59 6 Studyonthe Intra-Annual Distribution Characteristics of the WaterBudgetinthe HillyRegionofRedSoil in Northeast Jiangxi Province, China Junfeng Dai, Jiazhou Chen, Yuanlai Cui, and Yuanqiu He 6.1 INTRODUCTION In northeast Jiangxi Province, water balance is important for water resource utiliza- tion and agricultural regionalization. In this region, the red soil is affected by the subtropical monsoon climate in which rainfall is abundant, but unevenly distributed (Chen and Zhang 2002). The high intensity of rainfall leads to water loss during the rainy season. In addition, the time difference between the distribution of the rain- fall and its evaporation causes seasonal drought. Studying the characteristics of the intra-annual distribution of the water budget can help discover the occurrence rules of seasonal drought in the region and assist in adopting measures to increase water use efciency (Huang et al. 2004). Existing literature has examined aspects of the water problem in red soil areas. Yao (1996) described the water problem including seasonal drought and water stor- age capacity in a red soil area. Wang et al. (1996) analyzed the variations in rainfall, evaporation, and water storage and supply in low, hilly red soil areas. Xie et al. (2000) studied the status of water resources in 1,000-m 2 area using the runoff plot, neutron probe, and tensiometer. However, the water budget in the hilly region of red soil has rarely been investigated systematically. It is difcult to obtain or measure percolation, evapotranspiration, and runoff accurately in the large- to medium-scale basin, and even in the small-scale watershed. Therefore, water budget in a scale larger than a eld is not easily studied using traditional methods. Model simulation is a viable option for calculating the output of the water budget at a variety of scales. However, such studies © 2008 by Taylor & Francis Group, LLC 60 Wetland and Water Resource Modeling and Assessment have rarely been undertaken in the red soil areas of China. SWAT (Soil and Water Assessment Tool) (Arnold et al. 1998) is a distributed hydrological model based on physical mechanisms. Although SWAT is designed to predict the impact of manage- ment on water, sediment, and agricultural chemical yields in large ungauged basins, it is also valid in small catchments (Chanasyk et al. 2003, Mapfumo et al. 2004). In this study, characteristics of intra-annual distribution of water balance are investigated using runoff catchments and the SWAT99.2 model in order to provide appropriate directions for effective management and use of water resources in the hilly region of red soil. 6.2 BRIEF DESCRIPTION OF MODEL SWAT is a distributed and continuous time model developed by Dr. Jeff Arnold for the U.S. Department of Agriculture (USDA), Agricultural Research Service. It has undergone several versions with different interfaces. In this article, the results of water budget are calculated using the Windows interface for SWAT99.2. The SWAT model consists of three major components: sub-basin, reservoir routing, and channel routing. The sub-basin component consists of eight main sub- components dened as hydrology, weather, sedimentation, soil temperature, crop growth, nutrients, agricultural management, and pesticides. The following is a brief description of the main hydrology subcomponent. For a complete description, see Arnold et al. (1999). SWAT simulates surface runoff volumes using daily rainfall amounts. Surface runoff volume is computed using a modication of the SCS (Soil Conservation Ser- vice) curve number method (USDA Soil Conservation Service 1972), and canopy storage is taken into account in the surface runoff calculations. The percolation com- ponent of SWAT uses a storage routing technique to predict ow through each soil layer in the root zone. Percolation occurs when the eld capacity of a soil layer is exceeded and the layer below is not saturated. The ow rate is governed by the satu- rated conductivity of the soil layer. The model computes evaporation from soils and plants separately as described by Ritchie (1972). Potential