The effect of reforestation on stream flow in Upper Nan river basin using Soil and Water Assessment Tool (SWAT) model

The effect of reforestation on stream flow in Upper Nan river basin using Soil and Water Assessment Tool( SWAT) model Winai Wangpimool1 ,Kobkiat Pongput2 ,Chinnapat Sukvibool3 , Samran Sombatpanit4 ,and Philip W. Gassman5 Abstract Forests are an important natural resource,vital to all life. Forests are an important source of food,cloth鄄 ing,and medicines,as well as a place to live. Water released from forests drains into the soil causing groundwa鄄 ter to emerge as stream flow throughout the year. In Thailand,most forests have been encroached by logging, paper production,and housing construction. Population growth and the need for farming area for crop and live鄄 stock production have also caused forest encroachment. Technical tools are needed to support the decision mak鄄 ers and planners if they are to achieve objectives of water conservation,and development. These technical tools are needed for assistance in the engineering,socio鄄economic,and environmental planning. The Soil and Water Assessment Tool( SWAT) was used in the hydrological modeling in this study of the complex and dynamic problems of The Upper Nan river basin. This was a case study to evaluate the impact of changing conditions in the river basin affected by the stream flow due to reforestation. The watershed area was delineated into 5 sub鄄watersheds based on surface topography provided by the Digital Elevation Model( DEM) and the parameters of each of these watersheds were calculated. The land use data was processed and reclassi鄄 fied to match with the SWAT model land use code. Ten different categories of land use in the study area were used for SWAT processing. Types of land use consist of: mixed forest(33郾 7% ) ,disturbed forest(30郾 2% ) , evergreen forest(17郾 7% ) ,paddy field(7郾 1% ) ,orchard(3郾 7% ) ,range brush(2郾 7% ) ,field crop(1郾 7% ) , planted forest(1郾 7% ) ,urban(1郾 4% ) and water resources(0郾 4% ) . Twenty鄄two types of soil were found in the basin. The initial curve number values were assigned based on the land use type and soil hydrologic group for the average antecedent moisture condition of the curve number method. The potential evapotranspiration was computed using the Penman鄄Monteith method. The simulation was performed using three reforestation scenarios to assess stream flow:(1) improved dis鄄 turbed forest,(2) field crops and range grass,and(3) both disturbed forest and field crops. The results of refor鄄 estation from scenarios 1 and 3 can increase stream flow in the drought season and can also reduce the flow in the wet season in the main stream and its tributaries. For scenario 2,reforestation had no significant effect on the main stream. Key Words:Reforestation,Stream flow,SWAT model,Upper Nan river basin,Thailand

1摇 Introduction

Thailand蒺s area is divided into 25 major river basins or 254 sub鄄basins. The Nan river basin is one of the ma鄄

摇

1

摇

2

摇

3

摇

摇

4 5

Ph郾 D. Candidate, Department of Water Resources Engineering, Faculty of Engineering, Kasetsart University, Bangkok 10900, Thailand. E鄄mail:

winai郾 wangpimool@ gmail郾 com

Assoc. Prof郾 ,Department of Water Resources Engineering,Faculty of Engineering,Kasetsart University,Bangkok 10900,Thailand. Corresponding au鄄 thor: E鄄mail: kobkiat郾 p@ ku郾 ac郾 th

Land Development Department,Bangkok,Thailand

Past President and Membership Coordinator,World Association of Soil and Water Conservation( WASWAC) ,Bangkok,Thailand and Beijing,China Dr郾 ,Center for Agricultural and Rural Development,Iowa State Univ郾 ,Ames,Iowa,USA

