GUM uncertainty framework, Unscented transformation, and Monte Carlo approaches for the uncertainty evaluation of Brewer UV measurements

The software is implemented in R for the uncertainty evaluation of Brewer #150 UV measurements using three different approaches: an analytical one (the GUM uncertainty framework) and two numerical methods (the Unscented transformation and the Monte Carlo method). The software consists of several files. For the programs to work, the corresponding dataset must be downloaded (https://doi.org/10.5281/zenodo.12790731) and saved in a folder named 'Data' under the same directory as the code files.

"MCM.R": for the uncertainty evaluation based on the Monte Carlo method. It considers ten uncertainty sources (wavelength shift, noise, dark counts, dead time, distance, current, lamp irradiance, stability, temperature and cosine error) and returns the combined standard uncertainty as well as the uncertainties caused by each error source. 

"GUF.R": for the uncertainty evaluation based on the GUM uncertainty framework. It considers ten uncertainty sources (wavelength shift, noise, dark counts, dead time, distance, current, lamp irradiance, stability, temperature and cosine error) and returns the combined standard uncertainty as well as the uncertainties caused by each error source. 

"UT.R": for the uncertainty evaluation based on the Unscented transformation (propagation of the uncertainty using sigma points). It considers ten uncertainty sources (wavelength shift, noise, dark counts, dead time, distance, current, lamp irradiance, stability, temperature and cosine error) and returns the combined standard uncertainty as well as the uncertainties caused by each error source. 

"OAT.R": Function to study the contribution of each uncertainty source to the total uncertainty bugdet of Brewer UV measurements (one at a time). This is obtained by a Monte Carlo simulation with M trials. By default, M = 10 ^ 5

"OAT_unscented.R": Function to study the contribution of each uncertainty source to the total uncertainty budget of Brewer UV measurements (one at a time). This is obtained by defining the sigma points and propagating them through the nonlinear irradiance model. By default, beta = 2 (optimal for GAussian PDFs)

"calculate_irr.R": Function to calculate the measured irradiance by a Brewer.