soil water evaporation is estimated as a function of potential evapotranspiration and leaf area index. Actual soil water evaporation is estimated by using exponential functions of soil depth and water content. Plant transpiration is simulated as a linear function of potential evapotranspiration and leaf area index. The model offers three options for estimat- ing potential evapotranspiration: Hargreaves, Priestley-Taylor, and Penman-Monte- ith. SWAT utilizes a simplication of the EPIC crop model (Williams et al. 1984) to simulate all types of land covers. The model is able to differentiate between annual and perennial plants. 6.3 SITE SELECTION AND MODEL CALIBRATION 6.3.1 A PPLICATION SITE The study area is located in the articial forest and grass catchments of the Experi- ment Station of Red Earth Ecology, Chinese Academy of Science, Yingtan, Jiangxi Province (Figure 6.1). © 2008 by Taylor & Francis Group, LLC Study on the Intra-Annual Distribution Characteristics of the Water Budget 61 The catchments are at an altitude of 48 to 54 meters and the groundwater table is low. The catchments were established in 1988 and were isolated by concrete 5 cm wide and 1 m deep. The four catchments are relatively dependent. Consequently, each catchment is considered a watershed in the model and is not partitioned into sub-basins. Water budget in each catchment is calculated respectively. An observation house with a runoff pool was constructed at the exit of every catchment, and a water level gauge was installed in the runoff pool. Surface runoff of catchments could be calculated through the continuous recording of the water level change. The details of the four catchments are as follows: (1) Natural grass catchment: the vegetation recovered naturally after the original Pinus massoniana was cut; (2) Evergreen broad-leaved forest catchment: Schima Superba and C. ssa were planted after the Pinus massoniana was cut; (3) Mixed forest catchment: Q. chenil and Les- pedeza bicolor were planted in the original Pinus massonianas; (4) Coniferous for- est catchment: the original Pinus massoniana forestland remains. 6.3.2 DATA PREPARATION The model inputs require climate, soil, land use, and vegetation data. There is a mete- orological observation station near the forest and grass catchments. Daily precipita- tion and maximum and minimum air temperature observations from 2000 to 2001 are available, and they are input directly into the SWAT model. In the SWAT99.2 model with Windows interface, daily solar radiation, wind speed, and relative humid- ity are generated from average monthly values based on the historical statistical data. These statistics are generated using a weather generator module in SWAT. Thirteen years of observational data have been used to generate these statistics. The soil type of the catchments is red soil derived from the quaternary red clay. The properties of each soil layer are given in Table 6.1. Experiment Station of Red Earth Ecology, Yiangtan, Jiangxi, China N Forest and Grass Catchments FIGURE 6.1 Schematic map of the study area in China. © 2008 by Taylor & Francis Group, LLC 62 Wetland and Water Resource Modeling and Assessment Watershed attributes and vegetation characteristics of catchments are gained with investigation (Table 6.2). For the forest and grass catchments, a wet and a dry year (2000 and 2001) of daily runoff are obtained from the Experiment Station of Red Earth Ecology to calibrate and validate the model. 6.3.3 MODEL CALIBRATION AND VALIDATION The SWAT model is calibrated and validated against the observed runoff. In this study, the SCS curve number (CN 2 ) and soil evaporation compensation factor (ESCO) are adjusted until surface runoff is acceptable. CN 2 is a function of the soil’s perme- ability, land use, and antecedent soil water conditions. The parameter calibration of the model is displayed in Table 6.3. The correlation coefcient (R 2 ) and Nash-Sutcliffe coefcient (E ns ) (Saleh et al. 2000) are used to evaluate the variance between observed and simulated values. Comparing simulated daily and monthly surface runoff to observations, model ef- ciency is achieved (Table 6.4). It shows that the simulation model can be used as an analytical tool for calculating the hydrological cycle in a hilly region of red soil. TABLE 6.1 Soil properties of forest and grass catchments. Soil depth (mm) Parameter G Fb Fm Fc 200.00 Bulk density(g/cm 3 ) 1.45 1.29 1.31 1.42 Organic carbon content (%) 0.55 0. 