Editing Radar cross section (section)

==Calculation==
Quantitatively, RCS is calculated in three-dimensions as<ref name="Balanis 12"/>

:<math>\sigma = \lim_{r \to \infty} 4 \pi r^{2} \frac{S_{s}}{S_{i}}</math>

Where <math>\sigma</math> is the RCS, <math>S_{i}</math> is the incident [[power density]] measured at the target, and <math>S_{s}</math> is the scattered power density seen at a distance <math>r</math> away from the target.

In electromagnetic analysis this is also commonly written as<ref name=":0" />

:<math>\sigma = \lim_{r \to \infty} 4 \pi r^{2} \frac{|E_{s}|^{2}}{|E_{i}|^{2}}</math>

where <math>E_{s}</math> and <math>E_{i}</math> are the far field scattered and incident [[electric field]] intensities, respectively.

In the design phase, it is often desirable to employ a [[computer]] to predict what the RCS will look like before fabricating an actual object. Many [[iteration]]s of this prediction process can be performed in a short time at low cost, whereas use of a measurement range is often time-consuming, expensive and error-prone.
The linearity of [[Maxwell's equations]] makes RCS relatively straightforward to calculate with a variety of analytic and numerical methods, but changing levels of military interest and the need for secrecy have made the field challenging, nonetheless.

The field of solving [[Maxwell's equations]] through [[numerical analysis|numerical algorithms]] is called [[computational electromagnetics]], and many effective analysis methods have been applied to the RCS prediction problem.
RCS prediction software are often run on large [[supercomputer]]s and employ high-resolution [[Computer-aided design|CAD]] models of real radar targets.

[[High frequency approximation]]s such as [[geometric optics]], [[physical optics]], the [[geometric theory of diffraction]], the uniform theory of diffraction and the physical theory of [[diffraction]] are used when the [[wavelength]] is much shorter than the target feature size.

Statistical models include [[chi-square target models|chi-square]], [[rice distribution|Rice]], and the [[log-normal distribution|log-normal]] target models. These models are used to predict likely values of the RCS given an average value, and are useful when running radar [[Monte Carlo method|Monte Carlo]] simulations.

Purely [[numerical analysis|numerical]] methods such as the [[boundary element method]] ([[Method of moments (electromagnetics)|method of moments]]), [[finite difference time domain method]] ([[FDTD]]) and [[finite element]] methods are limited by computer performance to longer wavelengths or smaller features.

Though, for simple cases, the wavelength ranges of these two types of method overlap considerably, for difficult shapes and materials or very high accuracy they are combined in various sorts of [[hybrid method]].