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

53

jor river basins that have high encroachment of forest areas which has impacted watersheds of the river basin,and causes flooding almost every year,with frequent social and economic damages( Department of National Parks,Wild鄄 life and Plant Conservation,2007) . Land use change due to increased cultivated land has severely impacted the watershed area and decreased soil quality( Land Development Department,1987) . Moreover,both agricultural and ecological system reserves in the lower reaches of the Nan river basin have water shortage problems in the dry sea鄄 son( December鄄April) ( Department of Water Resources,2003,2006;Office of Natural Resources and Environmen鄄 tal Policy and Planning,2006) . The loss of forest area increases the flood potential and also increases drought im鄄 pacts. On the other hand,increasing forest area after returning agricultural lands to forest reduces the wet season stream flow and raises it in dry seasons,thus reducing flood potential in the wet season and drought severity in the dry season( Guo et al郾 ,2008) . A great number of studies have reported the impacts of land use on runoff( Costa et al郾 ,2003;Chen et al郾 ,2007;Xiaobo et al郾 ,2008;Ouyang et al郾 ,2008;Cao et al郾 ,2009;Mueller et al郾 ,2009; Mao and Cherkauer,2009;Mohammad and Adam,2010;Schilling et al郾 ,2010) . Ouyang et al. (2008) has indica鄄 ted that the highest peak runoff value is hillside cropland with slope larger than 15毅 which had been converted from forest,and the lowest is hillside cropland with a slope larger than 25毅 converted into forest when compared with the cropland and forest area conditions before the implementation of converting cropland to forest policy. Cao et al. (2009) has shown that the potential maximum plantation pine cover was expected to decrease moderately an鄄 nual total water yields,quick flow and base flow when compared with the current actual land use. In China蒺s loess plateau,pine woodland induced the largest water loss to surface runoff,followed by sloping cropland,alfalfa,semi鄄 natural grassland and shrub land( Chen et al郾 ,2007) . The Soil and Water Assessment Tool( SWAT) is a river basin scale model developed by Dr. Jeff Arnold for the USDA Agricultural Research Service( ARS) to predict the impact of land management practices on water,sediment and agricultural chemical yields in large and complex watersheds with varying soils,land use and management con鄄 ditions over long periods of time. SWAT is a public domain model,and is actively supported by the Grassland,Soil and Water Research Laboratory,and Black Land Research Center in Texas( Neitsch et al郾 ,2005) . SWAT is a spa鄄 tially distributed,physically based hydrological model,which can operate on a daily or monthly time step as well as annual steps for long鄄term simulations. SWAT uses spatially distributed data layers for elevation,land cover and soil types. Relational databases include soil attributes,weather and crop management data( Bingner,1996;Brown et al郾 ,1996; Arnold et al郾 ,1998; Zhang et al郾 ,2003; National Research Council,2004; Bouraoui et al郾 ,2005; Neitsch et al郾 ,2005;Easton et al郾 ,2008) . The fundamental concept of SWAT is the Hydrologic Response Unit ( HRU) . The HRUs are lumped land areas within the sub鄄basin that are of unique land cover,and soil and man鄄 agement combinations( Neitsch et al郾 ,2005) . The AVSWATX鄄2005 as a version of the SWAT software was de鄄 signed for GIS software ArcView鄄3郾 X;both of these were used in this study. The Upper Nan river basin has been taken as a case study because the deforestation situation in this area has caused enormous changes, and vastly impacted the watershed area. In this study, the effects of reforestation on stream flow are investigated. The land use in the year 2002 is used as a baseline year. Annual stream flows in the scenarios are also compared with the baseline year.

2摇 Methodologies

2郾 1摇 Description of study area

Nan province is located in the Northern region of Thailand. Most of the area in Nan is mountainous especially along the borders of the province,the flat area is located in the middle of the province. The Nan river originates in the north of the province and flows Southward to the Sirikit Dam in Uttaradit province,to join with other rivers to form Chao Phraya river. The Upper Nan river basin is in Nan province,90 km long,with an area of 5郯 663 km2 . The basin consists of five sub鄄basins namely鄄 Upper Nan,Nam Yao( W) ,Nam Yao( E) ,Nam Samun and Nam Nan part鄄2. Annual precipitation is 1,382 mm,and average temperature is about 25郾 6 degree Celsius. The locations of hydro鄄meteorological stations in the Upper Nan river basin are shown in Fig. 1( DDPM,2010) . Eighty鄄five percent of the land area in Nan Province is forested mountains and highlands. These abundant for鄄 est and catchment areas are the sources of numerous rivers, including the Nan River, which rises in Phu Wae Mountain in the north of the province. The sparse population of 477,000 consists of indigenous Lanna natives as well as other minority groups,most of them being hill tribes such as Hmong,Wa,Kamu,Luo,Tin and Mrabri,mak鄄 ing up 17 percent of the total population. Each group has different modes of production and a solid inheritance of culture and traditions. 54

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

Fig. 1摇 Location of Upper Nan basin,Thailand and its hydro鄄meteorological stations

The average income per person is the third lowest in Thailand,while the incidence of poverty(20 percent) is the third highest in the country and the second highest in northern Thailand after Mae Hong Son province( NES鄄 DB,2010) . One of the main underlying causes of poverty in Nan is the limited amount of arable land. Up to 87 percent of its area of 11郯 472 km2 is heavily forested with mountainous terrain,leaving only 12 percent for agricul鄄 ture and 1 percent for residential area.