60 0.58 0.59 Saturated conductivity (mm/h) 21.10 35.13 32.12 27.92 Available water capacity (mm/mm) 0.09 0.10 0.09 0.09 Clay content (%) 39.95 41.20 40.20 39.05 Silt content (%) 36.15 39.70 36.45 32.15 Sand content (%) 23.9 19.10 23.35 28.80 600.00 Bulk density (g/cm 3 ) 1.27 1.28 1.26 1.24 Organic carbon content (%) 0.30 0. 39 0.38 0.37 Saturated conductivity (mm/h) 4.48 8.78 6.78 3.80 Available water capacity (mm/mm) 0.10 0.08 0.09 0.09 Clay content (%) 48.45 50.95 49.75 50.05 Silt content (%) 37.55 37.55 37.30 33.85 Sand content (%) 14.00 11.50 12.95 15.80 2000.00 Bulk density (g/cm 3 ) 1.25 1.27 1.25 1.23 Organic carbon content (%) 0.20 0. 23 0.23 0.23 Saturated conductivity (mm/h) 1.24 2.76 1.72 1.08 Available water capacity (mm/mm) 0.08 0.08 0.07 0.08 Clay content (%) 52.65 52.05 52.10 53.10 Silt content (%) 33.70 37.00 36.20 34.00 Sand content (%) 13.65 10.95 11.70 12.90 Note: G: natural grass; Fb: broad-leaved forest; Fm: mixed forest; Fc: coniferous forest. © 2008 by Taylor & Francis Group, LLC Study on the Intra-Annual Distribution Characteristics of the Water Budget 63 6.4 RESULTS AND DISCUSSION 6.4.1 P RECIPITATION The daily rainfall record over a period of two years (Figure 6.2) indicates that the rainfall in the red soil area is uneven in its intra-annual distribution. The annual rainfall is heavy in the red soil area with 1,912.1 mm falling in 2000 and 1,482.4 mm for 2001. The daily rainfall record also shows that the heaviest rains fall from April to July. The April to June 2000 rainfall amounts account for about TABLE 6.2 Watershed attributes of forest and grass catchments. Parameter G Fb Fm Fc Area (m 2 ) 2672.2 3162.0 2483.6 3134.9 Slope length (m) 96.0 90.0 98.1 93.0 Slope steepness (m/m) 0.0573 0.0556 0.0535 0.0484 Surface litter (kg/hm 2 ) 172.1 5360.0 3294.4 1492.5 Land cover Perennial Trees Trees Trees Ground cover >75% >75% >75% >75% TABLE 6.3 Parameter calibration of model. G Fb Fm Fc Parameter Original Calibrated Original Calibrated Original Calibrated Original Calibrated CN 2 71 68 65 63 65 64 65 65 ESCO 0.95 0.80 0.95 0.70 0.95 0.75 0.95 0.75 TABLE 6.4 Evaluation of simulation efficiency. Year G Fb Fm Fc Daily Monthly Daily Monthly Daily Monthly Daily Monthly 2000 R 2 0.91 0.91 0.92 0.93 0.89 0.90 0.89 0.92 E ns 0.63 0.74 0.72 0.76 0.60 0.73 0.61 0.72 2001 R 2 0.86 0.90 0.73 0.85 0.78 0.93 0.80 0.89 E ns 0.74 0.73 0.66 0.75 0.65 0.74 0.67 0.73 © 2008 by Taylor & Francis Group, LLC 64 Wetland and Water Resource Modeling and Assessment 50.33% of annual total rainfall. High rainfall intensity is frequently observed dur- ing the rainy season (April to June) in the red soil areas. Over the two-year period, there were 10 days where the daily rainfall exceeded 50 mm, totaling 841.5 mm and representing 24.8% of the two-year total. In addition, 47.8% of the two-year rainfall occurs within a 38-day period, for a rain intensity of 42.7 mm d −1 . The rainfall pat- tern in the red soil area is hazardous to water resource utilization and crop growth. 6.4.2 SURFACE RUNOFF Daily observed and simulated surface runoffs are displayed in Figure 6.3. Both observed data and simulation results indicate that surface runoff scarcely takes place when the daily rainfall is less than 10 mm. The surface runoff can be observed when the daily rainfall is above 10–15 mm. The results show that temporal change of rainfall led to the variation of the surface runoff. Compared with the 2000 data, rainfall in 2001 decreased 22.74%, and the surface runoff of the four catchments was reduced between 62.11% and 74.04%. Furthermore, the surface runoff often occurs in the rainy season. The concentrated rainfall pattern induces heavy runoff. For instance, the rate of three months (April to June) surface