Dynamic simulation of future date palm plantation (Phoenix dactylifera L.) growth using CA–Markov model and FAO-LCCS data in Algerian dryland oases desert

Food and Agriculture Organization (FAO)-land cover classification system (LCCS) is one of the open access water productivity system products that is widely used in global mapping. In this study, we compared the performance of FAO-LCCS version 1.1 and version 1.2 (FAO-LCCS V. 1.1 and V. 1.2) images for monitoring and predicting LULC changes, mainly date palm plantation (DPP) pattern in Ziban region (Northeast Algeria). Three consecutive pairs of FAO-LCCS V. 1.1 and V. 1.2 images were used for monitoring LULC changes during the years 2009, 2012, and 2015. Decision tree algorithm classification was used to reclassify FAO-LCCS V. 1.1 and 1.2 images to five standard classes during the years 2009, 2012, and 2015. CA–Markov model was applied to simulate future DPP dynamics for the years 2018, 2021, 2024, 2027, and 2030. In case of FAO-LULC V. 1.2, there was no change in the five classes between all the studied years. In case of FAO-LULC V. 1.1, results revealed that from 2009 to 2012, 2012 to 2015, DPP cover increased with 0.44%, and ~ 1%, respectively, and it is expected to increase to 4.5% by 2030. For accuracy assessment, reclassification results and CA–Markov model were validated using kappa index statistics. More than 61% of overall accuracy was computed for each of the six FAO-LULC maps. CA–Markov model reached 88% of overall accuracy. The findings of this study showed reasonably good performance of the CA–Markov model. Besides. FOA-LCCS V. 1.1 data can be considered as an appropriate and useful data in studying LULC changes and vegetation growth at regional scale. These results represent a useful tool for limiting change land uses in the study area.