Bardsley Coefficient (B) fit function

Calculates the Bardsley coefficient between observed and modelled values, combining the coefficient of determination (\(R^2\)) and Nash-Sutcliffe Efficiency (\(NSE\)).

Usage

bardsley(obs, sim, na.rm = TRUE)

Arguments

obs

Numeric vector of observed values.

sim

Numeric vector of simulated (modelled) values.

na.rm

Logical; should missing values (NA) in either vector be removed pairwise before computation? Default is TRUE.

Value

numeric; Bardsley coefficient. Returns NA if undefined (e.g., zero observed variance or singular model fit).

Details

The Bardsley coefficient is calculated as: $$B = \frac{R^2}{2 - \text{NSE}}$$

It provides a model performance metric that jointly accounts for correlation and efficiency characteristics.

See also

nse for the Nash-Sutcliffe Efficiency component.

Examples

obs <- c(1, 2, 3, 4)
sim <- c(1.1, 2.1, 2.9, 4.2)
bardsley(obs, sim)
#> [1] 0.9771887