Available online at www.sciencedirect.com ScienceDirect

Procedia Engineering 154 ( 2016 ) 1064 – 1070

12th International Conference on Hydroinformatics, HIC 2016 Numerical Simulation of Land Surface Hydrological Process over Huaihe River Basin

CHENG Xing-wua, ZHAO Jina, ZHANG Honga, WANG Kaia*, XU Minb

aBureau of Hydrology of Huaihe River Commission, , Bengbu 233001, b Climate Center, ,Hefei 230031,China

Abstract

The BATS (Biosphere Atmosphere Transfer Scheme) model was used to simulate the thunderstorm and flood events in the mountains and plains of the Huaihe River Basin in 1998. And the inherent relationship among runoff, soil moisture content, soil texture and vegetation distribution were exposed via the simulation of water exchange between land and atmosphere. Results show that the effect of soil moisture content in root layer, soil texture, and soil color on the runoff in the mountains is similar to that in the plains, whereas with different sensitivities. However, the effect of soil moisture content in deep layer and vegetation canopy ratio on the runoff in the mountains is opposite to that in the plains. The meteorological and land surface conditions are different between the mountains and the plains, so their hydrological characteristics are different, showing the different atmospheric-hydrological relationship.

©© 2016 2016 The The Authors. Authors. Published Published by Elsevierby Elsevier Ltd. LtdThis. is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-reviewPeer-review under under responsibility responsibility of theof theorganizing organizing committee committee of HIC of 2016 HIC 2016.

Keywords:Huaihe River Basin; land surface hydrological process; BATS model; hydrological simulation

1. Introduction

In recent years, hydrologists and meteorologists pay more and more attention to the effects of hydrology in geosphere – biosphere circle. In the process of seeking the interaction and feedback mechanisms between land hydrological process and atmosphere, we must consider the interaction between land and atmosphere in the general

* Corresponding author. Tel.: +86-136-2552-7983; fax: +86-552-3093316. E-mail address:[email protected]

1877-7058 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of HIC 2016 doi: 10.1016/j.proeng.2016.07.597 Cheng Xing-wu et al. / Procedia Engineering 154 ( 2016 ) 1064 – 1070 1065 circulation model. In order to change the global climate model, we need to adjust meteorological model and study the changes of water resources through the authentic description of land surface hydrological process. A lot of land surface process model[1-2] have been developed in weather patterns, such as SiBUC which was used in the study of HUBEX-GEWEX by Fujiyoshi and Ding[3], DHSVM which was used in the simulation of Sea Luanhe River mountain area by Wang Shourong[4] and SHE, HBV, VIC which were mentioned by Chen Haishan and Sun Zhaobo[5], and they all obtained good results. Regional hydrological simulation research has been significantly improved with the development of land surface model, Ding Biao and Zeng Xinmin[6] has improved the surface water model according to the regional characteristics in China. In addition, a lot of research[7-8] aimed at short-term climate and land surface hydrology has also been carried out. In hydrometeorology coupling model, the description of hydrological cycle is the weakness of the current global climate model. The coupling model has been weighted heavily toward meteorological factors, whereas the description of the land surface process is too simple, especially for the local hydrological process [9-10]. From the most simple model—the bucket model to the multi-layer soil vegetation system at present, the effect of different terrain on the runoff yield and concentration has not been considered, real physical approach of water and heat quantity exchange between the atmosphere and soil has been accurately described though. The study which was carried out by Qi Dan[11] and Zhang Hongping[12] show that terrain and soil characteristics have an important influence on runoff. Therefore, atmosphere, land surface and surface water should be studied as a complete system[13] and the real description of the hydrological cycle process should be paid attention to, which can improve climate patterns change with time scale and its response to the water cycle on the one hand, on the other hand, the impact of climate on hydrology can be evaluated. The two difficult points in the hydro meteorological coupling model are: first, the fast changing process of the meteorological factors are not suited to the slow variation of the hydrological factors, second, the meteorological large scale uniform space field and the inhomogeneity of the hydrological space field is difficult to coordinate. In order to solve the coupling problem of the two, high-quality flux description between the land and atmosphere must be obtained. Therefore, considering the advantages of the BATS land surface model in the analysis of different terrain and land use, as well as the convenience of the usage of the available data, based on the Biosphere Atmosphere Transfer Scheme, this paper selected BATS surface process model for calculation. This paper selected two areas with typical hydrological characteristics, which are hilly area and plain area in Huaihe River Basin. Encryption data in 1998 of the global energy and water cycle test Huaihe River Basin experimental area (HUBEX) were used to simulate the storm flood process during flood season in 1998. The relation and interaction between atmosphere and land hydrology, as well as the intrinsic relation between runoff, soil moisture content, soil texture and vegetation distribution were revealed from the angle of land - atmosphere water exchange.

