J. Resour. Ecol. 2011 2(4) 322-327

DOI:10.3969/j.issn.1674-764x.2011.04.005

www.jorae.cn

December, 2011 Journal of Resources and Ecology

Received: 2011-06-21 Accepted: 2011-11-11Foundation: Supported by Ministry of Science and Technology of China (2008FY110300).* Corresponding author: XIE Gaodi. Email: xiegd@igsnrr.ac.cn.

Vol.2 No.4

Article

1 IntroductionThe ecosystem service of water supply refers to the function of vegetation to retain moisture and regulate runoff through redistributing precipitation (Li et al. 2010). Water supply services are affected by vegetation type, climatic change, soil properties, topography and geomorphology. Services show spatial and temporal heterogeneity because of the hydrologic cycle and water balance. It remains unclear how controlling factors affect water supply services at different spatial and temporal scales and this makes it difficult to evaluate water supply services at regional scales and to convert between different scales (Wang et al. 2010). The current evaluation methods for water supply services include soil water storage, water balance, precipitation storage and annual runoff (Zhang et al. 2009). Among them, the water balance method is the most efficient and has been widely applied (Lu et al. 2004; Zhang et al. 2008).

The InVEST model (Integrate Valuation of Ecosystem Services and Tradeoffs) is an ecosystem service appraisal tool developed in 2007 by Stanford University, the World

Wide Fund for Nature and The Nature Conservancy. It has been widely used across many regions such as Hawaii (Goldstein et al. 2010), Indonesia (Barano et al. 2010), Colombia (Goldman et al. 2010) and China (Zhou et al. 2010; Bai et al. 2011). The water yield module is a large spatio-temporal scale model developed on the basis of water balance principles. Taking advantage of this model, we evaluated water supply services at a regional scale in the Lancang River Basin. The aims of this study were: (i) define the spatial distribution of water supply service and the water supply service of the sub-watersheds; (ii) define water supply services of selected typical ecosystems; and (iii) identify hotspots of water supply service where the water supply capacity is more than 5000 m3 ha-1 y-1 in the basin.

2 Study areaThe Lancang River originates on the Qinghai-Tibet Plateau and is a famous transboundary river running throughout six countries. In China, the length of the main stream of the Lancang River is 2160 km with a drainage area of 164 000 km2. Annual runoff is 7.65E+08 m3, representing 14.5%

Modelling Ecosystem Water Supply Services across the Lancang River Basin

CHEN Long1,2, XIE Gaodi1*, ZHANG Changshun1, PEI Sha1,2, FAN Na1,2, GE Liqiang1 and ZHANG Caixia1

1 Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China; 2 Graduate University of Chinese Academy of Sciences, Beijing 100049, China

Abstract: Ecosystem services related to water supply are now a hot topic in ecology and hydrology. Here,

water supply service in the Lancang River basin was evaluated using the newly developed model InVEST.

We found the mean annual water supply in Lancang River basin is approximately 7.24E+10 m3 y-1 with

23.87% from main stream and 76.13% from the tributaries. There is an increasing trend downstream.

Grasslands and forests contribute 71.66% of the total water. A comparison of water supply capacity per

unit area for ecosystems of different composition indicates that there is a decreasing trend from broad-

leafed forest, mixed coniferous and broad-leafed forest, bamboo forest, coniferous forest, shrub forest and

grassland. Two-thirds of the total water is provided by an area covering 40% of the total basin area. This

study provides guidelines for the efficient management of water resources in the Lancang River basin.

Key words: water supply service; InVEST model; ecosystem services; hotspot; Lancang River

CHEN Long, et al.: Modelling Ecosystem Water Supply Services Across the Lancang River Basin 323

of the total river basin. The upper reach extends from the headwaters to Changdu, the middle reach from Changdu to Gongguoqiao, and the lower reach from Gongguoqiao to the Nana River estuary. Most of the major tributaries are located in upstream and downstream reaches (He et al. 2000, 2007). The Lancang River Basin (21°–34°N, 94°–102°E, Fig. 1) experiences large regional differences in climate as it flows from north to south. Temperature and precipitation increase from north to south, and decrease with increasing elevation. The diversity of physiographic settings results in a variety of soil types and ecosystems (Qiu 1996; Wang et al. 2007).

