شماره ركورد كنفرانس :
581
عنوان مقاله :
Application of Model Selection Techniques in Groundwatre Modeling
عنوان به زبان ديگر :
Application of Model Selection Techniques in Groundwatre Modeling
پديدآورندگان :
Ehtiat Majid نويسنده , Mousavi S. Jamshid نويسنده
كليدواژه :
Model selection , AIC , Groundwater Modeling , AICC
عنوان كنفرانس :
پنجمين كنفرانس مديريت منابع آب ايران
چكيده لاتين :
Modeling in general and groundwater modeling in particular are subject to issues of simplicity vs.complexity and accuracy vs. uncertainty. This recognition has led to a growing tendency among hydrologists to postulate several alternative hydrologic models for a given site. In this case, model selection techniques may be used as effective yet practical just-in-time approaches to rank these candidate models and select the ‘best’ one among them and/or attempt to construct a set of ‘good’ models. However, the application of these techniques must be with cognizance. In the present study, at first the steps of model selection process are briefly reviewed.Then, Akaike information criterion (AIC) and ‘consistent’ AIC model selection criterion (AICc) are used for alternative conceptualizations of a groundwater model of Dehloran aquifer located in southwest of Iran. The models differ in their number of calibrated parameters and observations. The model selection procedure is performed after both steady state and transient calibrations while the spatial distributions of hydraulicconductivity and specific yield parameters are determined separately in steady-state and transient modes, respectively. The model with more adjusted parameters in steady state calibration was evaluated by AIC as the best option, while AICc opted a simple model. Nonetheless, both AIC and AICc performed the same considering transient calibrated models. Results indicated that the ratio of number of observations to number of parameters affects the performance of the model selection criteria considerably.
شماره مدرك كنفرانس :
4462029