Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Flood Prediction in Ungauged Basins by Physical-Based TOPKAPI Model

Flood Prediction in Ungauged Basins by Physical-Based TOPKAPI Model Hindawi Advances in Meteorology Volume 2019, Article ID 4795853, 16 pages https://doi.org/10.1155/2019/4795853 Research Article Flood Prediction in Ungauged Basins by Physical-Based TOPKAPI Model 1 1 2 Xiangyi Kong , Zhijia Li, and Zhiyu Liu College of Hydrology and Water Resources, Hohai University, 1 Xikang Road, Nanjing 210098, China Ministry of Water Resources Information Center, 2 Lane, Baiguang Road, Beijing 100053, China Correspondence should be addressed to Xiangyi Kong; kysadeur@yeah.net Received 6 March 2019; Revised 23 May 2019; Accepted 12 June 2019; Published 5 September 2019 Academic Editor: Francesco Viola Copyright © 2019 Xiangyi Kong et al. *is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Scarce historical flood data in ungauged basins make it difficult to establish empirical and conceptual model forecast in these areas. *e physical-based distributed model TOPKAPI is introduced for flood prediction in an ungauged basin by parameter transplant. Five main parameters are selected, and the sensitivity is analyzed by the GLUE method. *e Xixian basin and Huangchuan basin in the upper Huaihe basin in China are chosen as study areas. *e Xixian basin is regarded as a gauged basin for parameter calibration, and the Huangchuan basin is regarded as an ungauged basin by ignoring the historical discharge data. *e model is calibrated in gauged Xixian basin, and then parameters are directly transplanted to adjacent “ungauged” Huangchuan basin to simulate flood forecast in an ungauged basin. *e sensitivity analysis shows that soil thickness and soil saturated water content are the most sensitive parameters, and the Manning coefficient of main channel with high Strahler also significantly affects forecast results. According to the simulation results, the TOPKAPI model exhibits good performance in building and the prediction of the ungauged basin, in which the qualified rate of volume and peaks reaches 69.23%, and the average NSE criterion is over 0.67, which is acceptable forecast accuracy and has positive implication for the hydrological forecasting research. using partial missing observation data [7]. *ough the 1. Introduction world’s construction of hydrological stations continues to *e hydrological model is a vital tool for the current flood develop, there are still many small basins without moni- forecasting work by simulating the hydrological process. To toring stations; these small basins are often the focus areas of estimate the outflow process at the exit section, it generates a flood disaster research. For this reason, the flood forecasting certain flood runoff process in the basin under the cir- of ungauged basins has become the focus and hot topic of cumstance of a certain structure and parameters [1]. *e existing research, as well as a challenge for all hydrologists hydrological model basing on physical theory, according to [8]. the description method and generalization degree of the *e study on ungauged basin flood forecasting aroused hydrological process, can be split into conceptual hydro- the attention from hydrologists in the end of last century. Vandewiele et al. [9] used the method of regionalization of logical model and physical-based model [2]. *e Xin’anjiang model proposed by Zhao of Hohai University is one of the model parameters to forecast monthly discharge in unga- conceptual hydrological models [3]. It is characterized by uged region in 1991. Despite using the conventional re- less model parameters and high forecasting accuracy, and it gression equations to get model parameters, Burn and has been widely used in flood forecasting in various river Boorman [10] employed the basin classification method to basins in China [4]. However, as hydrogeological conditions find similar basins. Since usually there are insufficient his- of basins are highly generalized [5], the parameters should be torical discharge data in ungauged basins, the commonest calibrated using historical flood data [6], making it difficult approach to build the hydrological model in ungauged is to build a forecast model in the ungauged basins or basins parameter transferring [11–13], suggesting that the 2 Advances in Meteorology [32]), fully considers the spatial heterogeneity of the un- parameters required for models were calibrated in nearby basins with sufficient discharge data and then transplanted derlying surface conditions, exhibiting the advantage of relatively simple structure, clear parameter meaning, large to ungauged basins. *is approach, however, is based on the similarity of these two basins [14]. Gunter ¨ [15] summarized spatial scale elasticity, as well as wide application fields [33]. several methods for ungauged basin prediction using both Flood simulation and forecasting by this model have been event models and explicit soil moisture accounting (ESMA) widely performed in many river basins in Italy, Germany, models. To solve the description of characteristics of Spain, and other countries; good results have been achieved ungauged basins, Murugesu [16] proposed the approach of as well [30]. physical controls (basin form and function) of the in- In this study, the capability to simulate flood in unga- uged basin of TOPKAPI model is tested to give suggestions stantaneous unit hydrograph (IUH), which has been latterly achieved by Ellouze-Gargouri [17], using geomorphological on flood forecast in ungauged basins. *e sensitivity of the main parameters of the TOPKAPI model was analyzed, instantaneous unit hydrograph (GIUH) and Copulas. In 2012, Moore et al. [18] made a prediction in ungauged basins laying the theoretical basis for the parameter transplantation to build forecast model in ungauged basins. *e Xixian basin in British Columbia (BC) by transposing model parameter from adjacent basin; they also achieved good results. Patil located in the upstream of Huaihe River in China and the and Stieglitz [19], using a model developed themselves, adjacent Huangchuan basin were taken as the research analyzed selectively transferring sensitive versus insensitive basins. *e Huangchuan basin was considered the ungauged parameters on flood forecasting in ungauged basins. Athira basin by ignoring the historical flood data. Parameters were et al. [20] used SWAT model to explore the appropriate calibrated in the Xixian basin and transplanted to the ad- functional relationship between the parameters and basin jacent Huangchuan basin for flood simulation to test the application of TOPKAPI model in ungauged basins without characteristics. Waseem et al. [21] compared different in- terpolation schemes on parameter spatial distribution in parameter calibration. ungauged basins. Hyun and Choi [22] analyzed the re- lationship between flooding index and rainfall pattern to 2. The TOPKAPI Model forecast flood severity in ungauged basins. Sahoo et al. [23] adopted a novel routing method VPMD at ungauged rivers 2.1. Model Introduction. *e TOPKAPI model refers to a to plot rating curves. Canovas ´ et al. [24], using the 2D distributed hydrological model based on physical charac- hydraulic model in Spain, conducted the research on flood teristics. It splits the study basin into detailed grids, and thus discharge in ungauged basins. the difference of physical characteristic in spatial distribu- With the advancement of geographic information tion can be fully considered under the calculation unit technology, remote sensing technology, and computer sci- subdividing [34]. It can also give the specific condition of soil ence, even in the absence of hydrological observatories, the saturation, runoff yield, and confluence depth at each cal- geography of the watershed can be obtained from remote culation unit point in the basin, which is of high reference sensing images, including digital elevation model (DEM) [5] implication in real-time flood forecasting, land use and and land use and soil classification maps [25]. To fully ex- environmental impact assessment, and hydrological process ploit the advantage of such development, the physical-based simulation in ungauged basins [35]. distributed hydrological models were created [2, 26]. *ey According to this model, the hydrological process in describe the underlying surface conditions of basins using each calculation unit is generalized into three nonlinear physical parameters from remote senses, making it possible reservoir equations, which are similar in structure, repre- to simulate ungauged basin and become a new topic in the senting drainage in soil, surface runoff, and channel runoff field of hydrological forecasting [27]. Among them, the on saturated soil and impermeable surface, respectively. *e TOPKAPI rainfall-runoff model refers to a distributed hy- finite difference method is used in the calculation, so the drological model with grid-based computational unit based correlated four surrounding grids are considered during the on physical basis [28]. calculation of each grid unit [36]. *e parameters used to Professor Ezio Todini from the University of Bologna in describe the underlying surface conditions (e.g., soil Italy proposed the TOPKAPI (Topographic Kinematic thickness, vertical and horizontal saturated hydraulic con- Approximation and Integration) hydrological model in 1995 ductivity, coefficient of nonlinear reservoir equation, and [29]. *is model refers to a physics-based distributed hy- Manning coefficient of surface and evaporation coefficient of drological model based on the study of rainfall-runoff re- vegetation) could be extracted directly from the soil type lationship to explore the potential of hydrological model classification map and land surface utilization map. Besides, prediction based on physical theory in mountain flood the elevation and gradient distribution of the basin could be forecast [29]. *e model consists of several modules (e.g., reflected by DEM (digital elevation model). evapotranspiration, snowmelt, soil flow, surface runoff, river In terms of specific mechanism, TOPKAPI model em- runoff, and groundwater) [30]. In this model, the whole ploys *ornthwaite evaporation formula to calculate the basin is split into cell grids, and each grid is considered a potential evaporation of different vegetation covers in dif- single calculation unit. Each calculation unit reflects the ferent growth stages; it also calculates the actual evaporation whole physical hydrological process [31]. *is model, according to the actual wettability of upper soil [28]. compared with conventional conceptual hydrological According to the change of soil water content in the upper models (e.g., Xin’anjiang Model [3] and Sacramento Model unsaturated region, the runoff in the soil and the surface Advances in Meteorology 3 runoff are first calculated and then aggregated to the total upper reaches of the Huaihe River were taken as research runoff. In the confluence part, the nonlinear reservoir area. *e Wangjiaba basin is located in the southern part of equation is