2. BATS land surface model

Land surface process model is composed of different physical processes, including the parameterization of the dynamic characteristics associated with vegetation shape, the transmission of long and short wave radiation in the canopy and rainfall interception, photosynthesis, transpiration associated with vegetable physiological process, physical process in soil with water and heat transfer, soil chemical process[13] and so on. The development and perfection of these physical processes have greatly enriched the contents of the land surface process model. The species diversity in ecological system and the difference of vegetation, topography lead to the complexity of climate ecology. More importantly, it is not possible to use consistent explicit physical equation constraints, increasing the difficulty of vegetation parameterization. In the development of the land surface process model, the BATS model has a certain representativeness, which uses the integrated model to simplify the land surface process model. The BATS model is very detailed in the description of the land, from bare soil to vegetation, from the snow cover to the canopy, from the land surface to the underground, using a large number of parameters (Fig.1). This model uses vegetation cover index to describe the vegetation, in which the process of rainfall interception, air hole block and the transmission of radiation use a whole set of parameters. Major formulas and parameters please refer to the references [14-16]. 1066 Cheng Xing-wu et al. / Procedia Engineering 154 ( 2016 ) 1064 – 1070

Fig.1 Structure of boundary layer and ground circumstance in the BATS model [9]

3. Model application

Huaihe River Basin is located in the eastern part of China, between the and Basin, with a total area of about 270,000km2, across the four provinces of , Anhui, Jiangsu and Shandong. The terrain of the Huaihe River Basin is generally inclining from the northwest to the southeast. Funiu Mountain and Tongbai Mountain area which are in the west of the basin, Dabie Mountain area in the south and Yimeng Mountain area in the northeast with elevation from 200 to 500 m form the mountain areas of the basin. Hilly regions are extension of the mountain areas mainly, mountain and hilly areas accounted for about one-third of the total area. The rest are vast plain areas, including the plain area in the north of Huaihe River, the western plain of Nansihu Lake and the river network plain of Lixiahe River, accounts for about two-thirds of the total area. Therefore, hilly areas and plain areas in the basin have formed different hydrological and meteorological characteristics [17]due to the different terrain conditions.

3.1. Study area

In this study, hilly areas and plain areas with representativeness in Huaihe River Basin were selected respectively: Sub Basin in the north of the and Sub Basin in the Huaibei Plain (Fig.2).

Fig.2 Shiguan River and Guo River in the Huaihe River basin of China Shiguan River is one of the main tributaries on the south bank of the Huaihe River, and the area of the Shiguan River Basin is 5930 km2withJiangjiaji as its outlet control station. The soils are mainly yellow brown soil and paddy soil in the Shiguan River Basin, in the south of this basin are high mountains, on both sides are hilly areas, and in the Cheng Xing-wu et al. / Procedia Engineering 154 ( 2016 ) 1064 – 1070 1067 north are vast plain areas. The vegetation in mountain areas is a mixed forest, and high stalk plants are the main vegetation along the river and in hilly or plain areas. The terrain is high in the south and low in the north of the baisn with large gradient ratio, so, once there is precipitation, it will produce large amount of runoff, and the runoff concentration of the river is fast. Therefore, this basin is an area with typical hydrological characteristics in mountain areas (Fig. 2). Guo River Basin, as another test area, is located in the central north of the Huaihe River Basin. The area of this basin is 10,575km2withBozhou as its outlet control station. This area is flat with yellow moist soil which has loose texture, and the main vegetation is dry crops (Fig.2). The terrain of the river has small gradient ratio, so, once there is precipitation,the runoff yield andconcentration is slow. Therefore, this basin is an area with typical hydrological characteristics in plain areas.

