Towards robust statistical inference for complex computer models. (30th March 2021)

• Record Type: Journal Article
• Authors: Oberpriller, Johannes; Cameron, David R.; Dietze, Michael C.; Hartig, Florian
• Editors: Coulson, Tim
• Is Part Of: Ecology letters. Volume 24:Number 6(2021)
• Page Start: 1251
• Page End: 1261
• Publication Date: 2021-03-30
• Subjects: Bayesian Inference -- bias correction -- biased models -- data imbalance -- robust inference; Ecology -- Periodicals; 577
• DOI: 10.1111/ele.13728
• Languages: English
• ISSNs: 1461-023X

Abstract

Abstract: Ecologists increasingly rely on complex computer simulations to forecast ecological systems. To make such forecasts precise, uncertainties in model parameters and structure must be reduced and correctly propagated to model outputs. Naively using standard statistical techniques for this task, however, can lead to bias and underestimation of uncertainties in parameters and predictions. Here, we explain why these problems occur and propose a framework for robust inference with complex computer simulations. After having identified that model error is more consequential in complex computer simulations, due to their more pronounced nonlinearity and interconnectedness, we discuss as possible solutions data rebalancing and adding bias corrections on model outputs or processes during or after the calibration procedure. We illustrate the methods in a case study, using a dynamic vegetation model. We conclude that developing better methods for robust inference of complex computer simulations is vital for generating reliable predictions of ecosystem responses. Abstract : Model error is a major problem for statistical inference with complex computer simulations due to their more pronounced nonlinearity and interconnectedness. Here, we propose a framework for robust inference including rebalancing of data and adding bias corrections on model outputs or processes during or after calibration. We conclude that methods for robust inference of complex computer simulations are vital for generating useful predictions of ecosystems responses.