adopted to merge surface runoff and un- Henan Province of China. It is a monsoon climate with an derground runoff for confluence calculation [28, 37]. *e average annual rainfall of 1060 mm [30]. About 50% of the main parts and approaches of TOPKAPI model are listed in rainfall concentrates in the flood season, and the Xixian and Figure 1. Huangchuan river basins are the origins of the upstream To describe the movement of water in surface, soil, and river. *e control area of Xixian Hydrological Station is 10 river channel, TOPKAPI model adopts three nonlinear 190 km . *ere are two large reservoirs in Nanwan (water reservoir equations with a similar structure. In the earliest catchment area 1058 km ) and Shishankou (water catch- version, the variable step five-stage Runge–Kutta numerical ment area 306 km ) [41]. After deducting two large reser- method is employed to solve the differential equations. Later, voirs, the catchment area is 8400 km [42]. *e elevation it was found that under appropriate approximation con- range of the Xixian County is 37–963 m [43]. *e terrain in ditions, these equations could be solved using analytical the northeast is relatively flat, being mountainous in the methods [28]. west, southwest, and south. *e soil type is mostly clay soil. It has soil types (e.g., loam, sandy loam, and clay). Most of the land is farmland, with a small portion of forest land and 2.2. Comparison with Conceptual Model. From the aspect of mixed forest land. *e Huangchuan basin outlet station has data required for the model building, the conceptual hy- a controlled area of 1989 km with Pohe reservoir built drological model should be based on the previous historical inside [44]. *e elevation range of the area is 36–984 m. *e hydrological data. Taking Xin’anjiang model as an example, watershed is mostly mountainous, the north is relatively flat, the parameters, including water storage capacity (WM), most of the soil types are clay and sandy loam, and the evapotranspiration coefficient (K), outflow and regression surface vegetation is primarily farmland [45, 46]. coefficient of the basin (KI and KG), should be calibrated *e digital elevation data of the basins could be obtained through historical floods [3]. In other words, conceptual from the SRTM global 90 m digital elevation model provided model is based on high-degree generalization of those by the Consultative Group on International Agricultural physical parameters to reflect the characteristics of basins. Research (Figure 3). GIS software was employed to extract *is model has the advantage of relatively flexible adjustable the simulated stream nets and basin boundaries in the Xixian parameters that can be fully adjusted to give optimized and Huangchuan basins. *e rainfall data of 13 rainfall forecast results and avoid some relatively complex mecha- stations and the discharge data of Xixian hydrological station nisms in the hydrological process [38]. However, the dis- and that of the two reservoirs were collected from 1980 to advantage lies in the reliance on historical flood data, which 2003 according to the distribution of stations in the basin. is particularly evident in some areas in the absence of his- *e discharge data, as well as the rainfall data of 5 rainfall torical rainfall and runoff data where it is almost likely to stations in the Huangchuan basin and the discharge data of calibrate parameters [39]. *e existing feasible method is to the Huangchuan hydrological station from 2003 to 2005, search for data basins with similar geological and hydro- were processed into 1 h interval. *e meteorological data of logical conditions for parameter transplantation, which will Xinyang, Zhumadian, and Guangshan meteorological sta- inevitably face systematic error of the model. In contrast, tions (including daily maximum and minimum tempera- most of the data required for describing hydrological con- tures, wind speed, humidity, and sunshine hours) were ditions and underlying surface conditions of watershed collected from 1980 to 2005. based on physical basis can be obtained through actual measurement (e.g., DEM, soil classification and soil thick- ness, hydraulic conductivity, and evaporation coefficient of 3.2. Underlying Surface Analysis. *e underlying surface land classification) [34]. *ese parameters are relatively easy parameters of the TOPKAPI model were transplanted on the to obtain online, including the Shuttle Radar Topography premise that the soil types and land cover of underlying Mission (SRTM) launched by the Consultative Organization surface of the two basins should be of the same type or have for International Agricultural Research (CGIAR), which can similar physical properties, so the distribution of soil obtain the global 90 m resolution digital elevation model free (Figure 4 and Table 1) and surface in the two basins (Figure 5 of charge from its official website; the accessible website of and Table 2) significantly affected the ungauged area pre- the University of Maryland (UMD) in the United States. On diction [47]. According to the collected data, the Xixian and this point, the hydrological model based on physical basis Huangchuan basins were located close to each other. *eir has obvious advantages over the conventional conceptual soil types were similar to clay, and most of the land surface hydrological model when it is relatively easy to obtain the was the same type of farmland. *e soil parameters and land parameters required for model building in ungauged areas use parameters calibrated by the Xixian basin in the [28, 40]. TOPKAPI model were applicable to the adjacent Huang- chuan basin. *e soil data are derived from the FAO soil map of the world, Global soil profile databases, which is 3. Study Area and Dataset available online at http://www.ngdc.noaa.gov/seg/eco/ cdroms/reynolds.htm. *e land cover map is derived 3.1. Study Area. *e neighboring Xixian and Huangchuan from National Administration of Surveying Mapping and river basins (Figure 2) above the Wangjiaba basin in the 4 Advances in Meteorology Meterological data Digital elevation Soil type map Land cover map (precipitation and temperature) model Spatial interpolation Calculation grid Evapotranspiration (Thornthwaite equation) Excess Percolation Soil water Surface water storage Subsurface flow Surface flow River channel Muskingum–cunge routing Channel outflow Figure 1: Main components and mechanism of TOPKAPI model. 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E ZhuMaDian Xixian NanWan HuangChuan XinYang ShiShanKou GuangShan PoHe 60 Kilometers 0 15 30 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E ◊ Meteorological stations Xixian basin Hydrological stations Reservoir Huangchuan basin (a) Figure 2: Continued. 31°30′0″N 32°0'0″N 32°30′0″N 33°0′0″N 31°30′0″N 32°0′0″N 32°30′0″N 33°0′0″N Advances in Meteorology 5 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E HuangGang XiaoCaoDiao DaPoLing TongBai HuJiaWan ChangTaiGuan LuoShan WuLiDian NanWan HuangChuan NanLiDian GuangShan PengXinDian XinDian DingYuanDian PoHe WuChenHe XinXian 0 15 30 60 Kilometers 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E Precipitation gauges Xixian basin Huangchuan basin Reservoir (b) Figure 2: Location of the study area Xixian and Huangchuan basins and the distribution of reservoirs and (a) meteorological and hy- drological stations and (b) precipitation gauges. Elevation (m) Figure 3: ƒe digital elevation model of study area. 31°30′0″N 32°0′0″N 32°30′0″N 33°0′0″N 32°0′0″N 32°30′0″N 33°0′0″N 6 Advances in Meteorology 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E N N 0 15 30 60 kilometers 0 5 10 20 kilometers 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E Sand clay loam Sand clay loam Loam Loam Clay Clay (a) (b) Figure 4: Soil types map of (a) Xixian basin and (b) Huangchuan basin. Table 1: Component of soil types in Xixian and Huangchuan basin. Soil type Code Percentage (%) Loam 3085 15.65 Xixian basin Sandy loam 3963 12.87 Clay 4326 71.48 Loam 3085 14.14 Huangchuan basin Sandy loam 3963 32.12 Clay 4326 53.74 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E N N 0 15 30 60 kilometers 0 5 10 20 kilometers 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E Farmland Forest Farmland Forest Grass Mixed forest Grass Mixed forest (a) (b) Figure 5: Land use map of (a) Xixian basin and (b) Huangchuan basin. 31°30′0″N 32°0′0″N 32°30′0″N 31°30′0″N 32°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N 32°30′0″N 31°30′0″N 32°0′0″N 32°30′0″N 31°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N Advances in Meteorology 7 Table 2: Component of land use in Xixian and Huangchuan basins. Land use type Code Percentage (%) Farmland 6 67.70 Forest 8 18.41 Xixian basin Mixed forest 10 2.81 Grass 11 11.09 Farmland 6 69.92 Forest 8 10.55 Huangchuan basin Mixed forest 10 5.03 Grass 11 14.50 Geoinformation of China (NASG), which is free down- a � r + Q , c c loadable for all Internet at http://www.globallandcover.com/ √�� 2/3 GLC30Download/index.aspx and the characteristic data for s (sin c) b � , each type of land cover are derived from Corinne Land 1/3 (4) 2/3 4/3 2 n (tan c) X Cover 2006 raster data by European Environment Agency, which is accessible at https://www.eea.europa.eu/data-and- maps/data/clc-2006-raster. c � , where r denotes the lateral inflow, including soil outflow to 4. Parameter Sensitivity Analysis surface and channel. *e calculation formulas show that the main influence in 4.1. Parameter Selection. In TOPKAPI model, the nonlinear the calculation of soil water originates from the upstream reservoir equation can be written as inflow, grid size, and hydraulic conductivity, in which hy- dy draulic conductivity as the unknown coefficient of the (1) � a − by , nonlinear equation will directly affect the solution of the dt unknown term. *us, hydraulic conductivity is considered where the variable y denotes average soil water content, river the key parameter to be investigated. *e calculation of storage, surface water depth, etc. Variables a, b, and c keep surface water on slope is affected by slope, precipitation and constant in a single time step. surface Manning coefficient, and surface runoff after In the calculation of groundwater, the variables a, b, and deducting infiltration is associated with soil conditions on c are expressed as follows: underlying surface. Accordingly, in the part of surface flow 2 u u calculation, the Manning coefficient of surface, the saturated pX + Q + Q 0 s a � , water content associated with soil characteristics, and the soil thickness describing soil volume were taken as the key (2) research parameters. In the calculation of channel flow using b � , nonlinear reservoir method, the Manning coefficient of river roughness was taken as the main calibration parameter. All c � a , the selected parameters are listed in Table 3. where p denotes the precipitation, X is the size of grid unit, u u Q and Q refer to surface and subsurface inflow of a grid 0 s 4.2. Sensitivity Analysis. *e GLUE method was first pro- unit, respectively, and C is the local hydraulic conductivity. posed by Beven and Binley in 1992 based on the RSA (re- In the calculation of surface water, the variables a, b, and gionalized sensitivity analysis) method proposed by c are expressed as follows: Hornberger and Spear in 1981 [48]. Beven and Binley 1 V proposed that this method is not the numerical value of a exf a � , single parameter, but the value of a group of parameters that XW dt significantly affect the prediction results of the model. In this 1/2 method, Monte Carlo method is adopted to sample a series tan (β) (3) b � , of parameter combinations in prior parameter range and n X substitute them into the model calculation. Likelihood function is selected as the evaluation and instruction of the c � a , parameter sets. *e parameter set whose likelihood function where W denotes the width of grid without river channel, value is higher than a certain threshold is considered to be V represents the net precipitation, and a is index in able to represent the characteristic of study basin (which is exf 0 Manning equation with a constant value of 5/3. called “behavioral”), whereas the parameter set with likeli- In the calculation of channel water propagation, the hood function value lower than the threshold is considered variables a, b, and c are expressed as follows: fail at reflecting it and be abandoned [49]. 