Organiccarbon and climate change on a regional scale

Safir, G.R., S.H. Gage, M. Colunga-Garcia, P.R. Grace, H. Yang, A. Dobermann, S. Rowshan, J. Qi, and G.P. Robertson.

Presented at the All Scientist Meeting (2004-10-08 )

Modeling at regional scales is complex, and thus it has only been attempted by research groups with the necessary combination of biology, modeling and computer science skills. A Modeling Applications Integrative Framework (MASIF) was developed to process the large amounts of spatial-temporal outputs from regional scale simulation experiments. MASIF incorporates existing commercial software capacities that provide the model analyst with the ability to rapidly summarize model results. One objective in developing MASIF was to provide an environment where a modeler can concentrate on developing a model that simulates real world processes without being encumbered with developing and integrating peripheral data manipulation and analytical tools that exist in the marketplace. MASIF enables visualization of model input data, model simulation results, and statistical patterns of information associated with model inputs and outputs. Two levels of utilization of MASIF for supporting regional analyses were implemented. One level was to evaluate the potential for incorporating models (SOCRATES, Sinclair’s MAIZE, HYBRID-MAIZE), that project carbon outputs directly into MASIF and to utilize MASIF as a primary geospatial data I/O and data analysis platform. The second level is the development of procedures to couple model output streams from geospatial simulation models to MASIF. This system enables comparison of model outputs with observed data (e.g. remotely sensed information, agricultural yield statistics), cross-model comparisons and evaluations and to utilize the powerful suite of data analysis and visualization capabilities incorporated in MASIF. This framework provides the ability to link any simulation model to some of the general-purpose utilities needed to analyze geospatial and temporal simulation data and output. The MASIF environment provides us with a reasonably powerful and flexible environment that allows us to conduct creative regional experiments with complex or simple regional models. There is value for both stand-alone and network-based analytical environments for model assessment and application.

Back to meeting | Show |
Sign In