runoff of G, Fb, Fm and Fc is 69.21%, 80.02%, 73.72%, and 69.41%, respectively, of the annual 2000 amount. Also, the surface runoff scarcely occurs in January, February, and December due to the shortage of rainfall. In addition, the average annual runoff coefcient of the forest and grass catch- ments is 0.14, which is different from the value of 0.1 in long-term cultivated red soils (Yang et al. 1993), and which also differs from the value ranging from 0.55 to 0.61 in the uncultivated or newly cultivated uplands of red soils (Ju and Wu 1995). These reveal that runoff in the red soil uplands changes remarkably with different types of utilization and vegetation. 6.4.3 EVAPOTRANSPIRATION Potential evapotranspiration is estimated using the Penman–Monteith equation in this article. Compared with the observed daily water surface evaporation, the calculated daily evapotranspiration in the forest and grass catchments is reasonable (Figure 6.4). Evapotranspiration mainly concentrates from April to August, especially between June and August, because of the ample soil water and high temperature. Between 0 50 100 150 200 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 11/1/00 3/1/011/1/01 5/1/01 7/1/01 9/1/01 11/1/01 Date Rainfall (mm) FIGURE 6.2 Daily rainfall in the study area. © 2008 by Taylor & Francis Group, LLC Study on the Intra-Annual Distribution Characteristics of the Water Budget 65 June and August 2000, the evapotranspiration of the catchments accounted for 45.69 to 50.87% of the annual total. Rainfall decreases sharply after August, but the actual evapotranspiration loss is still large. The time period of maximum evapotranspiration is not synchronous with that of maximum rainfall, which is the essential intra-annual distribution of the water budget in the red soil areas. FIGURE 6.3 Daily observed and simulated surface runoff for 2000 and 2001. © 2008 by Taylor & Francis Group, LLC 66 Wetland and Water Resource Modeling and Assessment In addition, the annual difference in evapotranspiration of forest and grass eco- systems is small. The results show that the evapotranspiration decreases less with rainfall than with surface runoff with rainfall. For instance, comparing 2001 data with that of 2000, rainfall decreases 22.47%, runoff of broad-leaved forest reduces 75.30% accordingly, but evapotranspiration only decreases 7.32%. 6.4.4 PERCOLATION FROM THE BOTTOM OF THE ROOT ZONE Figure 6.5 shows that percolation from the root zone of forestland and grassland often occurs in the rainy season. The excessive rainfall usually induces large daily percolation. The rate of four months’ (March to June) percolation is 67.00 to 79.35% of the annual rainfall over a period of two years. Percolation from the root zone scarcely occurs the rest of the year (July to October) due to insufcient rainfall and strong evapotranspiration. As a consequence of unequal rainfall, the rate of 2001 percolation of catchments is between 75.09 and 81.54% of the 2000 rate. In addition, the permeability of forest and grass catchments in red soil is better than that in bar- ren land in this area. This reveals that vegetation recovery in the red soil area can assist soil inltration and water resource utilization. 0 6 12 18 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 11/1/00 1/1/01 3/1/01 5/1/01 7/1/01 9/1/01 11/1/01 Date Evapotranspiration (mm) G Fb Fm Fc Water surface evaporation FIGURE 6.4 Simulated daily evapotranspirations in the forest and grass catchments. 