8 Advances in Meteorology Table 3: Main parameters of the TOPKAPI model. Modules Parameter Description Groundwater Horizontal hydraulic conductivity Reflecting water conduction rate in soil Calculating factors of water conduction rate on the Surface Manning coefficient surface When the soil water content reaches the saturated Saturated water content Surface runoff state, the volume percentage of water content *e soil depth between soil and bedrock reflects the Soil thickness total volume of soil on the underlying surface in the calculation unit Channel runoff Riverbed Manning coefficient Calculation factors of water conduction rate in rivers Several criterions could be used to describe the pre- saturated water content, and riverbed Manning coefficient. diction performance of a model, among which the Nash– Soil type with high coverage percentage such as clay (with Sutcliffe model efficiency coefficient (NSE) [50] has been 71.48% coverage) has significantly high standard deviation most widely used in model prediction performance judg- value for both thickness and saturated water content. *is is ment. In this study, the NSE is chosen as the likelihood the same for the land use Manning coefficient of farmland function. *e Nash–Sutcliffe model efficiency coefficient is (with 67.70% coverage) and the riverbed Manning co- efficient of V level river channel. expressed as T 2 t t 􏽐 Q − Q 􏼁 t�1 m o NSE � 1 − , (5) 5. Application in Ungauged Basins 􏽐 Q − Q 􏼁 t�1 o To verify the effectiveness of the model in ungauged basins, t t where Q and Q denote the modeled and observed dis- m o experiments were performed. *e Xixian basin and adjacent charge at time t, respectively, and Q is the average value of Huangchuan basin in Wangjiaba were taken as study areas. observed discharge over the whole time series. *e parameters of the model were calibrated in Xixian basin In this study, 3000 sets of parameters including those and subsequently transferred to Huangchuan basin, and selected in the previous part were generated by random flood simulations were carried out to verify the application Monte Carlo sampling methodology within the prior range effect of the model in ungauged basins. in Table 4. A threshold of 0.7 NSE is chosen to check whether *e calibration and validation results are evaluated by these parameter sets are behavioral. Parameter uncertainty “qualified rates” [52]. A flood forecast with flood volume and was analyzed by 8 floods in Xixian basin from 1980 to 2002 flood peak error less than 20% is considered to be “qualified,” in TOPKAPI model. A total of 283 sets of parameters were and the “qualified rate” means the percentage of qualified found to be behavioral achieving 0.7 NSE likelihood. flood simulations over the total number. *e whole forecast Posterior parameter distribution was derived from 283 scenario accuracy will be evaluated to a certain level behavioral parameter sets. To get visual impression of these according to Table 6. parameter sets, the histograms of all 18 parameters distri- bution are plotted in Figure 6. Histograms with sharp peaks or steep slopes means parameter are well identified, whereas 5.1. Parameter Calibration. Most parameters in physical- parameters with relative flat histograms are associated with based hydrological model could be obtained by actual more uncertainty [51]. Because all the behavioral sets have measurement; however, the measured values acquired on the performance over threshold in common, narrow and sharp point measurement might not be sufficiently representative peaks reflect relative strong relationship with model results at the calculation unit scale and might not reflect the and show that this parameter has more sensitivity, which temporal and spatial scale change within the unit. *us, the means variation in this parameter may significantly affect parameters obtained by measurement or reference to the model results. A statistic of parameter frequency was con- literature should be fine-tuned by trial before the simulation ducted in Table 5 by calculating standard deviation of fre- to compensate for the error caused by the spatial general- quencies of each parameter to get a numerical comparison. ization of the parameters and the time generalization of the In Table 5, the standard deviations of parameters fre- input parameters, making the results of the simulation closer quency distribution are listed. It can be clearly seen that to the actual situation. horizontal hydraulic conductivity has the least values and *e calibration parameters included soil horizontal the histograms are all relatively flat, meaning that this pa- water conductivity, saturated water content, land surface, rameter has low sensitivity and high uncertainty. *is sort of and riverbed Manning coefficient for Muskingum algorithm. parameters are difficult to be calibrated due to their weak In the previous GLUE assessment, 3000 sets of parameters relationship with model outputs, but they can only slightly have been generated. Several “best” performed sets of pa- affect the model performance. However, emphasis in cali- rameters with reasonable values and high NSE are chosen for bration should be put on those parameters with high manually adjustment. *ese parameters were calibrated by standard deviation and sensitivity such as soil thickness, the rainfall and discharge data of 16 floods from 1991 to 2003 θ Advances in Meteorology 9 Table 4: Prior range of parameters. Loam Sandy loam Clay Soil thickness (m) 1.3–1.8 0.4–1.1 0.8–1.6 − 6 Horizontal hydraulic conductivity (10 m/s) 2.26–9.05 0.40–1.62 0.17–0.71 Saturated water content 0.35–0.50 0.27–0.33 0.30–0.55 Farmland Forest Mixed forest Grass Surface Manning coefficient 0.06–0.25 0.20–0.35 0.12–0.27 0.06–0.25 I II III IV V Riverbed Manning coefficient 0.05–0.30 0.02–0.15 0.02–0.15 0.02–0.15 0.02–0.15 15 15 10 10 5 5 0 0 0 1.5 1.6 1.7 0.5 0.6 0.7 0.8 0.9 1 1.2 1.4 1.6 L (m) L (m) L (m) 5 5 0 0 0 48 6 0.5 1 1.5 2 34567 –6 –6 –7 k (m/s) ×10 k (m/s) k (m/s) ×10 ×10 sh sh sh 0 0 0 0.38 0.4 0.42 0.44 0.46 0.48 0.28 0.29 0.3 0.31 0.32 0.35 0.4 0.45 0.5 Loam Sandy loam Clay (a) 15 15 10 10 0 0 0 0 0.1 0.15 0.2 0.25 0.22 0.24 0.26 0.28 0.3 0.32 0.34 0.15 0.2 0.25 0.1 0.15 0.2 0.25 Farmland Forest Mixed forest Grass (b) Figure 6: Continued. Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) 10 Advances in Meteorology 0 0 0 0 0.1 0.15 0.2 0.25 0.04 0.06 0.08 0.1 0.12 0.05 0.1 0.15 0.05 0.1 0.15 0.02 0.04 0.06 0.08 0.1 0.12 I II III IV V (c) Figure 6: Frequency histograms of (a) soil thickness L, horizontal hydraulic conductivity (k ), and saturated water content of three types of sh soil, (b) surface Manning coefficient of fore types of land use, and (c) riverbed Manning coefficient of five river classifications. Table 5: Standard deviation of parameter frequency distribution. Loam Sandy loam Clay Soil thickness (m) 2.5123 3.2393 6.0154 − 6 Horizontal hydraulic conductivity (10 m/s) 0.7734 0.7244 0.5807 Saturated water content 2.7310 2.0651 5.4510 Farmland Forest Mixed forest Grass Surface Manning coefficient 3.7226 0.6719 0.7198 1.7073 I II III IV V Riverbed Manning coefficient 2.0449 1.4837 2.4079 4.5080 8.5721 applicability in this basin. According to the simulated flood Table 6: Forecast scenario classification level. volume, the qualified rate of TOPKAPI model flood volume Forecast level A B C was 81.3% in the 16 floods used for calibration and 93.8% in Qualified rate the flood peak simulation. From the results, it is revealed that QR> 85.0 85> QR> 70.0 70.0> QR> 60.0 (%) under the underlying elevation data, soil and land use NSE NSE> 0.90 0.90> NSE> 0.70 0.70> NSE> 0.60 distribution parameters and basin hydrometeorological data meet the demand, the model achieved relatively good simulation in the study area. *e average NSE were all above in Xixian basin. Run the TOPKAPI model with the mete- 0.7, achieving the accuracy of Class B forecasting [52], and orological data (including precipitation and temperature) of the average error between flood volume and flood peak was 13 floods in Huangchuan watershed from 2003 to 2005 controlled within 20%. *is is associated with the wet area together with the calibrated parameters and adjusted these where the research basin is located. *e full storage and parameter according to the outcome flood results. Fur- runoff yield models used in the model were applicable to the thermore, to verify the accuracy of the simulation and to wet area of Huaihe River, and the better the hydrogeological demonstrate the application of the model in ungauged conditions of the basin, the smaller the error of the model basins, the flood simulation results were compared with the confluence calculation will be. *us, TOPKAPI model ex- measured results. hibits good applicability in this basin. All the calibrated *e flood peak and NSE were taken as objective func- parameters are listed in Tables 8–10. tions to calibrate the parameters. *e percentage of initial From the process of model building, TOPKAPI model soil water content in the basin would significantly affect the did not adjust too much parameters in the calibration simulation results of the early floods in the basin. In this process, and the prediction results were good. Accordingly, calibration, to determine the initial soil moisture, the TOPKAPI model based on physical basis exhibited better “warm-up” method [53] was adopted to simulate the initial applicability when building model in adjacent areas without state through natural situation, which was initiated 30 days historical flood data. However, the prediction accuracy of before the first flood in advance. TOPKAPI model depended on the high-precision partition Figure 7 shows that, in the calibration, most of the of computational grids, whereas high-resolution grids would stream flow series are well reproduced. Floods with large increase