Predicting the Effects of Different Land-Use Scenarios on Water Availability Using a Hydrological Model

Predicting the Effects of Different Land-Use Scenarios on Water Availability Using a Hydrological Model

Ambika Khadka, MESc 20131


Water availability is a serious issue facing many communities and nations around the world due to rapid urbanization and population growth. Hydrological indicators can be used to assess how land-use changes affect water availability. I used the Soil and Water Analysis Tool (SWAT) to predict the effects of land-use changes on surface runoff, infiltration, and peak flow in the Xinjiang River, China, sub-watershed with six land-use scenarios. The major finding of the study is that the presence of forests helps regulate wet-season peak flow. The hydrograph with presence of forest is gradual, thus suggesting a reduction in flood potential in the wet season and water scarcity in the dry season. Installation of earthen ponds in the outlets of sub-basins with higher surface runoff provided the most significant hydrological improvements. The study will provide essential information to aid decision making in long-term land-use planning to protect water resources, and to mitigate the effects of water scarcity and flooding.


Rapid urbanization and other land-use changes related to population growth alter the hydrological regime by increasing the peak flow and volume of surface runoff, while decreasing infiltration (Sahin and Hall 1996). In the 1980s, after most of the existing vegetation was removed for firewood and industrial use, the Xinjiang River basin area was called the “red desert of southern China” (Zheng et al. 2008). The devastating floods of 1981 and 1998 in the Yangtze River were thought to be caused by removal of vegetation in the hills (Wei et al. 2008). In the last two decades, the population centers in southern China have expanded at the expense of forested areas and floodplains. This region has faced extreme water scarcity and flooding, which the government blamed on loss of tree cover as a result of rapid urbanization (Khadka et al. 2013). These catastrophic floods spurred hydrological research and motivated China to adopt the Natural Forest Conservation Program (NFCP). The objectives of the NFCP include restoring natural forests in ecologically sensitive areas such as headwaters of major rivers (ibid.). The NFCP program resulted in the reforestation of denuded hills in the region starting in 1998.

Knowing the hydrological responses to forest cover changes is essential to the question of how ongoing land-use changes influence water availability, potential of flooding, and water scarcity within a watershed. The purpose of this paper is to predict how different land-use scenarios affect the water availability in China’s Xinjiang River sub-watershed using a hydrologic model. The model represents and quantifies the water movement and distribution within the watershed which can be used to create optimal water management options for decision-making processes.


The Xinjiang River drains a 6,168 km2 watershed (Fig. 1) within the Jiangxi Province, located in southeastern China at 28.500° S and 117.500° E. This part of China is undergoing rapid changes in land-use that is likely to affect its water availability. Shangrao and Yintang are two major cities that are expanding within the watershed. During my fieldwork in the summer of 2012, I observed that many stretches of the river floodplain are used for human habitation. The current land-cover is a vast stretch of urban area along floodplain, and pine cultivation on hills which were once heavily denuded. The river is located in a narrow valley between two mountains, Wuyi and Huaiyu. The site is in a wet climatic zone with mean annual precipitation of 1,878 mm with an average surface evaporation of 1,044 mm for the period from 1953 – 2002. The annual mean temperature of the watershed is 18°C in October, with a mean temperature of 37°C in July (Guo et al. 2008). The terrain is rugged; elevations vary from below sea level to 2,100 m.

Figure 1. Project Site.


The study methodology involved compiling a Digital Elevation Model (DEM)2, soil properties, vegetation types, land-use and hydrological data of the Xinjiang River sub-watershed. The weather data for 2000 – 2010 from 15 stations around the study site was obtained from the National Center for Environmental Protection. The Soil and Water Assessment Tool (SWAT) was used to represent the Xinjiang River watershed in order to predict the effects of land-use changes on surface runoff, infiltration, and peak flow. SWAT is a physical based model developed “to predict the impact of land management practices on water, sediment and agricultural chemical yields in large complex watersheds with varying soils, land-use and management conditions over long periods of time” (Neitsch et al. 2009). In order to test the accuracy of the model prediction, the predicted flow from the model was calibrated against monthly flow measurements made over a two-year period using the SUFI2 automatic calibration program (Abbaspour 2011). The calibration was performed by iteration and permutation of eight model parameters to mimic the Xinjiang River sub-watershed in the model. The eight parameters were then used in a predictive simulation of hydrology for land-use of the year 2000 and five other land-use scenarios as shown in Table 1. The year 2000 was used as a base year because it was the land-cover map generated by survey conducted by the Land Management Bureau of Jiangxi Province for the year 2000.

Table 1. The predictive land-use scenarios used in the model for simulation.

