Link budget, which accounts for the various losses, is differerent than the mathematical (statistical or random) modeling of the waveforms involved; the latter are used for the communication systems' performance trade-offs (usually in SNR vs BER) and the former, under ceratin assumptions, is used to determine required Tx power for the operational scenario at hand (which includes ferquencies, antennas etc) and can be thought of as corresponding to mean values.

The existence (or lack) of multipaths, along with their nature determines the appropriate mathematical modeling of the (aggregated) communication channel/propagation environment.

Thanks Pronios , You mean we still not have any model which has atmospheric losses as well as wireless channel both combined. Do we combine these mathematically ? e.g in one PDF which having most main losses as well as some wireless channel e.g Rician,Rayleigh or nakagami. please correct me .

If I say that we have shadowed rician Land mobile satellite model proposed by abdi then can we add these atmospheric losses to it in order to determine more accurate values at the ground receiver point.please correct me. does we need to separately deal with these  losses. please explain in detials

You can bring an understanding as if there is LOS & less no. of scatters then Rician model can be assumed otherwise Rayleigh model can be assumed.

Rest other parameters like the number of antennas, power, modulation etc. can also be calculated( refer Wireless Communication by Vishwanathan)

Are you doing satellite to mobile earth terminal, or something more like satellite to fixed terminal.  In any case, the NASA Deep Space Network manuals (called 810-005) has a very complete discussion of the various propagation factors involved at S, X and Ka band, with a lot of modeling equations that may help. http://deepspace.jpl.nasa.gov/dsndocs/810-005/

At a more tutorial basis Joe Yuen's book "Deep Space Telecommunications Engineering" covers a lot of this.

http://descanso.jpl.nasa.gov/DSTSE/DSTSE.pdf

There's a variety of handbooks and references on propagation available

http://descanso.jpl.nasa.gov/Propagation/propagation.cfm?force_external=0

Thanks  you very much Lux  for a reply. I am shall be very grateful if you further guide.I am actually looking for latest research issues  in satellite to fixed terminals propagation channel in general. I am specifically interested to know since Signal to noise ratio in random variable .what are the factors which can affect the received SNR in satellite to fixed channel. like there are many fading/attenuation models for these satellite to fix terminal channel which incorporate random shadowing,random enevelope, random phase Probability distributions function which helps to evaluate the system performance e.g. shadowed rician model is available in terms of hyper geometric function for satellite to fix terminal on ground . are there other factors on which Signal to Noise ratio depends in these channels. are there any other probability distribution functions available which includes atmospheric loss, rain attenuations, polarization mismatch, leo satellite distance variables as well as in their PDFs. How can we generalize to one composite PDF , Can we do that?

Yuen's book (downloadable from JPL) discusses a lot of the factors, and their probability distributions, which is important in making a link budget that is not overly conservative. Off hand, there's absorption type things (oxygen, water vapor, etc) for which good models exist, and there's a lot of statistical data for the model (typically, there is a radiometer colocated with the earth terminal to make accurate measurements of atmospheric loss).  Then there's short term random fluctuations: scintillation in the ionosphere (there was an issue of IEEE Proceeds back in the 70s or 80s about this), turbulence in the troposphere resulting in refractive errors, etc.  There's also things like mechanical loads on the antenna (gravity loads deforming the antenna, wind blowing, asymmetric heating from the sun).  And then there's a whole bunch of effects from hydrometeors (rain, ice, snow), for which you can get statistical data from radar handbooks, or if you have a particular location in mind, you can get climatology and calculate using various models.

If you are in an urban environment, or your propagation path is low elevation angle, or you're in a forest or something similar, then you have multipath (Rayleigh or Rician distributions, typically, depending on whether there's a direct path or not).  The various land-mobile/terrestrial microwave models (Longley Rice, etc.) do a nice job of this, assuming you have good inputs for the model parameters (terrain roughness, etc.), and you're aware of the model limitations.

Ultimately, you could do a raytracing model (this is moderately common in land mobile radio/broadcast situations). There are a variety of programs (and companies) out there that can do this, with varying sophstication of the propagation model, and the geometry model (SRTM DEM data is a start, but doesn't have buildings or vegetation)

noise is your answer.

it means satellite link design.check the base band S/N ratio. and off course system noise temperature, check attenuation