3 Method and data resource3.1 MethodThis study evaluates the water supply services of vegetation in the Lancang River Basin at a regional scale. Runoff can be used to estimate water supply services based on the principle of water balance (Li 2008; Egoh et al. 2008; Nie 2009, 2010), and was calculated by the water yield component in InVEST as follows (Tallis et al. 2011):

(1 )*xjjx x

x

AETY P

P= −

(1)

11

1

xj x xj

xx xj

xj

AET R

P RR

ω

ω

+=

+ + (2)

xx

x

AWCZ

Pω = (3)

xj xxj

x

k EToR

P

⋅= (4)

where Yjx is annual water yield on pixel x with ecosystem j; AETxj is annual actual evapotranspiration on pixel x with ecosystem j; Px is annual precipitation on pixel x; AETxj/Px is an approximation of the Budyko curve developed by Zhang et al. (2001); ωx is a non-physical parameter to characterize the natural climatic-soil; Rxj is the dimensionless Budyko Dryness index on pixel x with ecosystem j, defined as the ratio of potential evapotranspiration to precipitation (Budyko 1974); Z is a seasonality coefficient that represents the seasonal rainfall distribution and rainfall depths ranging from 1 to 10 (Zhang et al. 2001); AWCx is the volumetric (mm) plant available water content; Kxj is the plant (vegetation) evapotranspiration coefficient associated with the ecosystem j on pixel x; ETox is the reference evapotranspiration from pixel x.

3.2 DataPrecipitation (P) from 35 meteorological stations during 1960–2005 was retrieved from the Meteorological Center of China Meteorological Administration. The mean annual value is interpolated in ArcGIS v9.3 by Kriging. Reference evapotranspiration (ETo) is calculated by the modified FAO Penman-Monteith method (FAO 1998a). The evapotranspiration coefficient (K) followed FAO (1998b) and Tallis et al. (2011). The coefficient (Z) proposed by Zhang was considered to be 3, at which the results will have the least error (He 1995; Qiu 1996).

Ecosystem types (ET) and sub-types are derived from land use maps, vegetation maps and grass maps (Table 1). The land use maps (1:100 000) were provided by Data Sharing Infrastructure of Earth System Science, vegetation maps (1:1 000 000) were provided by the Institute of Botany (Chinese Academy of Sciences), and grass maps (1:1 000 000) were provided by the Institute of Geographic Sciences and Natural Resources Research (Chinese Academy of Sciences).

The DEM data was downloaded from the Earth Resources Observation and Science Center (EROS) (http://eros.usgs.gov/#/Find_Data/Products_and_Data_Available/gtopo30/hydro) and the division of watershed and sub-watersheds were operated in ArcSWAT.

Soil depth was obtained from the soil properties database of China, compiled by the Data Center of

Fig. 1 Location of Lancang River Basin.

Yunnan

Sichuan

Tibet

Qinghai

N

km0 300 400

EW

S

100 20050

River

Research area

Province boundry

Journal of Resources and Ecology Vol.2 No.4, 2011324

Resources and Environmental Science (Chinese Academy of Sciences). Plant available water content (AWC) was provided by Zhou (2003), and maximum root depth was obtained from Canadell et al. (1996).4 Results

4.1 Distribution of water supply servicesThe distribution of water supply in the Lancang River basin is shown in Fig. 2. Annual water supply in the whole river basin is 7.24E+10 m3: 9.85E+09 m3 from the upper river basin (13.60%), 1.18E+10 m3 from the middle river basin (16.26%), and 5.08E+10 m3 from the lower river basin (70.15%). Here, we define water supply capacity as the annual water supply per unit area. The water supply capacity in the three different reaches is 1856.69, 3199.82 and 6967.72 m3 ha-1 y-1, respectively, increasing from upstream to downstream.