3.2. Data

Observed encryptedhydrological and meteorological data during flood season (June to August) in 1998of Shiguan River Basin and Guo River Basin were selected as model simulation data. Meteorological observation data near land surface in the two basins were selected as the forcing factors of the atmospheric boundary layer, and the time step of the simulation is 1 hour. Temperature values observed every 6 hour are interpolated to every 1 hour, and other meteorological data are updated every 6 hour. Empirical formula raised by Zillman[18] was used for radiation calculation. Observed encrypted data of every 1 hour were selected as rainfall data, and average of 15 stations were selected as basin surface average rainfall in Shiguan River basin and average of 19 stations were selected in Guo River basin. Aggregate method was adopted for runoff simulation, that is, the hilly and the plain areas were regarded as a grid of BATS respectively, the data of the meteorological stations in Gushi and represent the average conditions of the two sub basins respectively, discharge hydrograph and runoff observation data of the outlets of the two basins were regarded as true values for simulation verification. Observed soil temperature values in Gushi and Bozhou were regarded as reference values for earth temperature simulation verification.

3.3. Numerical simulation

Runoff is regarded as a parameter in the climate model in BATS, and soil freezing is one of the main factors for its change. The parameter is the surface runoff coefficient, which was taken as a fixed value, when the soil is frozen, n=1, otherwise, n=4[19].For the problem of rainfall and runoff when it is not frozen, the influence of topography and geomorphology is mainly reflected in the runoff coefficient[19].In order to consider the effect of the topography, n is regarded as a parametric variable of the topography influence. Runoff yield and concentration simulation test [20] in the basin indicates that when the soil vegetation parameters are determined, the gradient ratio of the topography is larger, the smaller the index n is, and vice versa. Runoff simulation can be improved by changing index n, for mountain areas, simulated runoff volume is 485mm when n=1.5, which is more close to the true value of 481mm; for plain areas, simulated runoff is more close to the true value when n=4, see table 1 in detail. Table1Simulatedrunoffvolumeovermountainandplainareasfordifferentindicesmm Simulatedrunoffvolume Observedrunoffvolume n=1.5 n=4.0 Mountainareas 142 236 158 Plainareas 26 18 15 In addition to topography factors, the initial value of soil water had a relatively large impact on runoff yield, and the average precipitation in June was 148 mm in Guo River basin, while the observed runoff of the outlet section was less than 2 mm. Apart from small precipitation intensity, uneven temporal and spatial distribution, small topography gradient ratio and the long-time runoff concentration which result in large loss of the evaporation quantity, small previous rainfall and large soil infiltration are also important factors. Comparison of observed and simulated runoff in flood seasons in mountain and plain areas was shown in Figure 3. 1068 Cheng Xing-wu et al. / Procedia Engineering 154 ( 2016 ) 1064 – 1070

Fig.3 Comparison between observed runoff and simulated one for different surfaces (a) mountain area˗ (b) plain area

4. Results analysis

It is a complex process from rainfall to runoff, rain drops from the sky to the ground and turns into net rain after evaporation, vegetation interception and infiltration. Runoff is divided into surface runoff and underground runoff.In general, surface runoff is much larger than underground runoff. When the net rain is allocated to surface runoff and underground runoff, there is also a part to supplement groundwater.

4.1. Effect of soil on runoff in mountain areas

The topography in Shiguan River basin is complex with large gradient ratio and high vegetation cover ratio, so, there are many factors which have an impact on runoff. Analysis results showed that soil moisture content in root layer, soil texture and vegetation cover ratio were proportional to runoff, while the soil color, soil moisture content in deep layer were in inverse proportion to runoff, but their sensitivity was not the same. (1) The contribution of soil moisture content in root layer to runoff is enormous, on the one hand, soil moisture is provided to vegetation evaporation through root system, and also provided to upper soil evaporation through the capillary; on the other hand, the root system is close to the upper soil, and has a great influence on the surface runoff. The total runoff of mountain areas in Shiguan River basin increased significantly along with the increase of soil moisture content in root layer. However, when the moisture content in root layer exceeds 500mm, change ofthe total runoff is not obvious, which is because the moisture content in this layer is too saturated, and the rich water has become a part of gravity drainage. Therefore, the sensitive area of the soil moisture content in root layer in Shiguan River basin was 100 ~ 500 mm (Fig.4). (2) The soil moisture content in deep layer is in inverse proportion to runoff with a relatively large effect. When the soil moisture content in deep layer increased from 500 mm to 2300 mm, the runoff of mountain areas decreased from 590 mm to 520 mm (Fig.5). (3)The influence of vegetation on runoff is very significant. When the vegetation cover was 0̚50%, its impact on runoff was very small; when the vegetation cover ratio increased to 60%, the runoff of mountain areas increased along with vegetation cover ratio. Therefore, vegetation cover ratio has significant impact on runoff of mountain areas in the range of 60%̚100%. Different vegetation types also have different influence on runoff. Generally speaking, runoff decreased with the increase of vegetation height, and when the land surface was frozen, the runoff reacheda maximum value of 600mm; the second is the short grass vegetation and its runoff was 564 mm; the runoff of the deciduous broad-leaved forest is the smallest, which was 425 mm.



