Comparing tightly coupled and loosely coupled paradigms for modeling hydrologic systems
AGU Fall Meeting Abstracts
Loosely coupled modeling architectures enable model developers to represent an environmental system by incorporating processes from multiple scientific disciplines. In contrast, models developed using the more common tightly coupled architecture approach maximize computational performance, but are more restrictive in how they allow modelers to incorporate processes not already implemented within that modeling framework. As the scope of hydrologic models continues to expand, loosely coupled modeling will become a more attractive option for representing complex hydrologic systems. Because loosely coupled modeling approaches have not been widely implemented within the hydrologic community, it is necessary to develop tests to ensure they produce accurate results and have similar performance metrics compared to tightly coupled models. For this reason, the objective of this study is to compare and contrast the process of implementing tightly-coupled and loosely-coupled models of hydrologic systems. The approach is to predict the streamflow from a precipitation event using the Hydrologic Engineering Center's (HEC) Hydrologic Modeling System (HMS) and custom programmed modeling components implemented using the Open Modeling Interface (OpenMI). Each of these two modeling approaches were set up to model the same event using the same process equations over the same spatial and temporal domains. The results showed that the OpenMI components can reproduce the solution of the tightly-coupled HEC-HMS model. There was minimal computational overhead introduced by the OpenMI component communication protocol. The results of this study suggest that loosely-coupled OpenMI components can be used in place of existing, tightly coupled models without compromising predictive accuracy or introducing restrictive computational demand. This work, however, was limited to a simple case study. Future work will be directed at expanding the example to more complex systems to understand how accuracy and performance are effected by model complexity.
Castronova, Anthony M. and Goodall, Jonathan L., "Comparing tightly coupled and loosely coupled paradigms for modeling hydrologic systems" (2009). Civil and Environmental Engineering Faculty Publications. Paper 1229.