A multiple criteria decision support system for testing integrated environmental models

Abstract

Spatial models of ecological and hydrological processes are widely used tools for studying natural systems over large areas. However, these models lack specific mechanisms for reporting output uncertainty contributed by model structure, and so testing their suitability for studying a large range of problems is difficult. This paper describes a method of evaluating the uncertainty contributed by underlying assumptions used in constructing integrated environmental models from two or more sub-models that were developed for different purposes. Integrated environmental models are typically constructed from many individual process-based models. Conflicting assumptions between these sub-models, e.g., spatial scale differences, are easily overlooked during model development and application. This “semantic error” cannot be predicted prior to simulation, as it may only emerge through the interaction of sub-models applied to a particular set of data used to drive a simulation. Model agreement is proposed and demonstrated as a way to detect problems of model integration at the state variable level within an integrated ecosystem model. This model agreement is then propagated to model response variables using multiple criteria to examine their sensitivity, predictability, and synchronicity to the measured uncertainty in state variables. These three properties are combined under fuzzy logic in order to provide decision support on where, for a given time during simulation the sub-models agree on a particular response variable. This paper describes the details of the approach and its application using an existing integrated environmental model. The results show that, for a given set of model inputs and application, integrated environmental models may have spatially variable levels of agreement at the sub-model level. The results using RHESSysD, a spatially integrated ecosystem hydrology model, indicate that semantic error in estimates of plant available soil moisture are consistent with observations of the need for resetting events, such as flooding, to initialize the model to a point where further simulation results can be trusted. These results suggest that a dynamic selection of sub-models may be warranted given a reasonable method of determining sub-model disagreement during simulation. Fuzzy set theory may be a useful tool in arriving at such a model selection process as it allows for a relatively straightforward synthesis of numerous model evaluation criteria with a large quantity of output from the model.