This paper presents a hybrid automatic calibration method for transition potential based Cellular Automata land-use models by integrating two calibration methods, process-specific and optimisation-based, into a single hybrid approach, combining the advantages of these two methods. The hybrid approach features the detailed exploration of a large population of possible model parameterisations achieved using optimisation with valuable understanding of land-use systems and their dynamics commonly utilised in process-specific methods to better enhance the plausibility of the results obtained. The utility of the proposed hybrid approach is tested through an application to Madrid, Spain, and outperforms the two other methods (conventional multi-objective optimisation and process-specific) in terms of objective performance, quality of simulated output maps based on visual assessment, and parameter estimates that are more consistent with process understanding.

本文通过将两种基于过程的优化和基于优化的校准方法集成到单个混合方法中，结合了这两种方法的优点，提出了一种基于过渡势的Cellular Automata土地利用模型的混合自动校准方法。混合方法的特点是详细探索大量可能的模型参数化，这些参数化是通过对土地利用系统及其动力学的宝贵理解而进行优化而实现的，这些参数通常用于特定于过程的方法中，以更好地增强所获得结果的合理性。拟议的混合方法的效用通过在西班牙马德里的应用进行了测试，在目标性能方面优于其他两种方法（传统的多目标优化和针对特定过程），