2郾 2摇 Description of SWAT

The Soil and Water Assessment Tool ( SWAT) model was developed by the U郾 S. Department of Agriculture (USDA) Agriculture Research Service( ARS) and represents a continuation of about 40 years of modeling efforts ( Williams et al郾 ,2008) . SWAT is a public domain watershed scale model developed to predict the effects of land management on water,sediment,nutrients,pesticides,and agricultural chemicals in small to large complex basins ( Neitsch et al郾 ,2005) . It is a physically based,semi鄄distributed parameter model with a robust hydrologic and pollution element that has been successfully employed in a number of watersheds. Applications of SWAT have ex鄄 panded worldwide over the past decade,especially in the US and Europe( Gassman et al郾 ,2007) . The major components of SWAT are climate,hydrology,erosion,land cover and plant growth,nutrients,pesti鄄 cides,and land management( Neitsch et al郾 ,2005) . SWAT was used to simulate the hydrologic processes of the study watershed. Simulations of the hydrology of a watershed can be separated into two major divisions. The first di鄄 vision is the land phase of the hydrologic cycle,and the second division is the water or routing phase of the cycle. 摇 For the Land phase,the hydrologic cycle is based on the water balance equation: SW t = SW o +

t

移 ( R day i=1

- Q surf - E a - w seep - Q gw )

(1)

where SW t is the final soil water content( mm H2 O) ,SW o is the initial soil water content on day i( mm H2 O) ,t is the time( days) ,R day is the amount of precipitation on day i( mm H2 O) ,Q surf is the amount of surface runoff on day

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

55

i( mm H2 O) ,E a is the amount of evapotranspiration on day i( mm H2 O) ,w seep is the amount of water entering the vadose zone from the soil profile on day i( mm H2 O) ,and Q gw is the amount of return flow on day i( mm H2 O) . For water routing, SWAT uses Manning蒺s equation to define the rate and velocity of flow. Water is routed through the channel network using the variable storage routing method or the Muskingum river routing method. Manning蒺s equation for uniform flow in a channel is used to calculate the rate and velocity of flow in a reach seg鄄 ment for a given time step: A ·R2ch/ 3 ·slp1ch/ 2 q ch = ch (2) n R2 / 3 ·slp1ch/ 2 (3) v c = ch n where q ch is the rate of flow in the channel( m3 s -1 ) ,A ch is the cross section area of flow in the channel( m2 ) ,R ch is the hydraulic radius for a given depth of flow ( m) , slp ch is the slope along the channel length ( m m -1 ) , n is Manning蒺s coefficient for the channel,and v c is the flow velocity( m s -1 ) ( Neitsch et al郾 ,2005) . The conceptual process of the SWAT model is shown in Fig. 2.

Fig郾 2摇 Processing and display concept for SWAT model

2郾 3摇 Data collection

The SWAT data input files for modeling the effect of reforestation on stream flow includes a digital elevation model( DEM) and land use,soil,climate,and runoff data. 2郾 3郾 1摇 Digital elevation model( DEM) data SWAT requires extensive topographic data. Spatial topographic data required for the SWAT application were obtained from the Royal Thai Survey Department( RTSD) . DEM is a digital representation of the ground surface to鄄 pography or terrain. The DEM data is characterized by a 30 m 伊 30 m(1 颐 50郯 000 scale) resolution,with mini鄄 mum,maximum,and average elevations of 230 m,2郯 118 m,and 567 m,respectively,as shown in Fig. 1. 2郾 3郾 2摇 Land use data The land use data for year 2002 was provided by the Land Development Department( LDD) . The main land use was classified into 10 categories,which are forest evergreen(17郾 7% ) ,forest mix(33郾 7% ) ,disturbed forest (30郾 2% ) ,planted forest (1郾 65% ) ,range鄄brush (2郾 7% ) ,water resources(0郾 4% ) ,urban area (1郾 4% ) ,paddy field(7郾 1% ) ,field crop(1郾 7% ) and orchard(3郾 7% ) with total area about 5郯 663 km2 as shown in Table 1 and Fig. 3. 2郾 3郾 3摇 Soil data Soil data obtained from the Land Development Department consists of 22 groups. Three major soil groups are: (1) Slope complex;Sc,is the soil mixture in the steep areas with more than 30% hill slopes generally covered by forest,with slow permeability,and a high risk of erosion. (2) Pak Chong;Pc,is the soil from many decayed reflec鄄 tions of various materials with fine鄄grain,high permeability,steep areas have a tendency to have land slide prob鄄 lems,and(3) Chiang Khan;Ch,is a soil from river sediments. It sedimentary rock weathering;it has a high permea鄄 bility,and in steep areas,it easily collapses. The three major soil groups occupy areas of about 78郾 9% ,9郾 01% , 56

