Evaluation and Calibration of an Agent Based Land use Model Using Remotely Sensed Land Cover and Primary Productivity Data

Identifying and reducing uncertainties in future land use projections are becoming critical in integrated assessments of the climate and social change scenarios. Here, we quantified correspondence between remotely sensed land cover and a model-derived projection of European land use to build a calibration and evaluation framework for land use projection models. For an eight-year period (2006–2013), we compared simulated land uses from a model (CRAFTY-EU), defined as agent functional types, against remotely sensed MODIS land cover. Information between two datasets and spatial complexity are calculated, which allowed the evaluation of the CRAFTY model and calibration of the model parameters. The computational cost was high. Thus more efficient searching algorithms are called for. The evaluation framework holds promise for better calibration of land use decision models to increase model usability and improve the value of future land cover projections.