rootsumsquares

RootSumSquares (RSS) is the operation of taking the square root of the sum of squares of a set of quantities. In metrology and many branches of science and engineering, it is used to combine independent standard uncertainties by calculating uc = sqrt(sum_i u_i^2). When considering a vector with components a_i, the magnitude is also given by the root of the sum of squares, |A| = sqrt(sum_i a_i^2), which is the Euclidean norm.

RSS is commonly described as quadrature addition when the components are statistically independent and uncorrelated. It is a standard method for uncertainty propagation and is often recommended in measurement guidance such as the Guide to the Expression of Uncertainty in Measurement (GUM). It provides a compact way to represent the combined effect of several error sources on a measured quantity.

Limitations include the need for independence among the contributing uncertainties. If uncertainties are correlated, the simple RSS must be augmented with covariance terms, expressed as uc = sqrt(u^T Cov u). In such cases RSS alone may underestimate or misrepresent the total uncertainty, especially for strong correlations or non-Gaussian error distributions.

Example: two independent standard uncertainties of 3 and 4 units combine to 5 units via RSS. In practice, RSS is widely used to combine instrument resolution, repeatability, and other independent error sources, and is common in experimental physics, engineering, metrology, and sensor fusion. It differs from the root-mean-square (RMS), which uses the mean of squares rather than the sum.