Scenario Land-use Type Condition
Scenario 1 Land-use of the year 2000 Land-use map of the year 2000 from Land Management Survey Bureau, Jiangxi Province, 2000.
Scenario 2 100% broad leaf mixed forest When scenario 1 is changed entirely to broad leaf mixed forest uses.
Scenario 3 100% urban When scenario 1 is changed entirely to urban uses.
Scenario 4 50% broad leaf mixed forest, 20% agriculture, 20% urban, and 10% rangeland When scenario 1 is managed as 5:2:2:1 (broad leaf mixed forest: agricultural land: urban area: rangeland).
Scenario 5 1km riparian buffer strips are created along the main river as well as all tributaries. The 1km buffer is forested with broad leaf mixed forest. The 1km riparian buffer is an addition to the scenario 1.
Scenario 6 80,000m2 earthen irrigation ponds are installed in outlet of sub-basins with maximum runoff. The earthen irrigation ponds of various sizes are installed to collect 60% of the precipitation during flood months June and July for 60 days.


Table 2 shows the findings for land-use in 2000 and five other predictive land-use scenarios to show the effect of each land-use regime on hydrological components. The hydrological response of six land-use scenarios is represented as hydrological indices in Table 3.

Table 2. Effect of six land-use scenarios on hydrological components.

Landuse scenario Surface runoff in mm/year Infiltration in mm/year Evaporation in mm/year Peak flow in m3/second
Scenario 1 341 386 677 125
Scenario 2 149 454 733 120
Scenario 3 382 384 667 126
Scenario 4 230 433 705 123
Scenario 5 195 625 665 119
Scenario 6 183 554 666 112

Table 3. Hydrological indices of six land-use scenarios.

Land-use scenario Surface runoff/ precipitation Infiltration/ precipitation Evaporation/ precipitation Peak flow/precipitation
Scenario 1 0.49 0.18 0.33 0.66
Scenario 2 0.35 0.22 0.35 0.63
Scenario 3 0.5 0.18 0.32 0.66
Scenario 4 0.43 0.21 0.34 0.64
Scenario 5 0.32 0.33 0.35 0.6
Scenario 6 0.29 0.36 0.35 0.58


From Table 2, all scenarios with forest cover show a reduction in surface runoff, peak flow and an increase in infiltration. Surface runoff and peak flow are positively correlated with unforested land in a watershed. Peak flow is the maximum flow of the season which directly contributes to floods. The model has predicted that with Scenario 1, 49% of total precipitation flows as surface runoff, 66% flows as peak flow to Xinjiang River, and only 18% of precipitation infiltrates into the soil in the form of groundwater (Table 3). Scenario 1 represents the condition in the Xinjiang River sub-watershed where most of the natural forests had been heavily deforested and replaced by scattered pines (Pinus massoniana) and grasses with low nutritional value (Cheng-Fan 1990) as a part of the NFCP project. Hills with no vegetation contributed to high surface runoff and peak flow into the Xinjiang River. The amount of precipitation that infiltrated in the form of groundwater was very low which suggests that less groundwater will be available to be used during dry season. As less water is available, the problem of water scarcity during the dry season still persists. When the entire watershed is altered to urban land-use (Scenario 3), the model predicts that the peak flow and infiltration rate remain the same and surface runoff increases by just 1%. The results are not very different from Scenario 1.

Many studies have indicated that soil compaction as a result of urban growth is more likely to influence flood responses than the presence of forests. On the small watershed scale the effects of human interventions such as deforestation, conversion to agricultural land, and urban area can be directly documented in terms of higher peak flows. Forest have reduced surface runoff by allowing water to soak into the ground, thus reducing flash floods and improving groundwater recharge in Scenario 2. Increased groundwater recharge increases water availability for the dry period of the year. However, the percentage of precipitation that becomes flow is reduced by 3% as compared to Scenario 1 (Table 3). The reduction in the flow is lower than expected, which might be due to steep slopes of the hills.

Infiltrated water on steep slopes moves rapidly through the soil due to gravity, still contributing to immediate flow and floods. Thus reforestation on steep slopes does not contribute to reduction in peak flow or floods. The model prediction suggests that a 1-km riparian buffer of mixed broad leaf forest (Scenario 5) has higher hydrological impacts by lowering the peak flow by 6% as compared to Scenario 1 (Table 3), thus reducing immediate downstream flooding. Forests on gentle slopes and floodplains are effective at reducing surface runoff contribution to the Xinjiang River and reducing immediate downstream flooding. Likewise, building ponds in sub-basins (Scenario 6) that contribute higher surface runoff proves very effective at holding water during the flooding season. The water holding capacity of the ponds reduces the peak flow by 8% as compared to Scenario 1 (Table 3). When the water is held back it percolates, thus reducing immediate floods downstream. The ponds provide water storage for irrigation and domestic uses during the dry season. The presence of forest in floodplains, gentle slopes, and building ponds in areas with high slope help to regulate immediate flood downstream.