Annual water supply was also calculated for first order tributaries with an area larger than 1000 km2. The main stream produces 1.73E+10 m3 y-1 with a contribution of 23.87%, while the tributaries supply 5.50E+10 m3 y-1 with a contribution of 76.13% (Fig. 3). The annual water supply in the Buyuan River was the highest among tributaries (7.46E+09 m3 y-1, 10.32%), followed by the Weiyuan River (8.11%) and the Heihui River (7.84%). Water supply capacity in the Buyuan River is also the highest among tributaries, with a value of 9459.70 m3 ha-1 y-1. The water supply capacity of the Nanlei and Hei Rivers is 8933.67 and 8793.96 m3 ha-1 y-1, respectively.

Table 1 Ecosystem types in the Lancang River basin.

Ecosystem Sub-ecosystem Code Area (m2) Percentage (%)Grassland High-cold meadow HM 5.26E+10 32.08

Warm-temperate herbosa WH 1.82E+9 1.11Tropical herbosa TJ 2.07E+10 12.63Montane meadow MM 5.37E+9 3.28

Desert High-cold bare desert HB 7.38E+9 4.5High-cold sparse desert HD 3.17E+9 1.93

Forest Dry valley sparse scrub and grass DS 6.25E+9 3.81Alpine and subalpine shrub forest AS 8.48E+9 5.17Alpine and subalpine broad-leafed forest AB 1.47E+9 0.9Alpine and subalpine mixed coniferous and broad-leafed forest AM 7.44E+7 0.05Alpine and subalpine coniferous forest AN 8.09E+9 4.93Tropical and subtropical bamboo forest TB 3.15E+8 0.19Tropical forest and monsoon rain forest TM 2.41E+9 1.47Subtropical shrub forest SS 4.94E+9 3.01Subtropical broad-leafed forest SB 5.64E+9 3.44Subtropical mixed coniferous and broad-leafed forest SM 2.68E+9 1.63Subtropical coniferous forest SN 5.32E+9 3.24

Farmland Dryland crops DC 1.79E+10 10.93Paddy field PF 7.69E+9 4.69

Human settlement Human settlement HS 2.59E+8 0.16Wetland Wetland WL 1.40E+9 0.85

Fig. 2 Spatial distribution of water supply service in Lancang River Basin.

High: 6259.56

Low: -2622.31No data

km0 330 440110 22055

High: 13562.2

Low: 890.84Subwater sheds

N

EW

S

Water supply (m3 ha-1)

CHEN Long, et al.: Modelling Ecosystem Water Supply Services Across the Lancang River Basin 325

4.2 Water supply services of typical ecosystemsThe study area is dominated by grassland and forest ecosystems, accounting for 49.10% and 27.85% of the region. The water supply services for these two ecosystems are shown in Fig. 4.

The grassland ecosystem supplies 2.28E+10 m3 y-1 (31.26%). The high-cold meadow widely occurs in the upper and middle river basins, which supplies 1.11E+10 m3 (15.40%), followed by tropical herbosa (13.17%) and montane meadow (2.10%); warm-temperate herbosa was the lowest (0.96%). The water supply capacity for grassland ecosystems is 2836.77 m3 ha-1 y-1. Within grassland ecosystems the water supply capacity of tropical herbosa is the highest, with a value of 4545.36 m3 ha-1 y-1, followed by warm-temperate herbosa (3817.66 m3 ha-1

y-1) and montane meadow (2836.77 m3 ha-1 y-1); high-cold meadow is the lowest at 2112.65 m3 ha-1 y-1.