0 6 12 18 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 11/1/00 1/1/01 3/1/01 5/1/01 7/1/01 9/1/01 11/1/01 Date Percolation (mm) G Fb Fm Fc FIGURE 6.5 Simulated daily percolations from the root zone in forest and grass catchments. © 2008 by Taylor & Francis Group, LLC Study on the Intra-Annual Distribution Characteristics of the Water Budget 67 6.4.5 WATER SURPLUS-DEFICIT STATUS In the study area, rainfall is the water input, and water outputs include surface runoff, percolation, and evapotranspiration. The water surplus-decit status in the catchments is shown in Figure 6.6. Water decits mainly occur from June to September, but in other months, water input is generally larger than water output. The results indicate that annual change of soil water storage accounts for about 1.0% of rainfall. From February to April, increasing rainfall and weak evaporation potential con- tribute to the increase in soil water storage, although a water decit can exist. From May to September, the soil water storage evinces a gradual decreasing trend. Espe- cially from July to September, the forestlands have a serious decit of soil water due to lack of rainfall and strong evaporation potential, and the depleted water is recharged by the soil water storage. Furthermore, in the autumn sowing period, a serious water shortage also occurs on farmland. This conict is detrimental to water utilization and crop growth. From October to the following February, the soil water storage gradually reaches the proper status. As a whole, the annual change of soil water storage is quite small. This indicates that the water budget in the hilly region of red soil has a dynamic balance. The results also show that the evapotranspiration in forest and grass catch- ments accounts for a majority of the water output in the red soil area. The pro- portion between annual evapotranspiration and rainfall is above 0.5 in forestlands, and approximately 0.4 in the grass catchment. Furthermore, in the forest and grass catchments where vegetation has grown for twelve years or more, percolation from the root zone is bigger than surface runoff, and the condition seems more remarkable in the wet year. 6.5 CONCLUSIONS The main characteristics of the water budget in a hilly region of red soil investigated with runoff catchments and model simulation revealed the following information. The temporal distribution of rainfall is uneven in red soil areas. High rain- fall intensity is frequently observed during the rainy season (April to June), which –25 0 25 50 75 100 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 11/1/00 1/1/01 3/1/01 5/1/01 7/1/01 9/1/01 11/1/01 Date Surplus-deficit Status (mm) G Fb Fm Fc FIGURE 6.6 Simulated daily water surplus-decit status of forest and grass catchments. © 2008 by Taylor & Francis Group, LLC 68 Wetland and Water Resource Modeling and Assessment induces heavy runoff. Maximum evapotranspiration occurs from July to September, which is not synchronous with maximum rainfall. This conict is detrimental to water utilization. In the forest and grasslands where vegetation has grown for twelve years or more, evapotranspiration is the largest water loss component, followed by percolation, and then surface runoff. This differs from the long-term cultivated lands, where percolation is the largest water loss component, and also from the uncultivated uplands, where runoff is the largest component. The variation in weather conditions, especially rainfall, is the most important factor that results in the difference of water balance in the red soil area. Of the water outputs, the surface runoff and percolation are most affected by rainfall, followed by evapotranspiration, and the annual uctuation of the soil water storage is very small. According to the intra-annual distribution characteristics of the water bud- get in the red soil, important measures such as storing rainfall through the rainy season, constructing the irrigation works, agroforestry system or silvopastoral system techniques, and agricultural water-saving techniques should be taken to increase the water use efciency and relieve the seasonal drought in the hilly red soil region. This study also conrms that the modeling tool is an effective method for inves- tigating the water balance of the hilly red soil region. The modeling tool is demon- strated in the forest and grass catchments in this research, but it could be adapted to the cropland catchments or large-scale watersheds in the red soil areas of China. The advantages of the modeling tool are not limited to its capability to simulate hydrological processes. Many management scenarios, such as fertilization, irriga- tion, water decits, and population control, could also be analyzed. ACKNOWLEDGMENTS This study was nancially supported by the National Natural Science Foundation of China (No. 40301019) and Chinese Ministry of Education Project (No. NCET-04-0664). REFERENCES Arnold, J. G., R. Srinivasan, R. S. Muttiah, and J. R. Williams. 