the requirement of computational ability. *us, to water volume are simulated better than those with low obtain higher prediction accuracy, the model requires better volume. From the results of model calibration in Table 7, the operation equipment and long calculation time. simulation results of the model in Xixian basin were good, and the average NSE were above 0.8. Among the 16 floods used in the calibration, 14 floods had NSE over 0.8, taking up 5.2. Parameter Validation. *e parameters obtained by the 87.5% and 6 floods over 0.9. In terms of Nash–Sutcliffe calibration in adjacent basin were directly applied in model efficiency coefficient, the model exhibited certain Huangchuan basin. Besides, the flood discharge process of Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Advances in Meteorology 11 1000 2000 3000 4000 5000 6000 7000 8000 Simulation steps (1 h ) Observation Simulation Precipitation Figure 7: Calibration hydrograph from 1991 to 2003. Table 7: Calibration results in Xixian basin. Flood volume Flood peak Flood no. NSE 4 3 3 Observation (10 m ) Simulation error percentage (%) Observation (m /s) Simulation error percentage (%) 10523199 39933.54 8.87 1300 1.77 0.76 10530199 46156.68 3.72 1670 − 2.86 0.89 12106199 110651.16 − 24.55 5060 − 28.19 0.83 10629199 73803.71 15.33 2960 − 1.97 0.91 10805199 92912.93 − 10.04 4420 − 3.97 0.83 10021992 19089.25 − 7.37 579 12.61 0.84 18199305 14123.72 3.36 516 3.86 0.7 19950707 42848.26 − 1.2 2300 2.68 0.78 17199607 81739.98 − 22.99 4450 − 19.46 0.89 19960802 25732.97 − 5.79 875 1.87 0.71 19970629 24946.02 2.81 1220 12.41 0.91 19980630 54968.35 16.86 2510 − 12.89 0.94 20020622 97511.4 14.1 5080 − 0.62 0.9 20020627 61223.48 1.34 2820 − 8.32 0.85 20020723 55287.15 13.27 2790 − 6.49 0.84 20030629 107940.6 6.67 3900 − 13.52 0.77 Abs. average 8.64 8.34 Average 2.15 − 3.94 0.84 Qualified rate 93.75% 93.75% 100.00% Table 8: Calibrated soil parameters. Table 10: Calibrated channel Manning coefficient. Loam Sandy loam Clay Order I Order II Order III Order IV Order V L (m) 1.7675 0.78856 1.4715 N 0.15918 0.085962 0.058017 0.046145 0.03347 Channel k (m/s) 4.55e‒ 6 8.4e‒ 7 4.13e‒ 7 ∗ sh N : channel Manning coefficient. Channel θ 0.48511 0.307 0.49886 L: soil thickness; k : horizontal hydraulic conductivity; θ : saturated water sh s Table 11, Figure 8, and the box plot Figure 9 of calculation content. and results, it is revealed that the overall NSE was relatively lower because of parameter transplantation. *e average Table 9: Calibrated surface Manning coefficient. NSE was 0.67, and the qualified rate of flood volume and peak simulation was 69.23%. *is was probably because Farmland Forest Mixed forest Grass Xixian is in the warm temperate zone and belongs to the N 0.182871 0.274005 0.239738 0.11128 Surf semihumid basin, and the overall basin is dominated by N : surface Manning coefficient. Surf excess of storage mode, which is consistent with the basic assumption of TOPKAPI model runoff generation. *e the outlet section was calculated using the rainfall data of 13 accuracy forecast in ungauged basins, compared with the floods in 2003–2005 as input to simulate the application of calibration results, decreased in varying degrees. *is was the model in the ungauged basin. From the statistical because the parameter transplantation inevitably brings Discharge (m /s) Precipitation (mm) 12 Advances in Meteorology Table 11: Application results in ungauged basin. Flood volume Flood peak Flood no. NSE 4 3 4 3 Observation (10 m ) Simulation error percentage (%) Observation (10 m ) Simulation error percentage (%) 20030622 6872.8 41.84 705 30.78 0.65 20030626 4765.9 38.09 380 54.42 0.8 20030629 28032.16 − 7.07 2180 2.48 0.83 20030707 39802.79 − 25.56 1750 − 5.62 0.79 19200307 6711.57 − 34.86 419 − 18.61 0.48 15200308 4060.62 − 13.1 173 39.97 0.63 20040716 15826.68 2.48 1070 11.64 0.72 20040801 11528.6 5.97 825 5.6 0.67 20040813 14726.48 − 0.84 1200 28.18 0.85 20050726 9259.52 − 16.38 600 − 5.13 0.62 20050820 8599 − 14.45 402 19.06 0.5 20050828 9145.8 − 7.8 733 − 6.9 0.59 20050902 11192.11 − 17.65 666 2.81 0.57 Abs. average 17.39 17.78 Average − 3.79 12.21 0.67 Qualified rate 69.23% 69.23% 38.46% 2500 0 100 200 300 400 500 600 700 800 900 1000 1100 Simulation steps Observation Simulation Precipitation (a) 2000 0 100 200 300 400 500 600 700 800 900 1000 1100 Simulation steps Observation Simulation Precipitation (b) Figure 8: Continued. 3 3 Discharge (m /s) Discharge (m /s) Precipitation (mm) Precipitation (mm) Advances in Meteorology 13 1000 0 100 200 300 400 500 600 700 800 900 1000 Simulation steps Observation Simulation Precipitation (c) Figure 8: Hydrograph of simulation in ungauged basins in 2003, 2004, and 2005. –40 (a) (b) (c) (d) Figure 9: Box plot error percentage of (a) calibration flood volume, (b) calibration flood peak, (c) ungauged flood volume, and (d) ungauged flood peak. about inconsistencies with the actual conditions, which is analysis and comparison of calibration simulation results also the most important problem facing the prediction of with measured data and prediction results of the model in ungauged basins. adjacent similar ungauged basins, parameter transplantation It was found that the flood process of TOPKAPI model was tested. Based on the results, the following conclusions had steep rise and fall just, as shown in the hydrographs. were drawn: Accordingly, following the flood peak priority calibration (1) *e description of basin characteristics in TOPKAPI strategy, the systematic error of overall smaller flood volume model can be characterized by underlying surface appeared. *e water confluence process was faster than the data from land use and soil type obtained using actual situation, which is also one of the parts of the model to remote sensing technology, which is easy to obtain be further optimized. Nevertheless, TOPKAPI model still and use. In this study, it was assumed that reflects good transplantability in flood peak prediction in Huangchuan basin is an ungauged basin without any ungauged basins. Since remote sensing data could reflect hydrological stations, whereas underlying surface actual topography, vegetation, and soil types as much as data can still be acquired from satellite or remote possible, the eligibility rate of flood peak prediction reached sensing equipment, making it likely to describe the about 70%. *is is the advantage brought by detailed hy- basin in hydrological model. drological process based on the physical basis of the dis- (2) According to the results of the sensitivity analysis of tributed model, fully displaying the physical conditions of the model parameters, the soil thickness and surface underlying surface. Manning coefficient, with a large proportion of coverage, significantly affected the prediction results 6. Discussion and Conclusion of the model. *e horizontal hydraulic conductivity In this study, a physically based distributed hydrological has the least sensitivity. Since the classification of model TOPKAPI was applied in upper Huaihe, Xixian, and river depends on the area of the whole river basin adjacent Huangchuan basin. *e model uncertainty caused and other factors, some differences would exist by spatial and temporal generalization of model parameters among different river basins for the same order. In was evaluated using GLUE method. *rough uncertainty this study, this is also a reason for the reduction of Error percentage (%) Discharge (m /s) Precipitation (mm) 14 Advances in Meteorology prediction accuracy of ungauged river basins. *e References rules of parameter transplantation in prediction [1] R. Johnston and V. Smakhtin, “Hydrological modeling of research should be elucidated further. large river basins: how much is enough?,” Water Resources (3) *e model exhibits good applicability in the upper Management, vol. 28, no. 10, pp. 2695–2730, 2014. Huaihe basin. *e base assumptions of TOPKAPI [2] M. Cai, S. Yang, H. Zeng, C. Zhao, and S. Wang, “A dis- model made it feasible in humid semihumid area, tributed hydrological model driven by multi-source spatial and thus the flood volume flood peak as long as NSE data and its application in the Ili river basin of Central Asia,” all achieved high accuracy in calibration. Water Resources Management, vol. 28, no. 10, pp. 2851–2866, (4) In the prediction of ungauged basins, it was found [3] R. J. Zhao and X. R. Liu, “*e Xinanjiang model,” Computer that under premise of sufficient remote sensing data Models of Watershed Hydrology, vol. 135, pp. 371–381, 1995. of underlying surface, the transplantation of pa- [4] Z. Ren-Jun, “*e Xinanjiang model applied in China,” Journal rameters among similar basins could achieve good of Hydrology, vol. 135, no. 1–4, pp. 371–381, 1992. forecasting results. [5] A. H. A. Suliman, M. Jajarmizadeh, S. Harun, and I. Z. Mat Darus, “Comparison of semi-distributed, GIS-based hydro- Besides, the high prediction accuracy of TOPKAPI logical models for the prediction of streamflow in a large model should be based on detailed meshing of computing catchment,” Water Resources Management, vol. 29, no. 9, units. For instance, the grid resolution adopted in this study pp. 3095–3110, 2015. was 200 m, and the prediction results were relatively good. [6] A. Kushwaha and M. K. Jain, “Hydrological simulation in a However, the calculation period took a long time. If the forest dominated watershed in Himalayan Region using model is applied in a larger watershed, the problem will be SWAT model,” Water Resources Management, vol. 27, no. 8, more prominent. *us, how to make an optimal trade-off pp. 3005–3023, 2013. between the calculation time of the model and the prediction [7] A. C. Guzha and T. B. Hardy, “Application of the distributed accuracy needs further studies. hydrological model, TOPNET, to the big darby creek wa- tershed, Ohio, USA,” Water Resources Management, vol. 24, no. 5, pp. 979–1003, 2010. Data Availability [8] W. *orsten and M. Alberto, “Convergence of approaches *e digital elevation model data can be acquired from toward reducing uncertainty in predictions in ungauged basins,” Water Resources Research, vol. 47, no. 6, pp. 453–460, CGIAR (http://srtm.csi.cgiar.org/). *e soil data are derived from the FAO soil map of the world, Global soil profile [9] G. L. Vandewiele, C.