A formal likelihood function for parameter and predictive inference of hydrologic models with correlated, heteroscedastic, and non‐Gaussian errors

By G. Schoups, Jasper A. Vrugt. 2010

570 citations recorded by the source corpus.

Subfields: Global and Planetary Change, Ocean Engineering, Water Science and Technology

View Paper: https://www.semanticscholar.org/paper/880b6cc4c4a24af6ce992aa606472775d67b542f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Global and Planetary Change

In Ocean Engineering

In Water Science and Technology

Citations in the corpus

Confronting uncertainty in model-based geostatistics using Markov Chain Monte Carlo simulation (2011, methodology): “This is our experience in time series analysis when fitting hydrologic models to time series of discharge observations (Schoups and Vrugt, 2010).”
Bridging the gap between GLUE and formal statistical approaches: approximate Bayesian computation (2013, methodology): “This includes the 7-parameter hmodel described in detail in Schoups and Vrugt (2010a) and the 13-parameter SAC-SMA model (Burnash et al., 1973).”
Toward diagnostic model calibration and evaluation: Approximate Bayesian computation (2013, methodology): “In the past years, various improvements to equation (7) have been proposed to handle nontraditional residual distributions [Schoups and Vrugt, 2010; Smith et al., 2010].”
A log‐sinh transformation for data normalization and variance stabilization (2012, background): “Various heteroscedastic error models have been proposed for hydrological applications [e.g., Schoups and Vrugt, 2010; Thyer et al., 2009].”
The stationarity paradigm revisited: Hypothesis testing using diagnostics, summary metrics, and DREAM(ABC) (2015, methodology): “This value of the RMSE would be substantially lower if the SAC-SMA model was fitted directly against the observed discharge data using a n-variate Gaussian likelihood function [Schoups and Vrugt, 2010].”
Facets of uncertainty: epistemic uncertainty, non-stationarity, likelihood, hypothesis testing, and communication (2016, background): “…just a matter of finding an appropriate statistical distribution or, alternatively, some non-parametric probabilistic structure for the model residuals (e.g. Schoups and Vrugt 2010, Sikorska et al. 2014), especially when the sample of possible arbitrary uncertainties (or surprises) might be small.”
An adaptive sparse‐grid high‐order stochastic collocation method for Bayesian inference in groundwater reactive transport modeling (2013, methodology): “The informal likelihood function is designed carefully to implicitly account for errors in measurements, model inputs, and model structure and to avoid overfitting to measurement data [Beven and Binley, 1992; Smith et al., 2008; Schoups and Vrugt, 2010; Smith et al., 2010)].”
An uncertain future, deep uncertainty, scenarios, robustness and adaptation: How do they fit together? (2016, background): “…can be expressed as distributions (Beyer and Sendhoff, 2007; Birge and Louveaux, 2011) and the effect of measurement error in observations when estimating parameters is taken into account by specifying properties of the errors, notably in terms of a likelihood function (Schoups and Vrugt, 2010).”
Flow Prediction in Ungauged Catchments Using Probabilistic Random Forests Regionalization and New Statistical Adequacy Tests (2019, background): “A key complication in estimating the quality of flow predictions in ungauged catchments is that observed flow data typically used in posterior diagnostics and model verification (e.g., Gneiting et al., 2007; McInerney et al., 2017; Schoups & Vrugt, 2010, and many others) are not available.”
High‐dimensional posterior exploration of hydrologic models using multiple‐try DREAM(ZS) and high‐performance computing (2012, methodology): “of the snooker update within the context of DE-MC is described by ter Braak and Vrugt [2008]. [11] Despite significant enhancements in the efficiency of MCMC methods, it remains typically difficult to solve very high-dimensional posterior exploration problems involving hundreds or thousands of param”
Development of a formal likelihood function for improved Bayesian inference of ephemeral catchments (2010, background): “[2009] and Schoups and Vrugt [2010]. If the residuals belong to the assumed distribution, the QQ plot should be close to linear.”