Abstract: 【Objective】The aims of this paper were to quantify the balance of supply and demand situation of carbon sequestration service in regional ecosystem(hereinafter referred to as "carbon sequestration service"), simulate the spatial flow of carbon sequestration service, reveal the function change spatial rule of carbon sequestration service in regional ecosystem, and give the regional spatial layout optimization strategy which will provide an intuitive scientific reference for guiding the low carbon development of regions.【Method】Population spatial density was simulated with a multi-source data fusion model, and then the demand quantity of carbon sequestration service was estimated in Guanzhong-Tianshui Economic Region. The supply quantity of biologic carbon sequestration service in study area was calculated using CASA model, the soil carbon sequestration service supply was estimated using remote sensing model of carbon cycle. Based on these, the regional carbon balance was quantified with current ratio and the flow process of carbon sequestration service from generation to the use of land was elaborated with spatial visualization method. The conditional probability of environmental variables to the supply of carbon sequestration was calculated rely on Bayes principle, and the key variables were screened by entropy reduction model, the uncertainty of carbon sequestration pattern was discussed by using the distribution of key factor subsets with optimal state, and at last, the carbon sequestration space layout optimization strategy was given.【Result】(1) Carbon sequestration service in the study area overall oversupply, the spatial difference of balance between supply and demand is obvious, the high value areas of demand are mainly distributed in the Guanzhong plain high population areas, the high value areas of supply are mainly distributed in the areas along the Qinling Mountains and the Beishan mountains. (2) The spatial flow of carbon sequestration service in ecosystems is obviously different. According to the current ratio distribution, the study area can be divided into three carbon source concentration centers: Tianshui carbon source concentration center (R_i>0.04), Binxian carbon source concentration center (R_i>0.04), and Xi’an multi-level carbon source concentration center, with the highest value (R_i>0.20). The carbon sequestration service space flow in the study area can be divided into four groups: the flow from the Middle and East of Qinling Mountains, Yongshou County in Beishan Muontain to Guanzhong City Group, the flow from the west of Qinling Mountains to Tianshui city, the flow from Linyou county and Xunyi County to Binxian County, the flow from Tongchuan City, Chengcheng County, Hua County to Pucheng County. (3) According to the conditional probability and entropy calculation, taking set{DEM=3, PET=1} as a key variable optimal state subset of biology carbon pool, which is mainly distributed in Baoji City in southern Qinling Mountains and the southwest corner of Tianshui, where the optimal biological carbon fixation probability can reach 54.36%; taking set {NPP=3, DEM=3} as a key variable in the optimal state subset of soil carbon pool, which is mainly distributed in the Qinling Mountains, along the southwest corner of Tianshui and the northeast corner of the city of Xianyang, where the optimal probability is as high as 92.84%. The suitable areas of biology carbon pool mainly distributed in Tianshui and Wushan County, Qinzhou District and the suitable areas of soil carbon pool are mainly distributed in the middle part of Qinling Mountains district.【Conclusion】In general, carbon sequestration service demand is less than the supply in study area with obvious spatial flow. Taking Tianshui City and the middle areas of Qinling Mountains as the main region to optimize the carbon sequestration function can get a higher probability of good carbon sequestration pattern.

Key words: carbon sequestration, ecosystem services, spatial flow, Bayes, Guanzhong-Tianshui Economic Region