My analysis shows the hydrological effects of land-use in 2000 and five predictive land-use scenarios on water availability. The model’s findings suggest that the presence of forests and vegetation, especially on riparian floodplain and gentle slopes, reduces surface runoff by enhancing infiltration by increasing soil water content and recharging groundwater. This increases the water availability in the form of groundwater for use during the dry season. The four scenarios (2, 4, 5, and 6) showing changes in land-use within the watershed have a range of potential benefits for improving water availability and reducing the risk of floods downstream. Scenario 5 yields the most significant hydrological importance in terms of improving water availability by reducing surface runoff and enhancing infiltration. Scenario 6 yields the most significant hydrological improvement in terms of regulating flow and reducing immediate flood downstream. Thus the optimal water management scenarios that yield the most significant hydrological improvement are restoration and forestation of riparian buffers, and installation of earthen ponds on steep sub-basins. Restoration of riparian buffers at the expense of cities along riparian areas is impossible, so installation of earthen ponds is more realistic. They will be less disruptive to cities and could maintain the agricultural base within the watershed.


I gratefully acknowledge funding received from the Tropical Resources Institute (TRI), Carpenter-Sperry Research Fund and Mr. Ambrecht. I am also grateful to the students of at Nanchang University, and Juijiang Hydrological Center for field assistance. Thanks also to Jing Ma for accompanying me to the field during summer of 2012. I am grateful to Prof. Chad Oliver, James Saiers, and Dr. Moe Myint for their advice and support in taking this study in the right direction.


Abbaspour, Karim. 2011. SWAT Calibration and Uncertainly Programs – A User Manual. Eawag Swiss Federal Institute of Aquatic Science and Technology.

Calder, Ian. 2000. Land-use impacts on water resources. Keynote paper presented at FAO Electronic-workshop on Land-Water Linkages in Rural Watersheds, 18 September to 27 October 2000.

Cheng-Fan, Xi. 1990. Better land-use and reclamation of red soil hilly region of southern China. Geo Journal 20, 365-368.

Guo, Hua, Qi Hu, and Tong Jiang. 2008. Annual and seasonal streamflow responses to climate and land-cover changes in the Poyang Lake basin. Journal of Hydrology 355, 106-122.

Khadka, A., C. Fu, M. Myint, C. Oliver and J. Saiers. 2013. Effect of land-cover changes and other remediations on hydrology of Xinjiang River sub-watershed. Journal of Environmental Science and Engineering B 2, 416-425.

Neitsch, S.L., J.G. Arnold, J.R. Kiriry, and J.R. Williams. 2009. Soil and Water Assessment Tool Theoretical Documentation Version 2009. Grassland, Soil and Water Research Laboratory, Agricultural Research Service, Texas. August 2009.

Sahin V. and M.J. Hall. 1996. The effects of afforestation and deforestation on water yields. Journal of Hydrology 178, 293-309.

Wei, X., G. Sun, S. Jiang, H. Zhou, L. Dai. 2008. Forest-streamflow relationships in China: A 40-year retrospect. Journal of the American Water Resources Association 44, 1076-1085.

Zheng, H., F. Chen, Z. Ouyang, N. Tu, W. Xu, X. Wang. 2008. Impacts of reforestation approaches on runoff control in the hilly red soil region of Southern China. Journal of Hydrology 365, 174-184.


Khadka, A. 2014. Predicting the Effects of Different Land-Use Scenarios on Water Availability Using a Hydrological Model. Tropical Resources Bulletin 32-33, 72-77.

  1. Ambika Khadka is a Research Officer at International Water Management Institute, Nepal. She received her MSc in Environmental Science from Yale School of Forestry and Environmental Studies. During her Master’s study she published a paper titled “Effects of Landcover changes and other Remediations on hydrology of Xinjiang River sub-watershed.” She was a Fulbright fellow to Yale and a recipient of the Andrew Sabin Environmental fellowship. She has more than three years of experience as researcher with WWF- Nepal, IUCN-Nepal and Hydro-Solution Private Limited. She can be contacted at //= 0; i=i-1){ if (l[i].substring(0, 1) == ' ') output += “&#”+unescape(l[i].substring(1))+”;”; else output += unescape(l[i]); } document.getElementById('eeEncEmail_ymutDYSsvZ').innerHTML = output; //]]> , //= 0; i=i-1){ if (l[i].substring(0, 1) == ' ') output += “&#”+unescape(l[i].substring(1))+”;”; else output += unescape(l[i]); } document.getElementById('eeEncEmail_CTjNrS9eHs').innerHTML = output; //]]> .

  2. Digital Elevation Model (DEM) is the data files that contain elevation of the terrain over a specified area, usually at a fixed grid interval over the “Bare earth”. The interval between each of the grid points will always be referenced to some geographical coordinate system either latitude-longitude or UTM coordinate system (