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

and 2郾 72% ,respectively of the Upper Nan river basin. All the soil series are shown in Table 2. Table 1

Land use types

Land use types in the Upper Nan river basin

Area

( km2 )

Disturbed forest( DTFR)

1,709郾 83

30郾 2

Orchard( ORCH)

207郾 53

3郾 7

151郾 29

2郾 7

Urban( URBN)

77郾 47

1郾 4

19郾 85

0郾 4

Paddy field( PDDY)

399郾 42

7郾 1

Forest evergreen( FRSE)

1,000郾 99

17郾 7

Field crop( FCRP)

94郾 50

% of Area

Land use types

Rang brush( RNGB)

Water Resources( WATR)

Area

% of Area

1郾 7

Forest mix( FRST)

Planted forest( PNFR) Total

( km2 )

1,908郾 50 93郾 63

5,663郾 00

33郾 7 1郾 7

100郾 0

Fig郾 3摇 Land use types for baseline in year 2002

2郾 3郾 4摇 Climate data Daily climate data were used,including precipitation,temperature,solar radiation,wind speed,and humidity data. The data were collected from distributed gauging stations within the Upper Nan river basin during the period from 1994 to 2008. The daily precipitation data were obtained at 11 gauging stations,while the daily temperature, solar radiation,wind speed and humidity data were collected from 1 major weather station of Thai Meteorological Department namely R28013 as shown in Fig. 1. 2郾 3郾 5摇 Runoff data The hydrologic runoff stations located in the basin areas that have complete daily runoff data were selected for model calibration and validation. Using four hydrologic runoff stations N郾 1,N郾 64,N郾 49 and N郾 65 from the Royal Irrigation Department,the two stations 090201 and 090203 from the Department of Water Resources were selected. International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

57

摇 摇 Table 2

Soil series in the Upper Nan river basin Area

Soil Series

( km2 )

Mae Sai( Ms) Tha Muang( Tm) Hang Dong( Hd) Nakhon Pathom( Np) Tha Yang( Ty) AC鄄pd Ban Mi( Bm) Chiang Khan( Ch) Pak Chong( Pc) Chiang Rai( Cr) Chatturat( Ct) Phen( Pn)

Area

% of Area

Soil Series

0郾 94 0郾 36 1郾 2 0郾 21 1郾 04 0郾 27 0郾 02 2郾 72 9郾 01 0郾 2 0郾 13 0郾 04

Phon Phisai( Pp)

53 20 68 12 59 15 1 154 510 11 7 2

% of Area

( km2 ) 9

0郾 16

Warin( Wn)

35

0郾 62

Roi Et( Re)

2

Wang Hai( Wi)

2

Kula Ronghai( Ki)

0郾 04

53

0郾 93 0郾 03

59

Kamphaeng Saen( Ks) Slope Complex( Sc)

4,468

Muak Lek( Ml)

98

Lat Ya( Ly)

1郾 04

78郾 9

15

Khao Yoi( Kyo)

0郾 27 1郾 74

8

Total

0郾 13

5郯 663

100

3摇 Application of SWAT 3郾 1摇 Watershed delineation

The delineations of the Upper Nan watershed for the SWAT simulations were performed using the ArcView SWAT( AVSWAT) interface developed by Di Luzio et al. (2002) . The study area was delineated into 28 sub鄄ba鄄 sins based on surface topography provided by the DEM,and the parameters of each of these were calculated using SWAT. The land use data was processed and reclassified to match the SWAT model with land use code. Ten differ鄄 ent categories of land use in the study area were used for SWAT processing. The type of land use consists of: mixed forest ( 33郾 7% ) , disturbed forest ( 30郾 2% ) , evergreen forest ( 17郾 7% ) , paddy field ( 7郾 1% ) , orchard (3郾 7% ) , range grass ( 2郾 7% ) , crop ( 1郾 7% ) , planting forest ( 1郾 7% ) , urban ( 1郾 4% ) and water resources (0郾 4% ) respectively. Twenty鄄two types of soil were found in the study area. The data were then converted and re鄄 classified to match the SWAT formats in order to support the model蒺s requirements. The initial curve number values were assigned based on the land use type and soil hydrologic group for the average antecedent moisture condition of the curve number method. The potential evapotranspiration was computed using the Penman鄄Monteith method.