The forest ecosystem provides 2.89E+10 m3 or 40.04% of the total. Annual water supply in the subtropical broad-leafed forest in the lower river basin is 4.75E+09 m3 y-1 (6.57%), in the subtropical coniferous forest is 4.20 E+09 m3 y-1 (5.81%), in the alpine and subalpine coniferous forest is 4.00 E+09 m3 y-1 (5.53%), and in the alpine and subalpine mixed coniferous and broad-leafed forest is 4.19E+07 m3 y-1 (0.06%). Annual water supply decreased with increasing elevation. Water supply capacity in forest ecosystems is 6333.08 m3 ha-1 y-1. It is the highest in tropical forest and monsoon rain forest (9071.59 m3 ha-1 y-1), followed by subtropical broad-leafed forest (8416.42 m3 ha-1 y-1), subtropical coniferous forest (7900.25 m3 ha-1 y-1),

and alpine and subalpine shrub forest (4216.93 m3 ha-1 y-1).Water supply capacity in broad-leafed forest is 8319.98

m3 ha-1 y-1, 7868.94 m3 ha-1 y-1 in mixed coniferous and broad-leafed forest, 7622.29 m3 ha-1 y-1 in bamboo forest, 6117.10 m3 ha-1 y-1 in coniferous forest, 5159.15 m3 ha-1 y-1

in shrub forest, and 2836.77 m3 ha-1 y-1 in grasslands.

4.3 Hotspot of water supply serviceThe study area was divided into four regions according to the water supply capacity of catchments (Table 2). The water supply index, defined as the ratio of the percent of water supply services to the percent of area, was used to characterize different regions (Fig. 5). Regions in the first class have a water supply capacity less than 2000 m3 ha-1 y-1, with a water supply index of 0.36. These regions are located in the upper and middle river basins. The water supply capacity in regions in the second class varies from 2000–5000 m3 ha-1 y-1 and the water supply index is 0.78, mainly in the Angqu River and Heihui River watersheds. Regions in the third class had a variable water supply capacity of 5000–8000 m3 ha-1 y-1 and water supply index of 1.45. These regions include Weiyuan River basin, Liusha River basin and Nana River basin. Regions in the fourth class are characterized by a water supply capacity of 8000 m3 ha-1 y-1 and a water supply index of 2.02, including the Buyuan River basin, Xiaohei River basin, Heihe River basin and Nanlei River basin.

The term “hotspot” was originally used for regions or sites of high species richness and has been widely applied to prioritize areas for biodiversity conservation (van

Table 2 Hotspot classes of water supply services in the Lancang River Basin.

Class Area Water supply Water supply index Water supply capacity(ha) (%) (m3) (%) (m3 ha-1)

1 5.11E+6 31.39 8.11E+9 11.2 0.36 <20002 4.62E+6 28.4 1.61E+10 22.23 0.78 2000–50003 4.18E+6 25.66 2.69E+10 37.21 1.45 5000–80004 2.37E+6 14.56 2.13E+10 29.36 2.02 >8000

Fig. 3 Water supply service of first grade tributary of Lancang River.

Fig. 4 Water supply service of different ecosystems in Lancang River Basin. The code of ecosystem types are shown in Table 1.

4000

3000

Area percentWater supply percent

Zhaq

u

Perc

ent (

%)

8

4

0

6

2 1000

2000

0

Wate

r sup

ply

per a

rea (

m3 ha

-1)8000

7000

5000

6000

10000

9000

16

12

14

10

Water supply per area

Buyu

anN

angu

oH

eihe

Wei

yuan

Bijia

ng

Xia

ohei

Men

gga

Luoz

haYi

njia

ngH

eihu

i

Den

gqu

Biqu

Tong

dian

Mai

quSe

quZi

quA

ngqu

Nan

laN

anla

nN

anle

iN

ana

Lius

ha

4000

3000

Area percentWater supply percent

HM

Perc

ent (

%)

20

10

0

15

51000

2000

0

Wate

r sup

ply

per a

rea (

m3 ha

-1)8000

7000

5000

6000

10000

900030

35

25

Water supply per area

SMSNSSAN DS

AB

AMTJ AS

WH

MM TMTBSB

Ecosystem type

Journal of Resources and Ecology Vol.2 No.4, 2011326

Jaarsveld et al. 1998). This study defines a hotspot as a river basin that has a water supply capacity of more than 5000 m3 ha-1 y-1 and therefore a significant region for water supply services. Generally, hotspots occur in the lower river basin (Fig. 5).