Fig.4Relationshipbetweensoilmoistureofrootzoneandr Fig.5Relationshipbetweensoilmoistureofdeeplayerand unoffinthemountainarea runoffinthemountainarea Cheng Xing-wu et al. / Procedia Engineering 154 ( 2016 ) 1064 – 1070 1069

4.2. Effect of soil on runoff in plain areas

Guo River basin is located in Huaibei Plain, its annual rainfall and flood season rainfall are both one order of magnitude smaller than mountain areas in Huainan, so its runoff is also much smaller than mountain areas. (1) With the increase of soil moisture content in root layer and soil texture, the total runoff in Guo River plain area is also increasing. When the soil moisture content in root layer was 20 ~ 60 mm, the runoff increased with the increase of soil moisture content in root layer; when the soil moisture content in root layer was 70 ~ 140 mm, the runoff almost had no change; when the soil moisture content in root layer was 150 ~ 200 mm, the runoff increased with the increase of soil moisture content in the root layer again (Fig.6). (2) The effect of soil moisture content in deep layer on runoff in plain areas is opposite to mountain areas (Fig.7). When the soil moisture content in deep layer was 400̚1600 mm, the change of runoff in plain areas is little; when the soil moisture content in deep layer was 1700̚2000 mm, the runoff increased significantly. This has a great relationship with the geomorphology of Huaibei plain. Under normal circumstances, the rainfall is relatively small in Huaibei plain, and the soil is sandy soil with little soil moisture content, so the rainfall will supplement underground water and surface water at first, so the runoff is produced when it reaches a certain order of magnitude(Fig.7). In addition, the gradient ratio of the topography is small, so the lag of runoff yield is relatively large. (3) With the increase of vegetation cover in Guo River basin, the runoff of the basin is decreasing. Therefore, the larger the vegetation cover ratio in plain areas, the greater the interception of rainfall, and the greater the vegetation evapotranspiration, which can reduce the effective rainfall and soil moisture and lead to the decrease of runoff. Impact of different vegetation types on runoff in Guo River basin is similar to the mountain areas.

  Fig.6 Relationship between soil moisture of root zone  Fig.7 Relationship between soil moisture of deep layer and and the runoff in the plain area runoff in the plain area

5. Conclusions

Runoff is the product of water and heat exchange between land and atmosphere under the circumstance of different soil, vegetation and topography. Runoff simulation can reflect soil water simulation under different conditions. (1) It demonstrate that the influence of soil moisture content in root layer and soil texture change on runoff has similarity between mountain areas in Shiguan River basin and plain areas in Guo river basin from simulation of the BATS model with just a different sensitivity. However, the influence of soil moisture content in deep layer and vegetation cover ratio on runoff in mountain areas and plain areas is totally opposite. (2)Soil moisture content in root layer, soil texture and vegetation cover ratio are proportional to runoff in mountain areas, while soil color and soil moisture content in deep layer are in inverse proportion to runoff. Among them, moisture content in root layer, moisture content in deep layer and vegetation cover ratio have relatively great impact on runoff. However, runoff is only proportional to soil moisture content in root layer and soil texture in plain areas. The influence of these factors on runoff is not very significant compared to mountain areas. (3) This paper verified the difference of microclimate and underlying surface between mountain areas and plain areas both theoretically and practically, as a result, these factors must be taken into account for flood forecasting scheme design and parameter calibration. 1070 Cheng Xing-wu et al. / Procedia Engineering 154 ( 2016 ) 1064 – 1070

Acknowledgements

The study was financially supported by Non-profit Industry Financial Program of MWR of China (201301066ǃ 201401027ǃ201501007)