3郾 2摇 Reforestation scenarios

The actual land use database,a digital map of land use in the year 2002 was defined as the baseline land use. Three scenarios were used to evaluate the effect of reforestation on stream flow in the Upper Nan river basin as shown in Table 3 and Fig. 4. The land use was changed for each scenario. The three scenarios were: Scenario鄄1 Forestry mix and Disturbed forest revival: All areas of forest mix and forest will be revived. Other land uses were the same as the 2002 pattern. Scenario鄄2 The Orchard and field crop will be reforestation and other land uses were the same as the 2002 pattern. Scenario鄄3 All of the Forestry mix,Disturbed forest,Orchard and field crop will be revived and other land u鄄 ses were the same as the 2002 pattern. Table 3

Items 1 2 3 4 5 6 7 8 9 10

58

Land use conditions in the three scenarios

Land use

categories Rang brush Water resources Urban Paddy field Field crop Forest evergreen Disturbed forest Forest mix Orchard Planted forest Total

Area

Baseline

( km2 )

151郾 3 19郾 9 77郾 5 399郾 4 94郾 5 1郯 001郾 0 1郯 709郾 8 1郯 908郾 5 207郾 5 93郾 6 5郯 663郾 0

Scenario鄄1

(% ) 2郾 7 0郾 4 1郾 4 7郾 1 1郾 7 17郾 7 30郾 2 33郾 7 3郾 7 1郾 7

100郾 0

Area

( km2 )

(% )

152郾 9 22郾 7 79郾 3 402郾 1 96郾 3 4郯 604郾 0

2郾 7 0郾 4 1郾 4 7郾 1 1郾 7 81郾 3

209郾 5 96郾 3

3郾 7 1郾 7

5郯 663郾 0

100郾 0

Scenario鄄2

Area

( km2 )

(% )

152郾 9 22郾 7 79郾 3 402郾 1

2郾 7 0郾 4 1郾 4 7郾 1

1郯 291郾 2 1郯 710郾 2 1郯 908郾 4

22郾 8 30郾 2 33郾 7

96郾 3

5,663郾 0

1郾 7

100郾 0

Scenario鄄3

Area

( km2 )

(% )

152郾 9 22郾 7 79郾 3 402郾 1

2郾 7 0郾 4 1郾 4 7郾 1

4郯 909郾 8

86郾 7

96郾 3

5郯 663郾 0

1郾 7

100郾 0

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

Fig郾 4摇 Reforestation scenario conditions in the Upper Nan river basin

3郾 3摇 Parameter sensitivity analysis

Sensitivity analysis was conducted to determine the influence of a set of parameters on predicting total flow to find the most sensitive parameters in the study watershed( Schmalz and Fohrer,2009) . The implementation of sen鄄 sitivity analysis procedures is helpful in calibration of hydrologic models and also for their transposition to different watersheds( Cibin et al郾 ,2010) . However,calibration and validation of the model is a key factor in reducing un鄄 certainty and increasing user confidence in its predictive abilities,which makes the application of the model effec鄄 tive. Information on sensitivity analysis,calibration,and validation of the multivariable SWAT model was provided to assist watershed modelers in developing their models to achieve watershed management goals ( White et al郾 , 2005) . In this study using SWATX 2005,sensitivity analyses were carried out through an automatic routine devel鄄 oped for the SWAT model( Van Griensven et al郾 ,2006) . Sensitivity analysis was performed on 27 different SWAT parameters for annual flow exports from the watershed. The top鄄ten range parameters resulting in the greatest values for each variable were shown in Table 4. Table 4

Top ten rank of parameters to sensitivity of model

Rank

Parameters

1

SURLAG

3

SOL_AWC

2

4

5

CN2

CH_K2 ESCO

6

ALPHA_BF

8

SOL_Z

7

9

10

SOL_K

CANMX SLOPE

Description Surface runoff lag coefficient

Initial SCS runoff curve number for moisture condition 域

Available water capacity of the soil layer( mm H2 O / mm soil)

Effective hydraulic conductivity in main channel alluvium( mm h -1 ) Soil evaporation compensation factor Base flow alpha factor( days)

Saturated hydraulic conductivity( mm h -1 )

Depth from soil surface to bottom of layer( mm) Maximum canopy storage( mm H2 O)

Slope of sub basin

摇 摇 Not all of the parameters identified by sensitivity analysis were modified during calibration,and some parame鄄 ters were modified during calibrations that were not identified during the sensitivity analysis. Parameters other than those identified during sensitivity analysis were used during calibration primarily due to the goal of matching the model as closely as possible to the naturally occurring processes in the watershed( White et al郾 ,2005) .