1998. Large area hydrologic modeling and assessment. Part I, model development. Journal of the American Water Resource Association 34(1):73–89. Arnold, J. G., J. R. Williams, R. Srinivasan, and K. W. King. 1999. Soil and water assessment tool theoretical documentation (version 99.2). http://www.brc.tamus.edu/swat. Chanasyk, D. S., E. Mapfumo, and W. Willms. 2003. Quantication and simulation of sur- face runoff from fescue grassland watersheds. Agricultural Water Management, no. 59: 137–153. Chen, Z. F., and X. X. Zhang. 2002. Effect of seasonal drought on forestry and fruit indus- try at red soil region in southern China [in Chinese]. Agro-environmental Protection 21(3):241–244. Huang, D. Y., K. L. Wang, M. Huang, H. S. Chen, J. S. Wu, G. P. Zhang, and T. B. Peng. 2004. Seasonal drought problems in the hilly, red soil region of the middle subtropical zone of China [in Chinese]. Acta Ecologica Sinica 24(11): 2516–2523. © 2008 by Taylor & Francis Group, LLC [...]... M S McFarland 2000 Application of SWAT for the upper north Bosque River watershed Transactions of the ASAE 43(5):1077–1087 U.S Department of Agriculture, Soil Conservation Service 1972 National Engineering Handbook Section 4, Hydrology Washington, DC: U.S Department of Agriculture, chaps 4–10 Wang, M Z., J B Zhang, and C S Zhao 19 96 Study on temporal and spatial variations of water resources and its... Intra-Annual Distribution Characteristics of the Water Budget 69 Ju, Z H., and S Z Wu 1995 The soil water balance conditions and improvement ways in central Jiangxi Province In Agricultural integrated development and countermeasure in red and yellow soil area in China [in Chinese], ed Y S Yang and L Q Xin Beijing: China Agricultural Science and Technology Publishing House, 143–147 Mapfumo, E., D S Chanasyk,... utilization in low hilly red soil area [in Chinese] Acta Agriculturae Jiangxi 8(1):47–58 Williams, J R., C A Jones, and P T Dyke 1984 A modeling approach to determining the relationship between erosion and soil productivity Transactions of ASAE 27:129–144 Xie, X L., K R Wang, and W J Zhou 2000 Status and management of water resources in red soil regions and hilly-sloping lands [in Chinese] Journal of... Chanasyk, and W D Willms 2004 Simulating daily soil water under foothills fescue grazing with the soil and water assessment tool model (Alberta, Canada) Hydrological Processes 8: 2787–2800 Ritchie, J T 1972 A model for predicting evaporation from a row crop with incomplete cover Water Resource Research 8:1204–1213 Saleh, A. , J G Arnold, P W Gassman, L M Hauck, W D Rosenthal, J R Williams, and A M S... Mountain Science 18(4):3 36 340 Yang, Y S., D M Shi, X X Lu, and Y Liang 1993 Analysis of sediment delivered from eroded sloping land in quaternary red region In Research on red soil ecosystem [in Chinese], ed M Z Wang, T L Zhang, and Y Q He Nanchang: Jiangxi Science and Technology Publishing House, 330–344 Yao, X L 19 96 Water problem of red soil and its management [in Chinese] Acta Pedologica Sinica 33(1):... and Y Q He Nanchang: Jiangxi Science and Technology Publishing House, 330–344 Yao, X L 19 96 Water problem of red soil and its management [in Chinese] Acta Pedologica Sinica 33(1): 13–20 © 2008 by Taylor & Francis Group, LLC . described the water problem including seasonal drought and water stor- age capacity in a red soil area. Wang et al. (19 96) analyzed the variations in rainfall, evaporation, and water storage and supply. Fc FIGURE 6. 6 Simulated daily water surplus-decit status of forest and grass catchments. © 2008 by Taylor & Francis Group, LLC 68 Wetland and Water Resource Modeling and Assessment induces heavy. FIGURE 6. 3 Daily observed and simulated surface runoff for 2000 and 2001. © 2008 by Taylor & Francis Group, LLC 66 Wetland and Water Resource Modeling and Assessment In addition, the annual difference