[Spatial autocorrelation analysis of land use and ecosystem service value in Maduo County, Qinghai Province, China at the grid scale].

Research on the spatial quantitative evaluation of land use and ecosystem service value in the source region of the Yellow River is of great significance for ensuring the ecological security of the river basin. Taking Maduo County in the source region of the Yellow River as an example, based on a 3 km × 3 km grid unit, the spatial autocorrelation method was applied to construct an evaluation model of ecosystem service value. The spatial autocorrelation pattern characteristics of Maduo County in 2015 was quantitatively evaluated, and the spatial information of ecosystem service value was visually expressed. The results showed that, at the examined grid scale, the area of grasslands with different coverages was large, and water grid area accounted for 42.9% of the total grids and was mainly distributed in the northwest of Maduo County. The construction land showed a "line-like" distribution from northeast to southwest, while the unutilized land was more distributed in southwest and less in northeast. In space, different land use type grids interacted with each other, with positive correlation and cluster distribution. The values of global Moran I and local Moran I of water area was the largest, with strongest spatial aggregation and high local connectivity. The global Moran I and local Moran I values of the construction land were 0.293 and 0.127, respectively, with the weakest spatial autocorrelation and clustering characteristics in a small range. In 2015, ecosystem service value (ESV) of Maduo County was 93.887 billion yuan, the mean ESV across all the grids was 3.20×107 yuan, with a maximum of 19.96×107 yuan. The water distribution grid had high ESV. On the whole, the ESV distribution pattern in Maduo County had a significant spatial positive correlation, with clustered ESV grids. The ESV grids of different land use types generally showed high-high cluster and low-low cluster, with the spatial pattern of high-low cluster and low-high cluster being sporadic. We proposed several possible strategies of land space planning and use control. First, the water and unused land should adopt a "centralized continuum" protection mode that emphasize the value of ecological spillovers. Second, grasslands with different coverage levels should adopt a "group-type" eco-governance model of primary and secondary division and zoning management. Third, construction land should adopt a small-scale intensive development and utilization model in a "corridor " distribution pattern.