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

59

4摇 Results and discussions 4郾 1摇 Model calibration and validation

The calibration and validation of the model was a key factor in reducing the uncertainty and increasing user confidence in its predictive abilities,which makes the application an effective model. Information on the sensitivity analysis,calibration,and validation of multivariable SWAT models was provided to assist watershed modelers in de鄄 veloping their models to achieve their watershed management goals( White et al郾 ,2005) . The SWAT simulation was executed for the 1993 2008 period to provide a two鄄year initialization period. Cal鄄 ibration of the SWAT model was performed with 1999鄄2003 data while 2004 to 2008 were used for validation. Mod鄄 el calibration was done to compare the flow ratio obtained from simulations and observations at the selected hydro鄄 logical stations on a daily basis. In this study,6 points were selected for the model calibration,four stations from the Royal Irrigation Department( N郾 49,N郾 64,N郾 65 and N郾 1) and two stations from the Department of Water re鄄 sources(090201 and 090203) . The model calibration method considered every sub鄄basin located in the upper se鄄 lected point. Then,in order to obtain the closest simulated flow ratio to the observed value at the selected point,the trial and error method was applied to find the suitable value for the parameters in each sub鄄basin. The model out鄄 put variance was expressed as the variance in model performance measures such as NSE( Nash鄄Sutcliffe efficien鄄 cy) and Root Mean Squared Error( RMSE) ( Moriasi,2007) . The NSE and RMSE were used in this study;the cali鄄 bration results obtained from selected points over the watershed area. After the satisfied results from model calibra鄄 ted points and appropriate values for parameter set in each basin. The results from baseline will be comparing in study scenarios. The calibrated results of the main river and its tributaries are shown in Table 5 and some station in Fig. 5. Table 5

Calibration and validation results at gauge stations in Upper Nan river basin

Station name

Description

N郾 64 N郾 1 N郾 49 N郾 65 090201 090203

Nam Nan at Pa Khuang Nam Nan at Muang Nam Yao at Pua Nam Yao at Pang Sa Nam Pua at Na Fang Nam Korn at Pa Dang

Monthly

calibrated(1999 2003)

NSE

0郾 89 0郾 92 0郾 82 0郾 45 0郾 68 0郾 89

RMSE 0郾 90 0郾 93 0郾 86 0郾 47 0郾 71 0郾 90

Monthly

validated(2004 2008)

NSE

0郾 76 0郾 82 0郾 79 0郾 40 0郾 67 0郾 84

RMSE 0郾 80 0郾 85 0郾 81 0郾 44 0郾 69 0郾 87

Fig. 5摇 Model calibrations at N郾 1 station in period 1999 to 2003

4郾 2摇 Stream flow response to reforestation

The reforestation changes in the Upper Nan river basin in each scenario and baseline were used in the SWAT simulation for investigating the effect on stream flow. The result of simulation in the whole basin of annual stream flow of baseline land use and three reforest scenarios are shown in Table 6. 60

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

Table 6 Scenarios Baseline

Annual stream flow for baseline and reforest scenarios Wet Season

Annual Stream Flow

MCM

4,152

MCM

3,558郾 25

( May

Dry Season

Oct郾 ) % 86

MCM

591郾 96

14

3,435郾 98

83

680郾 87

Scenario 3

4,217

3,479郾 40

83

737郾 97

3,529郾 37

86

% 14

4,117 4,121

Apr郾 )

593郾 75

Scenario 1 Scenario 2

( Nov郾

17 17

Fig郾 6摇 Average monthly stream flows for baseline and three land use scenarios

摇 摇 The effects of reforestation on stream flow,especially seasonal stream flow,are more significant. The altered land use has an influence on seasonal stream flow volumes. The volumes of seasonal stream flow of scenarios 1鄄3 and baseline are shown in Table 6. The results indicate that the volumes of seasonal stream flow of three scenarios decreased in the wet season and increased in the dry season compared with the baseline land use( Fig. 7) . For the percentage of seasonal stream flow changes in scenarios 1 and 3,they decreased in the wet season at about 3% and

Fig郾 7摇 Compared seasonal stream flows between baseline and three land use scenarios International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

61

increased in the dry season also at about 3% . These stream flows are more important for water resource manage鄄 ment in the dry season and also for flood reduction in the wet season.

