf. As well as the data size is large. I can't send you. Therefore I am redefining my problem so that you can understand better.

I have precipitation data in three dimension i.e. for some longitude, latitude and time. I want to find histogram and probability density function of that parameter. how to do that? As per my knowledge R can read netcdf data. I will take care of that. Please advice me what is the procedure mathematically. I couldn't understand whether to find histogram and probability density function with respect to space or time. If you have any example in R please write.

Hi Suman,

I only have some experience with a one-parameter sets, I'm sorry. Space (Lon or Lat) and time should be considered as second and/or third variable. A 3D histogram is basically possible with R packages like ggplot or rgl. Take a look at:

http://r.789695.n4.nabble.com/3D-histogram-td861674.html or

http://stackoverflow.com/questions/7684294/creating-a-3d-histogram-with-r

There are some examples over there but you might explore the manuals for the packages too. For better handling you should translate your netcdf file into a datamatrix following the instructions on: http://geog.uoregon.edu/bartlein/courses/geog607/Rmd/netCDF_01.htm. This dataframe can be handled by the mentioned packages. Another source, using the same package, and this page is probably the best to start with:

http://www.image.ucar.edu/GSP/Software/Netcdf/

I know this is a hard job...

wishing you good luck, 

Marcel 

Dear Dr. Suman, 

In many cases I tried the beta-distribution function. How to obtain the bell-like curve. You can see my papers related to UDFRF - universal distribution function for the relative fluctuations. You can find my recent papers published with Dr. A. Skorinkin (2014). 

You trying to find probability distribution  of precipitation data (X) collected  in three dimension (longitude, latitude and time). so variable of interest is X and you want ot fit a appropriate pdf to it. !!

Dear Subrata Chakraborty

Thank you very much for your query. Yes I want to do that. Consequently, if I don't want to fit any appropriate pdf but to find the pdf of the data itself is it possible? If yes how to do that?

Density Estimation e.g. Histogram and its more sophisticated generalization