-Y. Xu, and W. Huybrechts, “Region- databases, which is available online at http://www.ngdc. alisation of physically-based water balance models in Belgium. noaa.gov/seg/eco/cdroms/reynolds.htm. *e land cover Application to ungauged catchments,” Water Resources map is derived from National Administration of Surveying Management, vol. 5, no. 3-4, pp. 199–208, 1991. Mapping and Geoinformation of China (NASG), which is [10] D. H. Burn and D. B. Boorman, “Estimation of hydrological freely downloadable for all Internet at http://www. parameters at ungauged catchments,” Journal of Hydrology, globallandcover.com/GLC30Download/index.aspx and the vol. 143, no. 3-4, pp. 429–454, 1993. characteristic data for each type of land cover are derived [11] L. Oudin, V. Andreassian, ´ C. Perrin, C. Michel, and from Corine Land Cover 2006 raster data by European N. L. Moine, “Spatial proximity, physical similarity, regression Environment Agency, which is accessible at https://www. and ungaged catchments: a comparison of regionalization eea.europa.eu/data-and-maps/data/clc-2006-raster. approaches based on 913 French catchments,” Water Re- sources Research, vol. 44, no. 3, pp. 893–897, 2008. [12] L. Oudin, A. Kay, V. Andreassian, ´ and C. Perrin, “Are Conflicts of Interest seemingly physically similar catchments truly hydrologically similar?,” Water Resources Research, vol. 46, no. 11, 2010. *e authors declare that they have no conflicts of interest. [13] R. Singh, S. A. Archfield, and T. Wagener, “Identifying dominant controls on hydrologic parameter transfer from gauged to ungauged catchments—a comparative hydrology Acknowledgments approach,” Journal of Hydrology, vol. 517, pp. 985–996, 2014. *is work was supported by the National Key Research and [14] S. A. Archfield and R. M. Vogel, “Map correlation method: Development Program of China (grant no. 2018YFC1508103) selection of a reference streamgage to estimate daily streamflow at ungaged catchments,” Water Resources Re- and the Fundamental Research Funds for the Central Uni- search, vol. 46, no. 10, pp. 5613–5618, 2010. versities (grant no. 2016B04714). [15] B. Gu¨nter, “Rainfall-runoff modeling of ungauged catchments,” in Encyclopedia of Hydrological Sciences, M. G. Anderson and Supplementary Materials J. J. McDonnell, Eds., John Wiley & Sons, Ltd., Hoboken, NJ, USA, 2006. *e following data are available in supplementary materials: [16] S. Murugesu, “Prediction in ungauged basins: a grand chal- the precipitation data, temperature, and discharge data used lenge for theoretical hydrology,” Hydrological Processes, for parameter calibration in Xixian basin and the coordinate vol. 17, no. 15, pp. 3163–3170, 2010. of meteorological stations and precipitation gauges used in [17] E. Ellouze-Gargouri and Z. Bargaoui, “Runoff estimation for an this research. (Supplementary Materials) ungauged catchment using geomorphological instantaneous Advances in Meteorology 15 unit hydrograph (GIUH) and copulas,” Water Resources Digital Computers, U.S. Department of Commerce, National Management, vol. 26, no. 6, pp. 1615–1638, 2012. Weather Service, and State of California, Department of [18] R. D. D. Moore, J. W. Trubilowicz, and J. M. Trubilowicz, Water Resources, Sacramento, CA, USA, 1973. “Prediction of streamflow regime and annual runoff for [33] P. Deng, Z. Li, and Z. Liu, “Numerical algorithm of distributed ungauged basins using a distributed monthly water balance TOPKAPI model and its application,” Water Science & En- model,” Jawra Journal of the American Water Resources As- gineering, vol. 1, no. 4, pp. 14–21, 2008. sociation, vol. 48, no. 1, pp. 32–42, 2012. [34] Z. Liu, “Application of GIS-based distributed hydrological [19] S. Patil and M. Stieglitz, “Modelling daily streamflow at model to flood forecasting,” Journal of Hydraulic Engineering, ungauged catchments: what information is necessary?,” Hy- vol. 35, pp. 70–75, 2004. drological Processes, vol. 28, no. 3, pp. 1159–1169, 2014. [35] G. Coccia and E. Todini, “Recent developments in predictive [20] P. Athira, K. P. Sudheer, R. Cibin, and I. Chaubey, “Pre- uncertainty assessment based on the model conditional dictions in ungauged basins: an approach for regionalization processor approach,” Hydrology and Earth System Sciences, of hydrological models considering the probability distribu- vol. 15, no. 10, pp. 3253–3274, 2011. tion of model parameters,” Stochastic Environmental Research [36] S. Sinclair and G. G. S. Pegram, “A comparison of ASCATand and Risk Assessment, vol. 30, no. 4, pp. 1131–1149, 2016. modelled soil moisture over South Africa, using TOPKAPI in [21] M. Waseem, J.-y. Shin, and T.-W. Kim, “Comparing spatial land surface mode,” Hydrology and Earth System Sciences, interpolation schemes for constructing a flow duration curve vol. 14, no. 4, pp. 613–626, 2010. in an ungauged basin,” Water Resources Management, vol. 29, [37] L. I. Zhijia, X. Wang, L. Yanxiang, L. Chen, and L. I. Lanru, no. 7, pp. 2249–2265, 2015. “Application of TOPKAPI model and comparison with [22] A. J. Hyun and H. I. Choi, “A new flood index for use in Xin’anjiang model,” Water Power, vol. 11, 2013. evaluation of local flood severity: a case study of small [38] S. Yang, G. Dong, D. Zheng, H. Xiao, Y. Gao, and Y. Lang, ungauged catchments in Korea,” Jawra Journal of the “Coupling Xinanjiang model and SWAT to simulate agri- American Water Resources Association, vol. 49, no. 1, pp. 1–14, cultural non-point source pollution in Songtao watershed of Hainan, China,” Ecological Modelling, vol. 222, no. 20–22, [23] B. Sahoo, M. Perumal, T. Moramarco, and S. Barbetta, “Rating pp. 3701–3717, 2011. curve development at ungauged river sites using variable [39] H. Lu, ¨ T. Hou, R. Horton et al., “*e streamflow estimation parameter muskingum discharge routing method,” Water using the Xinanjiang rainfall runoff model and dual state- Resources Management, vol. 28, no. 11, pp. 3783–3800, 2014. parameter estimation method,” Journal of Hydrology, vol. 480, ´ ´ [24] J. A. B. Canovas, M. Eguibar, J. M. Bodoque, A. Dıez-Herrero, pp. 102–114, 2013. M. Stoffel, and I. Gutierrez-P ´ erez, ´ “Estimating flash flood [40] S. Sinclair and G. G. S. Pegram, “A sensitivity assessment of discharge in an ungauged mountain catchment with 2d hy- the TOPKAPI model with an added infiltration module,” draulic models and dendrogeomorphic paleostage in- Journal of Hydrology, vol. 479, pp. 100–112, 2013. dicators,” Hydrological Processes, vol. 25, no. 6, pp. 970–979, [41] P. Zhang, S. Jiang, H. Chen, M. Zhao, and M. Li, “Hydro- logical simulation capability of TRMM satellite precipitation [25] R. Xu and X. Huang, L. Li and S. Cai, A new grid-associated data in Xixian catchment, Huai River basin,” Journal of Water algorithm in the distributed hydrological model simulations,” Resources Research, vol. 6, no. 2, pp. 148–155, 2017. Science in China Series E: Technological Sciences, vol. 53, no. 1, [42] L. Liang, L. Zhao, D. Qi, C. Wang, H. Bao, and Y. Zhang, “*e pp. 235–241, 2010. experiment of hydrologic probabilistic forecast based on the [26] E. Sisay, A. Halefom, D. Khare, L. Singh, and T. Worku, precipitation forecast calibrated by bayesian model averag- “Hydrological modelling of ungauged urban watershed using ing,” Journal of Applied Meteorological Science, vol. 4, 2013. swat model,” Modeling Earth Systems and Environment, vol. 3, [43] P. Deng and L. I. Zhijia, “Comparison of three hydrological no. 2, pp. 693–702, 2017. models in flood simulation for Xixian basin of Huaihe River,” [27] M. Saber, T. Hamaguchi, T. Kojiri, K. Tanaka, and T. Sumi, “A Journal of Hohai University, vol. 41, pp. 377–382, 2013. physically based distributed hydrological model of wadi [44] Y. Y. Han and T. Cai, “*e impacts of different land use system to simulate flash floods in arid regions,” Arabian patterns on water volume in the Xixian watershed, China,” Journal of Geosciences, vol. 8, no. 1, pp. 143–160, 2015. Applied Mechanics and Materials, vol. 737, pp. 728–731, 2015. [28] E. Todini and L. Ciarapica, “*e TOPKAPI model,” in [45] P. Shi, C. Chen, R. Srinivasan et al., “Evaluating the SWAT Mathematical Models of Large Watershed Hydrology, model for hydrological modeling in the Xixian watershed and V. P. Singh, D. K. Frevert, and S. P. Meyer, Eds., pp. 471–550, a comparison with the XAJ model,” Water Resources Man- Water Resources Publications, Littleton, CO, USA, 2002. agement, vol. 25, no. 10, pp. 2595–2612, 2011. [29] L. Ciarapica and E. Todini, “TOPKAPI: a model for the [46] P. Shi, Y. Hou, Y. Xie et al., “Application of a SWATmodel for representation of the rainfall-runoff process at different hydrological modeling in the xixian watershed, China,” scales,” Hydrological Processes, vol. 16, no. 2, pp. 207–229, Journal of Hydrologic Engineering, vol. 18, no. 11, pp. 1522– 1529, 2013. [30] Z. Liu, M. L. V. Martina, and E. Todini, “Flood forecasting [47] Y. Y. Han and T. Cai, “*e impacts of land-use change using a fully distributed model: application of the TOPKAPI patterns on soil erosion in the Xixian basin, China,” Applied model to the upper Xixian catchment,” Hydrology and Earth Mechanics and Materials, vol. 737, pp. 762–765, 2015. System Sciences, vol. 9, no. 4, pp. 347–364, 2005. [48] G. M. Hornberger and R. C. Spear, “An approach to the [31] Z. Liu and T. Ezio, “Assessing the TOPKAPI non-linear preliminary analysis of environmental systems,” Journal of reservoir cascade approximation by means of a characteristic lines solution,” Hydrological Processes, vol. 19, no. 10, Environmental Management, vol. 12, no. 1, pp. 7–18, 1981. [49] R.-S. Blasone, J. A. Vrugt, H. Madsen, D. Rosbjerg, pp. 1983–2006, 2005. [32] R. J. C. Burnash, R. L. Ferral, and R. McGuire, A Generalized B. A. Robinson, and G. A. Zyvoloski, “Generalized likelihood Streamflow Simulation System; Conceptual Modeling for uncertainty estimation (GLUE) using adaptive Markov chain 16 Advances in Meteorology Monte Carlo sampling,” Advances in Water Resources, vol. 31, no. 4, pp. 630–648, 2008. [50] J. E. Nash and J. V. Sutcliffe, “River flow forecasting through conceptual models part I—A discussion of principles,” Journal of Hydrology, vol. 10, no. 3, pp. 282–290, 1970. [51] X. Jin, C.-Y. Xu, Q. Zhang, and V. P. Singh, “Parameter and modeling uncertainty simulated by glue and a formal bayesian method for a conceptual hydrological model,” Journal of Hydrology, vol. 383, no. 3-4, pp. 147–155, 2010. [52] Standards Press of China, “Standard for hydrological in- formation and hydrological forecasting (Gb/T 22482-2008),” in Ministry of Water Resources of the People’s Republic of China, Standards Press of China, Beijing, China, 2008. [53] C. Gabriele, M. Cinzia, E. Ortiz, and T. Ezio, “Application of the TOPKAPI model within the DMIP 2 Project,” in Pro- ceedings of the 23rd Conference on Hydrology, Phoenix, AZ, USA, January 2009. International Journal of The Scientific Advances in Advances in Geophysics Chemistry Scientica World Journal Public Health Hindawi Hindawi Hindawi Hindawi Publishing Corporation Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www www.hindawi.com .hindawi.com V Volume 2018 olume 2013 www.hindawi.com Volume 2018 Journal of Environmental and Public Health Advances in Meteorology Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com Applied & Environmental Journal of Soil Science Geological Research Hindawi Volume 2018 Hindawi www.hindawi.com www.hindawi.com Volume 2018 International Journal of International Journal of Agronomy Ecology International Journal of Advances in International Journal of Forestry Research Microbiology Agriculture Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 International Journal of Journal of Journal of International Journal of Biodiversity Archaea Analytical Chemistry Chemistry Marine Biology Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Advances in Meteorology Hindawi Publishing Corporation

Flood Prediction in Ungauged Basins by Physical-Based TOPKAPI Model

Download PDF
17 pages

Loading next page...