5 DiscussionThis study analyzed water supply service in the Lancang River Basin at a regional scale using the water yield module in InVEST. We conclude:

(1) An average volume of 7.24E+10 m3 y-1 of water is supplied in the Lancang River basin, and increases from upstream to downstream. The contribution of the main stream and tributaries was 23.87% and 76.13%. The Buyuan River supplied the highest volume of water both in total and unit area.

(2) Approximately 72% of the total water was produced by grasslands and forest ecosystems. High cold meadows supplied the highest volume of water in grassland ecosystems and subtropical broad-leafed forest the highest in forest ecosystems. Water supplying capacity per unit area in ecosystems with different composition and structure were compared, and the results showed a decreasing trend from broad-leafed forest, mixed coniferous and broad-

leafed forest, bamboo forest, coniferous forest, shrub forest and grassland.

(3) The study area was divided into four regions according to water supply per unit area. The hotspot covers 40% of the total basin area and provides two thirds of the total water.

This study evaluated the water supply service using the InVEST model, and mapped water supply spatial distribution patterns. The hotspot which produced most water supply services has practical significance for water resources and ecosystem management. However, several factors such as topography and geomorphology were not taken into account and some parameters such as evaporation coefficient and the root depth were defined according to the results of previous studies. Hence, there are some uncertainties in some regions. Nevertheless, this study provides a useful initial exploration into the research of ecosystem services at regional scales.

ReferencesBai Y, Zhuang C, Ouyang Z, et al. 2011. Spatial characteristics between

biodiversity and ecosystem services in a human-dominated watershed. Ecological Complexity, In Press, Corrected Proof.

Barano T, E McKenzie, N Bhagabati, et al. 2010. TEEB case: Integrating Ecosystem Services into Spatial Planning in Sumatra, Indonesia, available at: TEEBweb.org.

Budyko M I. 1974, Climate and Life. New York: Academic Press, 1–510. Canadell J, R B Jackson, J B Ehleringer, et al. 1996. Maximum rooting depth

of vegetation types at the global scale. Oecologia, 108(4): 583–595.Egoh B, B Reyers, M Rouget, et al. 2008. Mapping ecosystem services for

planning and management. Agriculture, Ecosystems and Environment, 127(1–2): 135–140.

FAO. 1998a. Crop Evapotranspiration: Guidelines for Computing Crop Requirements. Italy: Food and Agriculture Org.

FAO. 1998b. Guidelines for computing crop water requirements - FAO Irrigation and drainage paper 56. FAO, Rome, 1–300.

Goldman R L, S Benitez, A Calvache, et al. 2010. TEEB case: Water Funds for conservation of ecosystem services in watersheds, Colombia, available at: TEEBweb.org.

Goldstein J H., G Caldarone, C Colvin, et al. 2010. TEEB case: Integrating ecosystem services into land-use planning in Hawai’i, USA. available at: www.TEEBweb.org.

He D M, Feng Y, Hu J M, et al. 2007. Utilization of water resources and environmental conservation in the international rivers, southwest China. Beijing: Science Press, 1–225. (in Chinese)

He D M, Tang Q C. 2000. China international rivers. Beijing: Science Press, 1-209. (in Chinese)

He D M. 1995. Analysis of hydrological characteristics in Lancang-Mekong River. Yunnan Geographic Environment Research, 7(1): 58–74. (in Chinese)

Li S M, Xie G D, Zhang C X, et al. 2010. Flow process of water conservation service of forest ecosystem. Journal of Natural Resources, 25(4): 585–593. (in Chinese)

Li W H. 2008. Theories, methods and applications of ecosystem services valuation. Beijing: China Renmin University Press, 1–365. (in Chinese)

