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

doi:10.1007/s11442-019-1642-x

Abstract

Abstract:

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

Li ZENG, Jing LI. A Bayesian belief network approach for mapping water conservation ecosystem service optimization region[J].Journal of Geographical Sciences, 2019, 29(6): 1021-1038.

Figures/Tables 13

Figure 1

Table 1

Figure 2

Table 2

State classification of water conservation factors"

No.	Variables	State	Rank of state	Range
1	Precipitation (mm/a)	Highest	1	817-996
		High	2	681-817
		Medium	3	529-681
		Low	4	0-529
2	Vegetation type	Highest	1	Evergreen broad-leaved forest, evergreen coniferous forest, deciduous broad-leaved forest, deciduous coniferous forest, arbor garden, sparse forest, mixed coniferous and broad-leaved forest
		High	2	Shrubs, deciduous broad-leaved shrubs, arbors, sparse shrubs
		Medium	3	Herbaceous green space, herbaceous marsh, grass, dry land, paddy field, sparse grassland, grassland
		Low	4	Mining, industrial land, rivers, lakes, transportation, residential, open ground, bare soil, bare rock, reservoirs, canals, desert / sandy land
3	Land use	Cropland	1	Cropland
		Forest	2	Forest
		Grassland	3	Grassland
		Water area	4	Water area
		Urban land	5	Urban land
		Unused land	6	Unused land
4	Soil type	Highest	1	Swampy soil, paddy soil, silted black soil, tidal soil, gray-brown soil type silt, cinnamon soil type silt, slightly salinized silt
		High	2	Subalpine meadow soil, mountain brown soil, mountain leached cinnamon soil, dark brown soil, brown soil, mountain grassland soil, coarse bone soil, cinnamon soil, yellow-brown soil, mountain meadow grassland soil
		Medium	3	Cultivation of mountain leached cinnamon soil, cultivation of mountain cinnamon soil, cultivation of mountain carbonate cinnamon soil, cultivation of mountain meadow grassland soil, cultivation of mountain cinnamon soil, cultivation of mountain carbonate
		Low	4	Alluvial soil, new soil, purple soil, red soil, aeolian sand, yellow cinnamon soil, lithic soil, loessial soil, limestone soil
5	Evapotranspiration (mm/a)	Highest	1	10913-65535
		High	2	6003-10913
		Medium	3	4668-6003
		Low	4	1765-4668
6	Surface runoff (mm/a)	Highest	1	59-89
		High	2	34-59
		Medium	3	15-34
		Low	4	0-15
7	Water conservation (t/ha·a)	Highest	1	1103-1468
		High	2	917-1103
		Medium	3	163-917
		Low	4	0-163