 
/lp/hindawi-publishing-corporation/flood-prediction-in-ungauged-basins-by-physical-based-topkapi-model-cYWPjdRLS0

References (59)

Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2019 Xiangyi Kong et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ISSN
1687-9309
eISSN
1687-9317
DOI
10.1155/2019/4795853
Publisher site
See Article on Publisher Site

Abstract

Hindawi Advances in Meteorology Volume 2019, Article ID 4795853, 16 pages https://doi.org/10.1155/2019/4795853 Research Article Flood Prediction in Ungauged Basins by Physical-Based TOPKAPI Model 1 1 2 Xiangyi Kong , Zhijia Li, and Zhiyu Liu College of Hydrology and Water Resources, Hohai University, 1 Xikang Road, Nanjing 210098, China Ministry of Water Resources Information Center, 2 Lane, Baiguang Road, Beijing 100053, China Correspondence should be addressed to Xiangyi Kong; kysadeur@yeah.net Received 6 March 2019; Revised 23 May 2019; Accepted 12 June 2019; Published 5 September 2019 Academic Editor: Francesco Viola Copyright © 2019 Xiangyi Kong et al. *is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Scarce historical flood data in ungauged basins make it difficult to establish empirical and conceptual model forecast in these areas. *e physical-based distributed model TOPKAPI is introduced for flood prediction in an ungauged basin by parameter transplant. Five main parameters are selected, and the sensitivity is analyzed by the GLUE method. *e Xixian basin and Huangchuan basin in the upper Huaihe basin in China are chosen as study areas. *e Xixian basin is regarded as a gauged basin for parameter calibration, and the Huangchuan basin is regarded as an ungauged basin by ignoring the historical discharge data. *e model is calibrated in gauged Xixian basin, and then parameters are directly transplanted to adjacent “ungauged” Huangchuan basin to simulate flood forecast in an ungauged basin. *e sensitivity analysis shows that soil thickness and soil saturated water content are the most sensitive parameters, and the Manning coefficient of main channel with high Strahler also significantly affects forecast results. According to the simulation results, the TOPKAPI model exhibits good performance in building and the prediction of the ungauged basin, in which the qualified rate of volume and peaks reaches 69.23%, and the average NSE criterion is over 0.67, which is acceptable forecast accuracy and has positive implication for the hydrological forecasting research. using partial missing observation data [7]. *ough the 1. Introduction world’s construction of hydrological stations continues to *e hydrological model is a vital tool for the current flood develop, there are still many small basins without moni- forecasting work by simulating the hydrological process. To toring stations; these small basins are often the focus areas of estimate the outflow process at the exit section, it generates a flood disaster research. For this reason, the flood forecasting certain flood runoff process in the basin under the cir- of ungauged basins has become the focus and hot topic of cumstance of a certain structure and parameters [1]. *e existing research, as well as a challenge for all hydrologists hydrological model basing on physical theory, according to [8]. the description method and generalization degree of the *e study on ungauged basin flood forecasting aroused hydrological process, can be split into conceptual hydro- the attention from hydrologists in the end of last century. Vandewiele et al. [9] used the method of regionalization of logical model and physical-based model [2]. *e Xin’anjiang model proposed by Zhao of Hohai University is one of the model parameters to forecast monthly discharge in unga- conceptual hydrological models [3]. It is characterized by uged region in 1991. Despite using the conventional re- less model parameters and high forecasting accuracy, and it gression equations to get model parameters, Burn and has been widely used in flood forecasting in various river Boorman [10] employed the basin classification method to basins in China [4]. However, as hydrogeological conditions find similar basins. Since usually there are insufficient his- of basins are highly generalized [5], the parameters should be torical discharge data in ungauged basins, the commonest calibrated using historical flood data [6], making it difficult approach to build the hydrological model in ungauged is to build a forecast model in the ungauged basins or basins parameter transferring [11–13], suggesting that the 2 Advances in Meteorology [32]), fully considers the spatial heterogeneity of the un- parameters required for models were calibrated in nearby basins with sufficient discharge data and then transplanted derlying surface conditions, exhibiting the advantage of relatively simple structure, clear parameter meaning, large to ungauged basins. *is approach, however, is based on the similarity of these two basins [14]. Gunter ¨ [15] summarized spatial scale elasticity, as well as wide application fields [33]. several methods for ungauged basin prediction using both Flood simulation and forecasting by this model have been event models and explicit soil moisture accounting (ESMA) widely performed in many river basins in Italy, Germany, models. To solve the description of characteristics of Spain, and other countries; good results have been achieved ungauged basins, Murugesu [16] proposed the approach of as well [30]. physical controls (basin form and function) of the in- In this study, the capability to simulate flood in unga- uged basin of TOPKAPI model is tested to give suggestions stantaneous unit hydrograph (IUH), which has been latterly achieved by Ellouze-Gargouri [17], using geomorphological on flood forecast in ungauged basins. *e sensitivity of the main parameters of the TOPKAPI model was analyzed, instantaneous unit hydrograph (GIUH) and Copulas. In 2012, Moore et al. [18] made a prediction in ungauged basins laying the theoretical basis for the parameter transplantation to build forecast model in ungauged basins. *e Xixian basin in British Columbia (BC) by transposing model parameter from adjacent basin; they also achieved good results. Patil located in the upstream of Huaihe River in China and the and Stieglitz [19], using a model developed themselves, adjacent Huangchuan basin were taken as the research analyzed selectively transferring sensitive versus insensitive basins. *e Huangchuan basin was considered the ungauged parameters on flood forecasting in ungauged basins. Athira basin by ignoring the historical flood data. Parameters were et al. [20] used SWAT model to explore the appropriate calibrated in the Xixian basin and transplanted to the ad- functional relationship between the parameters and basin jacent Huangchuan basin for flood simulation to test the application of TOPKAPI model in ungauged basins without characteristics. Waseem et al. [21] compared different in- terpolation schemes on parameter spatial distribution in parameter calibration. ungauged basins. Hyun and Choi [22] analyzed the re- lationship between flooding index and rainfall pattern to 2. The TOPKAPI Model forecast flood severity in ungauged basins. Sahoo et al. [23] adopted a novel routing method VPMD at ungauged rivers 2.1. Model Introduction. *e TOPKAPI model refers to a to plot rating curves. Canovas ´ et al. [24], using the 2D distributed hydrological model based on physical charac- hydraulic model in Spain, conducted the research on flood teristics. It splits the study basin into detailed grids, and thus discharge in ungauged basins. the difference of physical characteristic in spatial distribu- With the advancement of geographic information tion can be fully considered under the calculation unit technology, remote sensing technology, and computer sci- subdividing [34]. It can also give the specific condition of soil ence, even in the absence of hydrological observatories, the saturation, runoff yield, and confluence depth at each cal- geography of the watershed can be obtained from remote culation unit point in the basin, which is of high reference sensing images, including digital elevation model (DEM) [5] implication in real-time flood forecasting, land use and and land use and soil classification maps [25]. To fully ex- environmental impact assessment, and hydrological process ploit the advantage of such development, the physical-based simulation in ungauged basins [35]. distributed hydrological models were created [2, 26]. *ey According to this model, the hydrological process in describe the underlying surface conditions of basins using each calculation unit is generalized into three nonlinear physical parameters from remote senses, making it possible reservoir equations, which are similar in structure, repre- to simulate ungauged basin and become a new topic in the senting drainage in soil, surface runoff, and channel runoff field of hydrological forecasting [27]. Among them, the on saturated soil and impermeable surface, respectively. *e TOPKAPI rainfall-runoff model refers to a distributed hy- finite difference method is used in the calculation, so the drological model with grid-based computational unit based correlated four surrounding grids are considered during the on physical basis [28]. calculation of each grid unit [36]. *e parameters used to Professor Ezio Todini from the University of Bologna in describe the underlying surface conditions (e.g., soil Italy proposed the TOPKAPI (Topographic Kinematic thickness, vertical and horizontal saturated hydraulic con- Approximation and Integration) hydrological model in 1995 ductivity, coefficient of nonlinear reservoir equation, and [29]. *is model refers to a physics-based distributed hy- Manning coefficient of surface and evaporation coefficient of drological model based on the study of rainfall-runoff re- vegetation) could be extracted directly from the soil type lationship to explore the potential of hydrological model classification map and land surface utilization map. Besides, prediction based on physical theory in mountain flood the elevation and gradient distribution of the basin could be forecast [29]. *e model consists of several modules (e.g., reflected by DEM (digital elevation model). evapotranspiration, snowmelt, soil flow, surface runoff, river In terms of specific mechanism, TOPKAPI model em- runoff, and groundwater) [30]. In this model, the whole ploys *ornthwaite evaporation formula to calculate the basin is split into cell grids, and each grid is considered a potential evaporation of different vegetation covers in dif- single calculation unit. Each calculation unit reflects the ferent growth stages; it also calculates the actual evaporation whole physical hydrological process [31]. *is model, according to the actual wettability of upper soil [28]. compared with conventional conceptual hydrological According to the change of soil water content in the upper models (e.g., Xin’anjiang Model [3] and Sacramento Model unsaturated region, the runoff in the soil and the surface Advances in Meteorology 3 runoff are first calculated and then aggregated to the total upper reaches of the Huaihe River were taken as research runoff. In the confluence part, the nonlinear reservoir area. *e Wangjiaba basin is located in the southern part of equation is adopted to merge surface runoff and un- Henan Province of China. It is a monsoon climate with an derground runoff for confluence calculation [28, 37]. *e average annual rainfall of 1060 mm [30]. About 50% of the main parts and approaches of TOPKAPI model are listed in rainfall concentrates in the flood season, and the Xixian and Figure 1. Huangchuan river basins are the origins of the upstream To describe the movement of water in surface, soil, and river. *e control area of Xixian Hydrological Station is 10 river channel, TOPKAPI model adopts three nonlinear 190 km . *ere are two large reservoirs in Nanwan (water reservoir equations with a similar structure. In the earliest catchment area 1058 km ) and Shishankou (water catch- version, the variable step five-stage Runge–Kutta numerical ment area 306 km ) [41]. After deducting two large reser- method is employed to solve the differential equations. Later, voirs, the catchment area is 8400 km [42]. *e elevation it was found that under appropriate approximation con- range of the Xixian County is 37–963 m [43]. *e terrain in ditions, these equations could be solved using analytical the northeast is relatively flat, being mountainous in the methods [28]. west, southwest, and south. *e