Land surface Verification Toolkit (LVT) - A generalized framework for land surface model evaluation

Abstract

Model evaluation and verification are key in improving the usage andapplicability of simulation models for real-world applications. In thisarticle, the development and capabilities of a formal system for land surfacemodel evaluation called the Land surface Verification Toolkit (LVT) isdescribed. LVT is designed to provide an integrated environment forsystematic land model evaluation and facilitates a range of verificationapproaches and analysis capabilities. LVT operates across multiple temporaland spatial scales and employs a large suite of in-situ, remotely sensed and other model and reanalysis datasetsin their native formats. In addition to the traditional accuracy-basedmeasures, LVT also includes uncertainty and ensemble diagnostics, informationtheory measures, spatial similarity metrics and scale decompositiontechniques that provide novel ways for performing diagnostic modelevaluations. Though LVT was originally designed to support the land surfacemodeling and data assimilation framework known as the Land Information System(LIS), it supports hydrological data products from non-LIS environments aswell. In addition, the analysis of diagnostics from various computationalsubsystems of LIS including data assimilation, optimization and uncertaintyestimation are supported within LVT. Together, LIS and LVT provide a robustend-to-end environment for enabling the concepts of model data fusion forhydrological applications. The evolving capabilities of LVT framework areexpected to facilitate rapid model evaluation efforts and aid the definitionand refinement of formal evaluation procedures for the land surface modeling community. © 2012 Author(s). CC Attribution 3.0 License.