Table 2

Table 3

Figure 3

Figure 4

Table 4

Table 5

Figure 5

Table 6

Figure 6

Figure 7

References 28

[1]	Brauman K A, Daily G C, Duarte T K et al., 2007. The nature and value of ecosystem services: An overview highlighting hydrologic services. Social Science Electronic Publishing, 32: 67-98.
[2]	Breyer B, Zipper S C, Qiu J X, 2018. Sociohydrological impacts of water conservation under anthropogenic drought in Austin, Texas (USA). Water Resources Research, 54(4): 3062-3080. doi: 10.1002/wrcr.v54.4
[3]	Chen S H, Pollino C A, 2012. Good practice in Bayesian network modelling. Environmental Modelling & Software, 37(17): 134-145.
[4]	Chen Y, Wang G J, 2005. Progress of the research in LUCC forcasting models. Journal of Liaoning Normal University (Natural Science Edition), 28(4): 484-487. (in Chinese)
[5]	Cheng X Y, Zhang N, Wu F F, 2014. Impacts of rainfall and land use on urban surface runoff: A case study of area surrounding the North Moat in Beijing, China. Journal of Natural Resources, 29(8): 1391-1402. (in Chinese)
[6]	Congalton R G, 1991. A review of assessing the accuracy of classifications of remotely sensed data. Remote Sensing of Environment, 37(1): 35-46. doi: 10.1016/0034-4257(91)90048-B
[7]	Daily G C, 1997. Nature’s services: Societal dependence on natural ecosystems.Pacific Conservation Biology, 6(2): 220-221.
[8]	Dang K B, Windhorst W,Burkhard B et al., 2019. A Bayesian Belief Network-based approach to link ecosystem functions with rice provisioning ecosystem services. Ecological Indicators, 100: 30-44. . 2018.04.055.
[9]	Fan Y N, Liu K, Chen S Set al., 2017. Spatial pattern analysis on water conservative fuctionality of land ecosystem in northern slope of Qinling Mountains. Bulletin of Soil and Water Conservation, 37(2): 50-56. (in Chinese)
[10]	Ferro N D, Quinn C, Morari F, 2018. A Bayesian belief network framework to predict SOC dynamics of alternative management scenarios. Soil and Tillage Research, 179: 114-124. doi: 10.1016/j.still.2018.01.002
[11]	Fu B J, Zhang L W, 2014. Land-use change and ecosystem services: Concepts, methods and progress. Progress in Geography, 33(4): 441-446. (in Chinese)
[12]	Gan H, Liu Y S, Wang D W, 2004. Simulation and analysis of the human driving factors of land use type conversion. Resources Science, 26(2): 88-93. (in Chinese)
[13]	Gonzalez R J, Luque S, Poggio Let al., 2016. Spatial Bayesian belief networks as a planning decision tool for mapping ecosystem services trade-offs on forested landscapes. Environmental Research, 144: 15-26. doi: 10.1016/j.envres.2015.11.009
[14]	Guo Y F, 2011. Land use change and its impacts on water conservation services in Jiangxi Province based on CLUE model [D]. Wuhu: Anhui Normal University. (in Chinese)
[15]	Landuyt D, Broekx S, D’Hondt Ret al., 2013. A review of Bayesian belief networks in ecosystem service modelling.Environmental Modelling & Software, 46(C): 1-11.
[16]	Landuyt D, Broekx S, Engelen Get al., 2016a. The importance of uncertainties in scenario analyses: A study on future ecosystem service delivery in Flanders. Science of the Total Environment, 553: 504-518. doi: 10.1016/j.scitotenv.2016.02.098
[17]	Landuyt D, Broekx S, Goethals P L M, 2016b. Bayesian belief networks to analyse trade-offs among ecosystem services at the regional scale. Ecological Indicators, 71: 327-335. doi: 10.1016/j.ecolind.2016.07.015
[18]	Li T, Li J, Wang Y Z, 2019. Carbon sequestration service flow in the Guanzhong-Tianshui Economic Region of China: How it flows, what drives it, and where could be optimized? Ecological Indicators, 96: 548-558. doi: 10.1016/j.ecolind.2018.09.040
[19]	Nelson E, Mendoza G, Regetz Jet al., 2009. Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales. Frontiers in Ecology and the Environment, 7(1): 4-11. doi: 10.1890/080023
[20]	Pamukcu P, Erdem N, Serengil Yet al., 2016. Ecohydrologic modelling of water resources and land use for watershed conservation. Ecological Informatics, 36(Suppl. C): 31-41. doi: 10.1016/j.ecoinf.2016.09.005
[21]	Pearl J, 1988. Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. San Francisco: Margan Kaufmann, 1022-1027.
[22]	Pérez-Miñana E, 2016. Improving ecosystem services modelling: Insights from a Bayesian network tools review.Environmental Modelling & Software, 85: 184-201.
[23]	Shang H, 2007. Study on the water resources and ecologieal environment in Weihe River. Research of Soil and Water Conservation, 14(5): 89-92. (in Chinese)
[24]	Vigerstol K L, Aukema J E, 2011. A comparison of tools for modeling freshwater ecosystem services. Journal of Environmental Management, 92(10): 2403-2409.
[25]	Wang S M, Zhang Z H, 2017. The systematic coupling of landscape pattern, ecological process and ecosystem services: A review of Fu Bojie’s ideas of landscape ecology. Journal of Poyang Lake,(2): 78-83, 127. (in Chinese)
[26]	Xiao H, Ouyang Z Y, Zhu Jet al., 2000. Forest ecosystem services and their ecological valuation: A case study of tropical forest in Jianfengling of Hainan Island. Chinese Journal of Applied Ecology, 11(4): 481-484. (in Chinese)
[27]	Xue J, Gui D W, Lei J Qet al., 2017. A hybrid Bayesian network approach for trade-offs between environmental flows and agricultural water using dynamic discretization. Advances in Water Resources, 110: 445-458. doi: 10.1016/j.advwatres.2016.10.022
[28]	Zhan C S, Qiao C, Xu Z Xet al., 2011. Ecological landscape patterns in Guanzhong part of the Weihe River Basin based on remote sensing. Resources Science, 33(12): 2349-2355. (in Chinese)

Vegetation Type	Precipitation	Evapotranspiration
Vegetation Type	Precipitation	Highest	High	Medium	Low
Highest	Highest	0.0647	52.15	38.991	8.794
Highest	High	0.104	55.123	37.712	7.061
Highest	Medium	0.16	58.139	35.251	6.451
Highest	Low	0.314	80.038	19.586	0.0628
High	Highest	1.036	48.705	26.425	23.834
High	High	1.104	25.153	49.202	24.54
High	Medium	1.139	22.322	47.41	29.129
High	Low	0.704	48.415	50.792	0.088
Medium	Highest	0.813	0.675	12.462	86.049
Medium	High	3.265	2.664	41.71	52.36
Medium	Medium	2.248	5.154	34.65	57.948
Medium	Low	1.249	8.005	90.68	0.0662
Low	Highest	4.66	1.553	9.709	84.078
Low	High	32.642	2.554	26.257	38.547
Low	Medium	15.281	2.488	23.099	59.133
Low	Low	17.687	4.082	77.751	0.68

Region	GDP growth rate (%)				Growth rate of permanent population (%)
Region	2015	Planning	Protection	Development	2015	Planning	Protection	Development
Xi’an	11.53	12	10	15	4.64	6	4.5	8
Weinan	7.71	7.5	6	9	3.46	4	3.5	6
Baoji	6.27	6.5	5.5	8	3.55	4	3.5	6
Tongchuan	0.65	0.6	0.5	1	3.79	4	3.8	4.5
Xianyang	12.08	13	11	15	3.98	5	4.0	6
Yangling	13.57	14	10	15	4.93	6	5.0	6.5
Tianshui	8.9	9	8	10	0.35	0.4	0.3	0.6

The rate of land use change	Protection (%)	Planning (%)	Development (%)
Cropland	-13.22	-8.42	-11.52
Forest	18.12	13.07	13.27
Grassland	0.73	-1.00	-0.94
Water area	-5.08	-5.05	-5.00
Urban land	0.38	2.25	5.04
Unused land	-0.92	-0.85	-0.85

Node name	Variance reduction	Relative percentage
Water conservation	1.49285	100
Precipitation	0.53235	35.7
Evapotranspiration	0.11014	7.38
Land use (2010)	0.10628	7.12
Surface runoff	0.05903	3.95
Vegetation type	0.02189	1.47
Soil type	0.00015	0.00975