Land use/cover change in the Yellow River Basin based on Google Earth Engine and complex network

The Yellow River Basin plays an important role in China′s economic and social development and ecological security. The natural evolution and human activities in past decades had a profound impact on the Basin. A systematic and scientific understanding of the environmental changes in the Yellow River Basin is an important premise to achieve the high-quality development of the Basin. Using landsat imagery（17080 scenes） and the classification and regression trees algorithm, mapped and analyzed the annual continuous land use/cover change（LUCC） of the Yellow River Basin based on the Google Earth Engine platform and complex network analysis methods, and provided a set of 90 m resolution continuous annual land use/cover maps from 1986 to 2018. The assessment based on 1000 independent validation points collected in Google Earth indicated that these maps achieved an overall accuracy of 82.6% for 7 first-degree classes and 74.7% for 15 second-degree classes. The analysis of the land system showed that the Yellow River Basin presented complex temporal and spatial changes during the study period. The main LUCC patterns included no change or little change, urban expansion with croplands loss, grasslands restoration, orchard and terrace expansion, and forests growth. In the land use/cover transfer network, high coverage grasslands, low-vegetated lands, and the mixed forests were the key land class nodes in the land system of the Yellow River Basin, with high degrees and frequent transfers with other classes. In addition, the nodes of orchard and terrace, croplands, and urban and built-up all had high diversity and closeness and form one relatively active network community structure. The transfers between them were more important land transfer types in the land system. Time series transfer network analysis showed that the land system of the Yellow River Basin had three unstable periods in 1993-1998, 2001-2007, and 2011-2014 with lower average shortest-path length and higher network density and network transitivity. The possible reasons for these three periods are the reclamation of a large amount of agricultural land, the implementation of the Grain for Green program, and the large increase in orchard and terrace, respectively. Based on the data and analysis results, we can further understand the role of the land system in the mutual feedback between society and the environment. The results of LUCC analysis revealed the land use/cover dynamics caused by frequent human ac