Characterizing the impact of model error in hydrologic time series recovery inverse problems
S.K. Hansen, J. He, V.V. Vesselinov
Advances in Water Resources2018DOI 10.1017/j.advwatres.2017.146.R2
Summary
Hydrologic models are commonly over-smoothed relative to reality, owing to computational limitations and to the difficulty of obtaining accurate high-resolution information. When used in an inversion context, such models may introduce systematic biases which cannot be encapsulated by an unbiased“observation noise”term of the type assumed by standard regularization theory and typical Bayesian formulations.