5摇 Conclusion

The study of the effect of reforestation on stream flow in the Upper Nan river basin by the Soil and Water As鄄 sessment Tool( SWAT) model application designated the land use in the year 2002 as the baseline land use,and assigned 3 reforestation land use scenarios. The estimated stream flow with the SWAT model showed annual aver鄄 age stream flow about 4,152 MCM. The average stream flow occurring during the rainy season( May October) was about 3,558 MCM(86% of the average annual stream flow) and in the dry season about 594 MCM(14% of the average annual stream flow) . The simulation was done by using three reforestation scenarios to assess stream flow: (1) improved disturbed forest,(2) field crop and range grass,and(3) both disturbed forest and field crop. The re鄄 sults of reforestation from scenarios 1 and 3 was predicted to increase annual stream flow in the drought season about 3% and to reduce the flow in the wet season about 3% in the main stream and its tributaries. For scenario鄄 2,reforestation was predicted to neither increase or decrease stream flow in either the wet or dry seasons. However, the effect of reforestation on water resources is both complex and uncertain;the consideration of many alternatives and evaluating the past will help in understanding the dynamic changes. This study result will be a guideline for decision鄄making about land use and water resources management in the Upper Nan and other river basins in the northern part of Thailand.

Acknowledgements

The authors would like to express their sincere thanks to the Department of Disaster Prevention and Mitigation for funding the developing a disaster management network for flash floods and landslides in Nan Province project in which this study was implemented. We would also like to thank the Thailand Research Fund through the Royal Golden Jubilee Ph郾 D. Program for their financial support.

References

Arnold,J郾 G郾 ,R. Srinivasan,R郾 S. Muttiah and J郾 R. Williams. 1998,Large area hydrologic modeling and assessment. Part I: Model de鄄 velopment. J. American water resources association,Vol. 34,pp. 73 89.

Bingner,R郾 L. 1996,Runoff simulated from Goodwin Creek watershed using SWAT. Transactions ASAE,Vol. 39,pp. 85 90.

Bouraoui,F郾 ,S. Benabdallah,A. Jrad and G. Bidoglio. 2005,Application of the SWAT model on the Medjerda river basin, Tunisia physics and chemistry of the earth,Vol. 30,pp. 497 507.

Brown,C郾 D. and J郾 M. Hollis. 1996,SWAT—A semi鄄empirical model to predict concentrations of pesticides entering surface waters from agricultural land. Pesticide Science,Vol. 47,pp. 41 50.

Cao,W郾 ,W郾 B. Bowden,T. Davie and A. Fenemor. 2009,Modeling impacts of land cover change on critical water resources in the Mo鄄 tueka river catchments,New Zealand. Water Resources Management,Vol. 23,pp. 137 151.

Chen,L郾 ,Z. Huang,J. Gong,B. Fu and Y. Huang. 2007,The effect of land cover / vegetation on soil water dynamic in the hilly area of the Loess Plateau,China. Catena,Vol. 70,pp. 200 208.

R. Cibin,K. P. Sudheer and I. Chuabey. 2010,Sensitivity and Identify ability of stream flow generation parameters of the SWAT mod鄄 el. Hydrological Processes,Vol. 24,pp. 1133 1148.

Costa,M郾 H郾 ,A郾 l. Botta and J郾 A. Cardille. 2003,Effects of large鄄scale changes in land cover on the discharge of the Tocantins River, Southeastern Amazonia. Journal of Hydrology,Vol. 283,pp. 206 217.

Department of National Parks,Wildlife and Plant Conservation. 2007,Investigated productivity and intrusion of forest resources to as鄄 sess the risk on forest,Bangkok.

Department of Water Resources. 2003,Integrated plans for water resources management in the Nan River Basin,Bangkok.

Department of Water Resources. 2006,Role in Thailand蒺s Water Management,Ministry of Natural resources and Environment,

Bangkok.

Department of Disaster Prevention and Mitigation( DDPM) . 2010,Final report of the Project for a development network of natural dis鄄 asters for flash floods and landslides. Institute of Science and Technology,Chiang Mai University.

Easton,Z郾 M郾 ,D郾 R. Fuka,M郾 T. Walter,D郾 M. Cowan,E郾 M. Schneiderman and T郾 S. Steenhuis. 2008,Re鄄conceptualizing the soil and water assessment tool( SWAT) model to predict runoff from variable source areas. Journal of hydrology,Vol. 348,pp. 279 291.