soil type is mostly clay soil. It has soil types (e.g., loam, sandy loam, and clay). Most of the land is farmland, with a small portion of forest land and 2.2. Comparison with Conceptual Model. From the aspect of mixed forest land. *e Huangchuan basin outlet station has data required for the model building, the conceptual hy- a controlled area of 1989 km with Pohe reservoir built drological model should be based on the previous historical inside [44]. *e elevation range of the area is 36–984 m. *e hydrological data. Taking Xin’anjiang model as an example, watershed is mostly mountainous, the north is relatively flat, the parameters, including water storage capacity (WM), most of the soil types are clay and sandy loam, and the evapotranspiration coefficient (K), outflow and regression surface vegetation is primarily farmland [45, 46]. coefficient of the basin (KI and KG), should be calibrated *e digital elevation data of the basins could be obtained through historical floods [3]. In other words, conceptual from the SRTM global 90 m digital elevation model provided model is based on high-degree generalization of those by the Consultative Group on International Agricultural physical parameters to reflect the characteristics of basins. Research (Figure 3). GIS software was employed to extract *is model has the advantage of relatively flexible adjustable the simulated stream nets and basin boundaries in the Xixian parameters that can be fully adjusted to give optimized and Huangchuan basins. *e rainfall data of 13 rainfall forecast results and avoid some relatively complex mecha- stations and the discharge data of Xixian hydrological station nisms in the hydrological process [38]. However, the dis- and that of the two reservoirs were collected from 1980 to advantage lies in the reliance on historical flood data, which 2003 according to the distribution of stations in the basin. is particularly evident in some areas in the absence of his- *e discharge data, as well as the rainfall data of 5 rainfall torical rainfall and runoff data where it is almost likely to stations in the Huangchuan basin and the discharge data of calibrate parameters [39]. *e existing feasible method is to the Huangchuan hydrological station from 2003 to 2005, search for data basins with similar geological and hydro- were processed into 1 h interval. *e meteorological data of logical conditions for parameter transplantation, which will Xinyang, Zhumadian, and Guangshan meteorological sta- inevitably face systematic error of the model. In contrast, tions (including daily maximum and minimum tempera- most of the data required for describing hydrological con- tures, wind speed, humidity, and sunshine hours) were ditions and underlying surface conditions of watershed collected from 1980 to 2005. based on physical basis can be obtained through actual measurement (e.g., DEM, soil classification and soil thick- ness, hydraulic conductivity, and evaporation coefficient of 3.2. Underlying Surface Analysis. *e underlying surface land classification) [34]. *ese parameters are relatively easy parameters of the TOPKAPI model were transplanted on the to obtain online, including the Shuttle Radar Topography premise that the soil types and land cover of underlying Mission (SRTM) launched by the Consultative Organization surface of the two basins should be of the same type or have for International Agricultural Research (CGIAR), which can similar physical properties, so the distribution of soil obtain the global 90 m resolution digital elevation model free (Figure 4 and Table 1) and surface in the two basins (Figure 5 of charge from its official website; the accessible website of and Table 2) significantly affected the ungauged area pre- the University of Maryland (UMD) in the United States. On diction [47]. According to the collected data, the Xixian and this point, the hydrological model based on physical basis Huangchuan basins were located close to each other. *eir has obvious advantages over the conventional conceptual soil types were similar to clay, and most of the land surface hydrological model when it is relatively easy to obtain the was the same type of farmland. *e soil parameters and land parameters required for model building in ungauged areas use parameters calibrated by the Xixian basin in the [28, 40]. TOPKAPI model were applicable to the adjacent Huang- chuan basin. *e soil data are derived from the FAO soil map of the world, Global soil profile databases, which is 3. Study Area and Dataset available online at http://www.ngdc.noaa.gov/seg/eco/ cdroms/reynolds.htm. *e land cover map is derived 3.1. Study Area. *e neighboring Xixian and Huangchuan from National Administration of Surveying Mapping and river basins (Figure 2) above the Wangjiaba basin in the 4 Advances in Meteorology Meterological data Digital elevation Soil type map Land cover map (precipitation and temperature) model Spatial interpolation Calculation grid Evapotranspiration (Thornthwaite equation) Excess Percolation Soil water Surface water storage Subsurface flow Surface flow River channel Muskingum–cunge routing Channel outflow Figure 1: Main components and mechanism of TOPKAPI model. 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E ZhuMaDian Xixian NanWan HuangChuan XinYang ShiShanKou GuangShan PoHe 60 Kilometers 0 15 30 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E ◊ Meteorological stations Xixian basin Hydrological stations Reservoir Huangchuan basin (a) Figure 2: Continued. 31°30′0″N 32°0'0″N 32°30′0″N 33°0′0″N 31°30′0″N 32°0′0″N 32°30′0″N 33°0′0″N Advances in Meteorology 5 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E HuangGang XiaoCaoDiao DaPoLing TongBai HuJiaWan ChangTaiGuan LuoShan WuLiDian NanWan HuangChuan NanLiDian GuangShan PengXinDian XinDian DingYuanDian PoHe WuChenHe XinXian 0 15 30 60 Kilometers 113°30′0″E 114°0′0″E 114°30′0″E 115°0′0″E 115°30′0″E Precipitation gauges Xixian basin Huangchuan basin Reservoir (b) Figure 2: Location of the study area Xixian and Huangchuan basins and the distribution of reservoirs and (a) meteorological and hy- drological stations and (b) precipitation gauges. Elevation (m) Figure 3: ƒe digital elevation model of study area. 31°30′0″N 32°0′0″N 32°30′0″N 33°0′0″N 32°0′0″N 32°30′0″N 33°0′0″N 6 Advances in Meteorology 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E N N 0 15 30 60 kilometers 0 5 10 20 kilometers 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E Sand clay loam Sand clay loam Loam Loam Clay Clay (a) (b) Figure 4: Soil types map of (a) Xixian basin and (b) Huangchuan basin. Table 1: Component of soil types in Xixian and Huangchuan basin. Soil type Code Percentage (%) Loam 3085 15.65 Xixian basin Sandy loam 3963 12.87 Clay 4326 71.48 Loam 3085 14.14 Huangchuan basin Sandy loam 3963 32.12 Clay 4326 53.74 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E N N 0 15 30 60 kilometers 0 5 10 20 kilometers 113°30′0″E 114°0′0″E 114°30′0″E 114°30′0″E 115°0′0″E Farmland Forest Farmland Forest Grass Mixed forest Grass Mixed forest (a) (b) Figure 5: Land use map of (a) Xixian basin and (b) Huangchuan basin. 31°30′0″N 32°0′0″N 32°30′0″N 31°30′0″N 32°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N 32°30′0″N 31°30′0″N 32°0′0″N 32°30′0″N 31°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N 31°30′0″N 32°0′0″N Advances in Meteorology 7 Table 2: Component of land use in Xixian and Huangchuan basins. Land use type Code Percentage (%) Farmland 6 67.70 Forest 8 18.41 Xixian basin Mixed forest 10 2.81 Grass 11 11.09 Farmland 6 69.92 Forest 8 10.55 Huangchuan basin Mixed forest 10 5.03 Grass 11 14.50 Geoinformation of China (NASG), which is free down- a � r + Q , c c loadable for all Internet at http://www.globallandcover.com/ √�� 2/3 GLC30Download/index.aspx and the characteristic data for s (sin c) b � , each type of land cover are derived from Corinne Land 1/3 (4) 2/3 4/3 2 n (tan c) X Cover 2006 raster data by European Environment Agency, which is accessible at https://www.eea.europa.eu/data-and- maps/data/clc-2006-raster. c � , where r denotes the lateral inflow, including soil outflow to 4. Parameter Sensitivity Analysis surface and channel. *e calculation formulas show that the main influence in 4.1. Parameter Selection. In TOPKAPI model, the nonlinear the calculation of soil water originates from the upstream reservoir equation can be written as inflow, grid size, and hydraulic conductivity, in which hy- dy draulic conductivity as the unknown coefficient of the (1) � a − by , nonlinear equation will directly affect the solution of the dt unknown term. *us, hydraulic conductivity is considered where the variable y denotes average soil water content, river the key parameter to be investigated. *e calculation of storage, surface water depth, etc. Variables a, b, and c keep surface water on slope is affected by slope, precipitation and constant in a single time step. surface Manning coefficient, and surface runoff after In the calculation of groundwater, the variables a, b, and deducting infiltration is associated with soil conditions on c are expressed as follows: underlying surface. Accordingly, in the part of surface flow 2 u u calculation, the Manning coefficient of surface, the saturated pX + Q + Q 0 s a � , water content associated with soil characteristics, and the soil thickness describing soil volume were taken as the key (2) research parameters. In the calculation of channel flow using b � , nonlinear reservoir method, the Manning coefficient of river roughness was taken as the main calibration parameter. All c � a , the selected parameters are listed in Table 3. where p denotes the precipitation, X is the size of grid unit, u u Q and Q refer to surface and subsurface inflow of a grid 0 s 4.2. Sensitivity Analysis. *e GLUE method was first pro- unit, respectively, and C is the local hydraulic conductivity. posed by Beven and Binley in 1992 based on the RSA (re- In the calculation of surface water, the variables a, b, and gionalized sensitivity analysis) method proposed by c are expressed as follows: Hornberger and Spear in 1981 [48]. Beven and Binley 1 V proposed that this method is not the numerical value of a exf a � , single parameter, but the value of a group of parameters that XW dt significantly affect the prediction results of the model. In this 1/2 method, Monte Carlo method is adopted to sample a series tan (β) (3) b � , of parameter combinations in prior parameter range and n X substitute them into the model calculation. Likelihood function is selected as the evaluation and instruction of the c � a , parameter sets. *e parameter set whose likelihood function where W denotes the width of grid without river channel, value is higher than a certain threshold is considered to be V represents the net precipitation, and a is index in able to represent the characteristic of study basin (which is exf 0 Manning equation with a constant value of 5/3. called “behavioral”), whereas the parameter set with likeli- In the calculation of channel water propagation, the hood function value lower than the threshold is considered variables a, b, and c are expressed as follows: fail at reflecting it and be abandoned [49]. 