Lu C X, Xie G D, Xiao Y, et al. 2004. Ecosystem diversity and economic valuation of Qinghai -Tibet Plateau. Acta Ecologica Sinica, 24(12): 2749–2755. (in Chinese)

Nie Y H, Gong B, Li Z. 2010. The spatial-temporal variations of water conservation capacity in Qinghai-Tibet Plateau. Earth Science Frontiers, 17(1):373–377. (in Chinese)

Nie Y H, Gong B, Yi X W. 2009. Water Conservation Valuation of Qinghai-Tibet Plateau. Research of Soil and Water Conservation, 16(5): 210–212, 281. (in Chinese)

Qiu G X. 1996. Research on environmental planning of Lancang River Basin

Fig. 5 Hotspot distribution of water supply service in Lancang River Basin.

1

3

km0 330 440110 22055

2

4subwatersheds

N

EW

S

hotspot class

CHEN Long, et al.: Modelling Ecosystem Water Supply Services Across the Lancang River Basin 327

in Yunnan Province. Kunming: Yunnan Environmental Science Press, 1–213. (in Chinese)

Tallis H T, T Ricketts, A D Guerry, et al. 2011. InVEST 2.0 beta User’s Guide. The Natural Capital Project, Stanford, 1–275.

van Jaarsveld A S, S Freitag, S L Chown, et al. 1998. Biodiversity assessment and conservation strategies. Science, 279, 2106–2108.

Wang J, Cui B S, Yao H R. 2007. Spatio-temporal dynamic study on landscape patterns in Lancang River watershed of Yunnan Province. Journal of Soil and Water Conservation, 21(4): 85–89, 97. (in Chinese)

Wang X X , LI X Y, Mo F, et al. 2010. Exploration of a new modeling method for forest water conservation based on cellular automata: concept and theoretical framework. Acta Ecologica Sinica, 30(20): 5491–5500. (in Chinese)

Zhang B, Li W H, Xie G D, et al. 2008. Characteristics of water conservation of forest ecosystem in Beijing. Acta Ecologica Sinica, 28(11): 5619–5624. (in Chinese)

Zhang B, Li W H, Xie G D, et al. 2009. Water conservation function and its measurement methods of forest ecosystem. Chinese Journal of Ecology, 28(3): 529–534. (in Chinese)

Zhang L, W R Dawes, G R Walker. 2001. Response of mean annual evapotranspiration to vegetation changes at catchment scale. Water Resources Research, 37(3): 701–708.

Zhou B, Yu X X, Chen L H. 2010. Soil Erosion Simulation in Mountain Areas of Beijing Based on InVEST Model. Research of Soil and Water Conservation, 17(6): 9–13, 19. (in Chinese)

Zhou W Z. 2003. A study on available water capacity of main soil types in China based on geographic information system. Nanjing: Nanjing Agricultural University, 1–97. (in Chinese)

澜沧江流域生态系统水源涵养功能研究

陈 龙1,2, 谢高地1, 张昌顺1, 裴 厦1,2, 范 娜1,2, 盖力强1 , 张彩霞1

1 中国科学院地理科学与资源研究所，北京100101；

2 中国科学院研究生院，北京100049

摘要：生态系统的水源涵养功能一直以来都是生态学和水文学研究的热点和难点问题。本文基于新开发的InVEST模型对澜沧

江流域的水源涵养功能进行了研究，得出以下结果：（1）澜沧江流域年均供水7.24E+10 m3 y-1，其中干流贡献率为23.87%，而支

流贡献率为76.13%，从上游至下游呈明显递增趋势。（2）草地生态系统和森林生态系统的水源涵养量贡献率共计达71.66%；对不

同组成结构生态系统类型的水源涵养能力进行比较，阔叶林>针阔混交林>竹林>针叶林>灌木林>草地。（3）澜沧江流域三分之

二的水由占总面积40%的热点区域提供。研究为澜沧江流域水资源的有效管理提供一定依据。

关键词：水源涵养；InVEST模型；生态系统服务；热点；澜沧江