Gassman P郾 W郾 ,M. Reyes,C郾 H. Green and J郾 G. Arnold. 2007,The Soil and Water Assessment Tool: Historical development,applica鄄

62

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

tions and future directions. Trans ASABE,Vol. 50,No. 4,pp. 1211 1250.

Guo,H郾 ,Q. Hu and T. Jaing. 2008,Annual and seasonal streamflow responses to climate and land鄄cover changes in the Poyang Lake basin,China. Journal of Hydrology,Vol. 355,pp. 106 122.

Land Development Department. 1987,Land use plans in Northern part of Thailand. Ministry of Agriculture and Cooperatives,Bangkok.

Mao,D. and K郾 A. Cherkauer. 2009,Impacts of land鄄use change on hydrologic responses in the Great Lakes region. Journal of Hydrolo鄄 gy,Vol. 374,pp. 71 82.

Di Luzio,M郾 ,R. Srinivasan, J郾 G. Arnold, S郾 L. Neitsch. 2002, ArcView Interface for SWAT2000 User蒺s Guide, Blackland Research Center – Agricultural Research Service,Texas – USA.

Mohammad,A郾 G. and M郾 A. Adam. 2010,The impact of vegetative cover type on runoff and soil erosion under different land uses. Ca鄄 tena,Vol. 81,pp. 97 103.

Moriasi,D郾 N郾 ,J郾 G. Arnold,M郾 W. Van Liew,R郾 L. Binger,R郾 D. Harmel and T. Veith. 2007, Model evaluation Moriasi guidelines for systematic quantification of accuracy in watershed simulations. Trans ASABE,Vol. 50,No. 3,pp. 885 900.

Mueller,E郾 N郾 ,T. Francke,R郾 J. Batalla and A. Bronstert. 2009,Modeling the effects of land鄄use change on runoff and sediment yield for a meso鄄scale catchment in the Southern Pyrenees. Catena,Vol. 79,pp. 288 296.

National Research Council. 2004,Utilization of GAME鄄T data and developed models for integrated resources management of the Chao Phraya Basin,Bangkok.

Neitsch,S郾 L郾 ,J郾 G. Arnold,J郾 R. Kiniry and J郾 R. Williams. 2005,Soil and Water Assessment Tool,Theoretical Documentation,version 2005,Temple,Texas.

Office of the National Economic and Social Development Board( NESDB) . 2010,Northern Development News. Northern of the National Economic and Social Development. ( In Thai)

Office of Natural Resources and Environmental Policy and Planning. 2006,Situations of natural resources in watershed areas in 10 years. Ministry of Natural Resources and Environment,Bangkok.

Ouyang,W郾 ,F鄄H. Hao,X鄄L. Wang and H鄄G. Cheng. 2008,Nonpoint source pollution responses simulation for conversion cropland to forest in mountains by SWAT in China. Environmental Management,Vol. 41,pp. 79 89.

Schilling,K郾 E郾 ,K鄄S. Chan,H. Liu and Y鄄K. Zhang. 2010,Quantifying the effect of land use land cover change on increasing discharge in the Upper Mississippi River. J. Hydrology,Vol. 387,No. 3 4,pp. 343 345.

Van Griensven,A郾 ,T. Meixner,S. Grunwald,T. Bishop,M. Diluzio,and R. Srinivasan. 2006b,A global sensitivity analysis tool for the parameters of multi ‐ variable catchment models. J. Hydrol. Vol. 324,No. 1 4,pp. 10 23.

Williams J郾 R郾 ,J郾 G. Arnold, J郾 R. Kiniry, P郾 W. Gassman and C郾 H. Green. 2008, History of model development at Temple, Texas. Hydrol. Sci. J郾 ,Vol. 53,No. 5,pp. 948 960.

White,Kati L. and Indrajeet Chaubey. 2005,Sensitivity analysis,calibration,and validations for a multisite and multivariable SWAT model. J. American Water Resources Association( JAWRA) ,Vol. 41,No. 5,pp. 1077 1089.

Xiaobo,J郾 ,C鄄H. Huang and F. Ruan. 2008,Impacts of land cover changes on runoff and sediment in the Cedar Creek watershed,St. Joseph River,Indiana,United States. J. Mt. Sci郾 ,Vol. 5,pp. 113 121.

Zhang,X鄄S郾 ,F鄄H. Hao,H鄄G. Cheng and D鄄F. Li. 2003,Application of SWAT model in the upstream watershed of the Luohe river. Chinese Geogr. Science,Vol. 13,pp. 334 339.

International Soil and Water Conservation Research,Vol郾 1,No郾 2,2013,pp. 53 63

63