Rain fade is usually estimated experimentally and also can be calculated theoretically using scattering theory of rain drops. Rain drop size distribution (DSD) is an important consideration for studying rain fade characteristics.[1] Various mathematical forms such as Gamma function, lognormal or exponential forms are usually used to model the DSD. Mie or Rayleigh scattering theory with point matching or t-matrix approach is used to calculate the scattering cross section, and specific rain attenuation. Since rain is a non-homogeneous process in both time and space, specific attenuation varies with location, time and rain type.

Total rain attenuation is also dependent upon the spatial structure of rain field. Horizontal as well vertical extension of rain again varies for different rain type and location. Limit of the vertical rain region is usually assumed to coincide with 0 degree isotherm and called rain height. Melting layer height is also used as the limits of rain region and can be estimated from the bright band signature of radar reflectivity.[2] The horizontal rain structure is assumed to have a cellular form, called rain cell. Rain cell sizes can vary from a few hundred meters to several kilometers and dependent upon the rain type and location. Existence of very small size rain cells are recently observed in tropical rain.[3]

Possible ways to overcome the effects of rain fade are site diversity, uplink power control, variable rate encoding, receiving antennas larger (i.e. higher gain) than the required size for normal weather conditions, and hydrophobic coatings. Only superhydrophobic, Lotus effect surfaces repel snow and ice.

The simplest way to compensate the rain fade effect in satellite communications is to increase the transmission power: this dynamic fade countermeasure is called uplink power control (UPC). Until more recently, uplink power control had a limited use since it required more powerful transmitters - ones that could normally run at lower levels and could be run up in power level on command (i.e. automatically). Also uplink power control could not provide very large signal margins without compressing the transmitting amplifier. Modern amplifiers coupled with advanced uplink power control systems that offer automatic controls to prevent transponder saturation make uplink power control systems an effective, affordable and easy solution to rain fade in satellite signals.

In terrestrial point to point microwave systems ranging from 11 GHz to 80 GHz, a parallel backup link can be installed alongside a rain fade prone higher bandwidth connection. In this arrangement, a primary link such as an 80 GHz 1 Gbit/s full duplex microwave bridge may be calculated to have a 99.9% availability rate over the period of one year. The calculated 99.9% availability rate means that the link may be down for a cumulative total of ten or more hours per year as the peaks of rain storms pass over the area. A secondary lower bandwidth link such as a 5.8 GHz based 100 Mbit/s bridge may be installed parallel to the primary link, with routers on both ends controlling automatic failover to the 100 Mbit/s bridge when the primary 1 Gbit/s link is down due to rain fade. Using this arrangement, high frequency point to point links (23 GHz+) may be installed to service locations many kilometers farther than could be served with a single link requiring 99.99% uptime over the course of one year.

## CCIR interpolation formula

It is possible to extrapolate the cumulative attenuation distribution at a given location by using the CCIR interpolation formula:[4]

Ap = A001 0.12 p−(0.546 − 0.0043 log10 p).

where Ap is the attenuation in dB exceeded for a p percentage of the time and A001 is the attenuation exceeded for 0.01% of the time.

## ITU-R frequency scaling formula

According to the ITU-R,[5] rain attenuation statistics can be scaled in frequency in the range 7 to 55 GHz by the formula

${\displaystyle {\frac {A_{2}}{A_{1}}}=\left({\frac {b_{2}}{b_{1}}}\right)^{1-1.12\cdot 10^{-3}{\sqrt {b_{2}/b_{1}}}(b_{1}A_{1})^{0.55}}}$

where

${\displaystyle b_{i}={\frac {f_{i}^{2}}{1+10^{-4}f_{i}^{2}}}}$

and f is the frequency in GHz.