References

[1] Deardorff J W. Efficient prediction of ground surface temperature and moisture with inclusion of a layer of vegetation[J]. J Geophys Res,1978,83(C4):1889-1903. [2] Mabuchi K,Sato Y, Kida H,etal.A biosphere-atmosphere interaction model(BAM) and its primary verification using grassland data[J]. Papers in Meteorol and Geophys,1997,47(3/4):115-140. [3] Fujiyoshi Y, Ding Y H. Final report of GAME/HUBEX-GEWEX Asian monsoon experiment/Huaihe River Basin experiment- HUBEX[M].Word, 2006:48-60. [4] Wang Shourong, Huang Ronghui, Ding Yihui. Improvements of a distributed hydrology model DHSVM and its climatological-hydrological off-line simulation experiments[J]. ActaMeteorologica Sinica,2002,60(3):289-299. (In Chinese) [5] Chen Hai-shan, Sun Zhao-bo. Review of land-atmosphere interaction and land surface model studies[J]. Journal of Nanjing Institute of Meteorology, 2002, 25(2):277-288. (In Chinese) [6] Ding Biao, Zeng Xinmin. Numerical experiments of a regional climate model with its surface runoff scheme improved[J]. Scientia meteorologicasinica ,2006, 26(1):31-38. (In Chinese) [7] Li Ning, Zeng Xinmin, Xi Zhaoli. Dynamical-forecast study of monthly-scale climate for east China, partช: establishment and evaluation of the regional model[J]. Scientia meteorologicasinica , 2006ୈ26ୄ5୅ୖ591-597. (In Chinese) [8] Zeng Xinmin, Zhang Qiang. Numerical study of the effects of disturbances of land surface parameters on the simulation of a heavy rainfall over Northwest China[J]. Scientia meteorologicasinica , 2009ୈ29ୄ3୅ୖ291-298. (In Chinese) [9] Dichinson R E,Henderson-Sellers A,Kennedy P J,etal.Biosphere-atmosphere transfer scheme (BATS) version le as coupled to the NCAR community climate model[R]//NACR Technical Note,August 1991,NACR/TN-387+STR.1991. [10] Sellers P J,MintzY,Sud Y C,etal.A simple biosphere model (SB) for use within general circulation models[J].J Atmos Sci,1986,43(6)ୖ506- 531. [11] Qi Dan, Zhao Ping, Tu Qi-pu, et al. Numerical simulation of the Run-off Scheme in the RegCM2[J]. Journal of Nanjing Institute of Meteorology, 2006ୈ29ୄ6୅ୖ782-789. (In Chinese) [12] Zhang Hong-ping, Zhou Suo-quan, Xue Gen-yuan, et al. Numerical experiment of the coupling of RegCM2 and a Hydrological Model[J]. Journal of Nanjing Institute of Meteorology, 2006ୈ29ୄ2୅ୖ158-165. (In Chinese) [13] Sun Shufen. Physical, biochemical and parametric model of land surface process[M]. Beijing: China Meteorological Press, 2005:27-69;125- 209. (In Chinese) [14] Zhou Suoquan, Dai Kan, Chen Tao, et al. The improvement and verification of land surface process model(BATS)[J]. ActaMeteorologicaSinica, 2003,61ୄ3୅:275-290. (In Chinese) [15] Dickinson R E.The force-restore model for surface temperature and its generalizations[J].J Climateୈ1988ୈ1ୄ11୅: 1086-1098. [16] Chen Haishan, Ni Donghong, Li Zhongxian, et al. Modeling of impacts of vegetation coverage change on land surface conditions[J]. Journal of Nanjing Institute of Meteorology, 2006ୈ29ୄ6୅ୖ725-734. (In Chinese) [17] Ning Yuan, Qian Min, Wang Yutai. Water conservancy manual of the Huaihe River basin[M]. Beijing: Science Press,2003:1-7. (In Chinese) [18] Zillman J W. A study of some aspects of the radiation and heat budgets of the sourthern hemisphere oceans[M].Canberra:Australian Government Publishing Service,1972:742-753. [19] Liu Chunzhen, Cheng Bin, Cheng Xinwu, et al. A case study on simulating runoff with BATS Model[J].Journal of China Hydrology, 1998ୄ1୅: 8-13. (In Chinese) [20] Zhao Renjun. Xinanjiang Model [M]. Beijing: Water Conservancy and Electric Power Press, 1984ୖ255-269. (In Chinese)