Simultaneously considering the spatial and temporal processes is essential for land cover simulation models. A cellular automaton (CA) usually simulates the spatial conversion of land cover through post-classification comparisons between the beginning and the end of the training period. However, such an approach does not consider the temporal evolution of land cover. As a result, a CA model fails to explain the realistic land cover change. This paper proposes a temporal-dimension-extension CA (TDE-CA) by integrating the temporal evolution of land cover with a CA. In the TDE-CA, the Breaks for Additive Season and Trend (BFAST) monitor algorithm was employed in the temporal evolution simulation module (TESM) to simulate the gradual evolution of land cover, and an optimized random forest CA (optimized RF-CA) was used to simulate the spatial conversion driven by many spatial variables. Subsequently, the Ensemble Kalman Filter (EnKF) was employed to integrate the TESM with the optimized RF-CA. The TDE-CA was then tested in the land cover simulation of Shendong mining area during the period 2005–2015. The TDE-CA was compared with a Null model, with its sub-models, and with the traditional CA models, including the Logistic-CA and the MLP-CA (Multilayer Perceptron CA) models. The results show that the TDE-CA is superior to the Null model. Furthermore, the overall accuracy and the Kappa coefficient of the TDE-CA were 79.84% and 71.61%, respectively; compared with the TESM and the optimized RF-CA, the values showed 17.14% and 4.48% improvements in the overall accuracies and 0.2167 and 0.0512 improvements in the Kappa coefficients, respectively. When compared with the Logistic-CA and the MLP-CA, we measured 8.41% and 8.25% improvements in the overall accuracies and 0.0985 and 0.0964 improvements in the Kappa coefficients. These experiments indicate that the TDE-CA not only provides an effective model for the spatiotemporal dynamical simulation of land cover, but also enhances the development of the existing simulation theory.
This is an open access article distributed under the Creative Commons Attribution License
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited