An adaptive surrogate modeling-based sampling strategy for parameter optimization and distribution estimation (ASMO-PODE)

- An adaptive surrogate modeling-based sampling strategy for distribution estimation was developed.
- The method was tested with 2 test problems and one land surface model.
- The effectiveness and efficiency of the ASMO-PODE method was evaluated.

Parameter distribution estimation has long been a hot issue for the uncertainty quantification of environmental models. Traditional approaches such as MCMC (Markov Chain Monte Carlo) are prohibitive to be applied to large complex dynamic models because of the high computational cost of computing resources. To reduce the number of model evaluations required, we proposed an adaptive surrogate modeling-based sampling strategy for parameter distribution estimation, named ASMO-PODE (Adaptive Surrogate Modeling-based Optimization - Parameter Optimization and Distribution Estimation). The ASMO-PODE can provide an estimation of the parameter distribution using as little as one percent of the model evaluations required by a regular MCMC approach. The effectiveness and efficiency of the ASMO-PODE approach have been evaluated with 2 test problems and one land surface model, the Common Land Model. The results demonstrated that the ASMO-PODE method is an economic way for parameter optimization and distribution estimation.

Illustration of the resampling procedure of the ASMO-PODE. (a) The blue points are initial sample points and the gray points are the estimated distribution generated by the Metropolis algorithm on a surrogate model. (b) The points of estimated distribution are divided into the following quartiles: top 25%, 25%-50%, 50-75% and the bottom 25%. Darker points have higher probabilities, and the black point in the center has the highest probability. (c) Red points are the selected resampled points that have the largest distance to their nearest neighbors. Select one point in each quartile range and also the most probable point in the center. (d) Resampled points after 5 iterations are shown as red points.139