Journal of Geographical Sciences ›› 2019, Vol. 29 ›› Issue (6): 1021-1038.doi: 10.1007/s11442-019-1642-x

• Regular Articles • Previous Articles     Next Articles

A Bayesian belief network approach for mapping water conservation ecosystem service optimization region

Li ZENG, Jing LI*()   

  1. School of Geography and Tourism, Shaanxi Normal University, Xi'an 710119, China
  • Received:2018-10-15 Accepted:2018-12-21 Online:2019-06-25 Published:2019-06-25
  • Contact: Jing LI
  • About author:

    Author: Zeng Li (1993-), Master, specialized in resources and environment remote sensing and GIS.E-mail:

  • Supported by:
    National Natural Science Foundation of China, No.41771198, No.41771576;The Fundamental Research Funds For the Central Universities, Shaanxi Normal University, No.2017CSY011


Water conservation is one of the most important ecosystem services of terrestrial ecosystems. Identifying the optimization regions of water conservation using Bayesian belief networks not only helps develop a better understanding of water conservation processes but also increases the rationality of scenario design and pattern optimization. This study establishes a water conservation network model. The model, based on Bayesian belief networks, forecasts the distribution probability of the water conservation projected under different land use scenarios for the year 2050 with the CA-Markov model. A key variable subset method is proposed to optimize the spatial pattern of the water conservation. Three key findings were obtained. First, among the three scenarios, the probability of high water conservation value was the largest under the protection scenario, and the design of this scenario was conducive to the formulation of future land use policies. Second, the key influencing factors impacting the water conservation included precipitation, evapotranspiration and land use, and the state set corresponding to the highest state of water conservation was mainly distributed in areas with high annual average rainfall and evapotranspiration and high vegetation coverage. Third, the regions suitable for optimizing water conservation were mainly distributed in the southern part of Maiji District in Tianshui, southwest of Longxian and south of Weibin District in Baoji, northeast of Xunyi County and northwest of Yongshou County in Xianyang, and west of Yaozhou District in Tongchuan.

Key words: water conservation, ecosystem services, Bayesian belief network, scenario analysis, spatial suitability, land use