8 Advances in Meteorology Table 3: Main parameters of the TOPKAPI model. Modules Parameter Description Groundwater Horizontal hydraulic conductivity Reflecting water conduction rate in soil Calculating factors of water conduction rate on the Surface Manning coefficient surface When the soil water content reaches the saturated Saturated water content Surface runoff state, the volume percentage of water content *e soil depth between soil and bedrock reflects the Soil thickness total volume of soil on the underlying surface in the calculation unit Channel runoff Riverbed Manning coefficient Calculation factors of water conduction rate in rivers Several criterions could be used to describe the pre- saturated water content, and riverbed Manning coefficient. diction performance of a model, among which the Nash– Soil type with high coverage percentage such as clay (with Sutcliffe model efficiency coefficient (NSE) [50] has been 71.48% coverage) has significantly high standard deviation most widely used in model prediction performance judg- value for both thickness and saturated water content. *is is ment. In this study, the NSE is chosen as the likelihood the same for the land use Manning coefficient of farmland function. *e Nash–Sutcliffe model efficiency coefficient is (with 67.70% coverage) and the riverbed Manning co- efficient of V level river channel. expressed as T 2 t t 􏽐 Q − Q 􏼁 t�1 m o NSE � 1 − , (5) 5. Application in Ungauged Basins 􏽐 Q − Q 􏼁 t�1 o To verify the effectiveness of the model in ungauged basins, t t where Q and Q denote the modeled and observed dis- m o experiments were performed. *e Xixian basin and adjacent charge at time t, respectively, and Q is the average value of Huangchuan basin in Wangjiaba were taken as study areas. observed discharge over the whole time series. *e parameters of the model were calibrated in Xixian basin In this study, 3000 sets of parameters including those and subsequently transferred to Huangchuan basin, and selected in the previous part were generated by random flood simulations were carried out to verify the application Monte Carlo sampling methodology within the prior range effect of the model in ungauged basins. in Table 4. A threshold of 0.7 NSE is chosen to check whether *e calibration and validation results are evaluated by these parameter sets are behavioral. Parameter uncertainty “qualified rates” [52]. A flood forecast with flood volume and was analyzed by 8 floods in Xixian basin from 1980 to 2002 flood peak error less than 20% is considered to be “qualified,” in TOPKAPI model. A total of 283 sets of parameters were and the “qualified rate” means the percentage of qualified found to be behavioral achieving 0.7 NSE likelihood. flood simulations over the total number. *e whole forecast Posterior parameter distribution was derived from 283 scenario accuracy will be evaluated to a certain level behavioral parameter sets. To get visual impression of these according to Table 6. parameter sets, the histograms of all 18 parameters distri- bution are plotted in Figure 6. Histograms with sharp peaks or steep slopes means parameter are well identified, whereas 5.1. Parameter Calibration. Most parameters in physical- parameters with relative flat histograms are associated with based hydrological model could be obtained by actual more uncertainty [51]. Because all the behavioral sets have measurement; however, the measured values acquired on the performance over threshold in common, narrow and sharp point measurement might not be sufficiently representative peaks reflect relative strong relationship with model results at the calculation unit scale and might not reflect the and show that this parameter has more sensitivity, which temporal and spatial scale change within the unit. *us, the means variation in this parameter may significantly affect parameters obtained by measurement or reference to the model results. A statistic of parameter frequency was con- literature should be fine-tuned by trial before the simulation ducted in Table 5 by calculating standard deviation of fre- to compensate for the error caused by the spatial general- quencies of each parameter to get a numerical comparison. ization of the parameters and the time generalization of the In Table 5, the standard deviations of parameters fre- input parameters, making the results of the simulation closer quency distribution are listed. It can be clearly seen that to the actual situation. horizontal hydraulic conductivity has the least values and *e calibration parameters included soil horizontal the histograms are all relatively flat, meaning that this pa- water conductivity, saturated water content, land surface, rameter has low sensitivity and high uncertainty. *is sort of and riverbed Manning coefficient for Muskingum algorithm. parameters are difficult to be calibrated due to their weak In the previous GLUE assessment, 3000 sets of parameters relationship with model outputs, but they can only slightly have been generated. Several “best” performed sets of pa- affect the model performance. However, emphasis in cali- rameters with reasonable values and high NSE are chosen for bration should be put on those parameters with high manually adjustment. *ese parameters were calibrated by standard deviation and sensitivity such as soil thickness, the rainfall and discharge data of 16 floods from 1991 to 2003 θ Advances in Meteorology 9 Table 4: Prior range of parameters. Loam Sandy loam Clay Soil thickness (m) 1.3–1.8 0.4–1.1 0.8–1.6 − 6 Horizontal hydraulic conductivity (10 m/s) 2.26–9.05 0.40–1.62 0.17–0.71 Saturated water content 0.35–0.50 0.27–0.33 0.30–0.55 Farmland Forest Mixed forest Grass Surface Manning coefficient 0.06–0.25 0.20–0.35 0.12–0.27 0.06–0.25 I II III IV V Riverbed Manning coefficient 0.05–0.30 0.02–0.15 0.02–0.15 0.02–0.15 0.02–0.15 15 15 10 10 5 5 0 0 0 1.5 1.6 1.7 0.5 0.6 0.7 0.8 0.9 1 1.2 1.4 1.6 L (m) L (m) L (m) 5 5 0 0 0 48 6 0.5 1 1.5 2 34567 –6 –6 –7 k (m/s) ×10 k (m/s) k (m/s) ×10 ×10 sh sh sh 0 0 0 0.38 0.4 0.42 0.44 0.46 0.48 0.28 0.29 0.3 0.31 0.32 0.35 0.4 0.45 0.5 Loam Sandy loam Clay (a) 15 15 10 10 0 0 0 0 0.1 0.15 0.2 0.25 0.22 0.24 0.26 0.28 0.3 0.32 0.34 0.15 0.2 0.25 0.1 0.15 0.2 0.25 Farmland Forest Mixed forest Grass (b) Figure 6: Continued. Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) 10 Advances in Meteorology 0 0 0 0 0.1 0.15 0.2 0.25 0.04 0.06 0.08 0.1 0.12 0.05 0.1 0.15 0.05 0.1 0.15 0.02 0.04 0.06 0.08 0.1 0.12 I II III IV V (c) Figure 6: Frequency histograms of (a) soil thickness L, horizontal hydraulic conductivity (k ), and saturated water content of three types of sh soil, (b) surface Manning coefficient of fore types of land use, and (c) riverbed Manning coefficient of five river classifications. Table 5: Standard deviation of parameter frequency distribution. Loam Sandy loam Clay Soil thickness (m) 2.5123 3.2393 6.0154 − 6 Horizontal hydraulic conductivity (10 m/s) 0.7734 0.7244 0.5807 Saturated water content 2.7310 2.0651 5.4510 Farmland Forest Mixed forest Grass Surface Manning coefficient 3.7226 0.6719 0.7198 1.7073 I II III IV V Riverbed Manning coefficient 2.0449 1.4837 2.4079 4.5080 8.5721 applicability in this basin. According to the simulated flood Table 6: Forecast scenario classification level. volume, the qualified rate of TOPKAPI model flood volume Forecast level A B C was 81.3% in the 16 floods used for calibration and 93.8% in Qualified rate the flood peak simulation. From the results, it is revealed that QR> 85.0 85> QR> 70.0 70.0> QR> 60.0 (%) under the underlying elevation data, soil and land use NSE NSE> 0.90 0.90> NSE> 0.70 0.70> NSE> 0.60 distribution parameters and basin hydrometeorological data meet the demand, the model achieved relatively good simulation in the study area. *e average NSE were all above in Xixian basin. Run the TOPKAPI model with the mete- 0.7, achieving the accuracy of Class B forecasting [52], and orological data (including precipitation and temperature) of the average error between flood volume and flood peak was 13 floods in Huangchuan watershed from 2003 to 2005 controlled within 20%. *is is associated with the wet area together with the calibrated parameters and adjusted these where the research basin is located. *e full storage and parameter according to the outcome flood results. Fur- runoff yield models used in the model were applicable to the thermore, to verify the accuracy of the simulation and to wet area of Huaihe River, and the better the hydrogeological demonstrate the application of the model in ungauged conditions of the basin, the smaller the error of the model basins, the flood simulation results were compared with the confluence calculation will be. *us, TOPKAPI model ex- measured results. hibits good applicability in this basin. All the calibrated *e flood peak and NSE were taken as objective func- parameters are listed in Tables 8–10. tions to calibrate the parameters. *e percentage of initial From the process of model building, TOPKAPI model soil water content in the basin would significantly affect the did not adjust too much parameters in the calibration simulation results of the early floods in the basin. In this process, and the prediction results were good. Accordingly, calibration, to determine the initial soil moisture, the TOPKAPI model based on physical basis exhibited better “warm-up” method [53] was adopted to simulate the initial applicability when building model in adjacent areas without state through natural situation, which was initiated 30 days historical flood data. However, the prediction accuracy of before the first flood in advance. TOPKAPI model depended on the high-precision partition Figure 7 shows that, in the calibration, most of the of computational grids, whereas high-resolution grids would stream flow series are well reproduced. Floods with large increase the requirement of computational ability. *us, to water volume are simulated better than those with low obtain higher prediction accuracy, the model requires better volume. From the results of model calibration in Table 7, the operation equipment and long calculation time. simulation results of the model in Xixian basin were good, and the average NSE were above 0.8. Among the 16 floods used in the calibration, 14 floods had NSE over 0.8, taking up 5.2. Parameter Validation. *e parameters obtained by the 87.5% and 6 floods over 0.9. In terms of Nash–Sutcliffe calibration in adjacent basin were directly applied in model efficiency coefficient, the model exhibited certain Huangchuan basin. Besides, the flood discharge process of Frequency (%) Frequency (%) Frequency (%) Frequency (%) Frequency (%) Advances in Meteorology 11 1000 2000 3000 4000 5000 6000 7000 8000 Simulation steps (1 h ) Observation Simulation Precipitation Figure 7: Calibration hydrograph from 1991 to 2003. Table 7: Calibration results in Xixian basin. Flood volume Flood peak Flood no. NSE 4 3 3 Observation (10 m ) Simulation error percentage (%) Observation (m /s) Simulation error percentage (%) 10523199 39933.54 8.87 1300 1.77 0.76 10530199 46156.68 3.72 1670 − 2.86 0.89 12106199 110651.16 − 24.55 5060 − 28.19 0.83 10629199 73803.71 15.33 2960 − 1.97 0.91 10805199 92912.93 − 10.04 4420 − 3.97 0.83 10021992 19089.25 − 7.37 579 12.61 0.84 18199305 14123.72 3.36 516 3.86 0.7 19950707 42848.26 − 1.2 2300 2.68 0.78 17199607 81739.98 − 22.99 4450 − 19.46 0.89 19960802 25732.97 − 5.79 875 1.87 0.71 19970629 24946.02 2.81 1220 12.41 0.91 19980630 54968.35 16.86 2510 − 12.89 0.94 20020622 97511.4 14.1 5080 − 0.62 0.9 20020627 61223.48 1.34 2820 − 8.32 0.85 20020723 55287.15 13.27 2790 − 6.49 0.84 20030629 107940.6 6.67 3900 − 13.52 0.77 Abs. average 8.64 8.34 Average 2.15 − 3.94 0.84 Qualified rate 93.75% 93.75% 100.00% Table 8: Calibrated soil parameters. Table 10: Calibrated channel Manning coefficient. Loam Sandy loam Clay Order I Order II Order III Order IV Order V L (m) 1.7675 0.78856 1.4715 N 0.15918 0.085962 0.058017 0.046145 0.03347 Channel k (m/s) 4.55e‒ 6 8.4e‒ 7 4.13e‒ 7 ∗ sh N : channel Manning coefficient. Channel θ 0.48511 0.307 0.49886 L: soil thickness; k : horizontal hydraulic conductivity; θ : saturated water sh s Table 11, Figure 8, and the box plot Figure 9 of calculation content. and results, it is revealed that the overall NSE was relatively lower because of parameter transplantation. *e average Table 9: Calibrated surface Manning coefficient. NSE was 0.67, and the qualified rate of flood volume and peak simulation was 69.23%. *is was probably because Farmland Forest Mixed forest Grass Xixian is in the warm temperate zone and belongs to the N 0.182871 0.274005 0.239738 0.11128 Surf semihumid basin, and the overall basin is dominated by N : surface Manning coefficient. Surf excess of storage mode, which is consistent with the basic assumption of TOPKAPI model runoff generation. *e the outlet section was calculated using the rainfall data of 13 accuracy forecast in ungauged basins, compared with the floods in 2003–2005 as input to simulate the application of calibration results, decreased in varying degrees. *is was the model in the ungauged basin. From the statistical because the parameter transplantation inevitably brings Discharge (m /s) Precipitation (mm) 12 Advances in Meteorology Table 11: Application results in ungauged basin. Flood volume Flood peak Flood no. NSE 4 3 4 3 Observation (10 m ) Simulation error percentage (%) Observation (10 m ) Simulation error percentage (%) 20030622 6872.8 41.84 705 30.78 0.65 20030626 4765.9 38.09 380 54.42 0.8 20030629 28032.16 − 7.07 2180 2.48 0.83 20030707 39802.79 − 25.56 1750 − 5.62 0.79 19200307 6711.57 − 34.86 419 − 18.61 0.48 15200308 4060.62 − 13.1 173 39.97 0.63 20040716 15826.68 2.48 1070 11.64 0.72 20040801 11528.6 5.97 825 5.6 0.67 20040813 14726.48 − 0.84 1200 28.18 0.85 20050726 9259.52 − 16.38 600 − 5.13 0.62 20050820 8599 − 14.45 402 19.06 0.5 20050828 9145.8 − 7.8 733 − 6.9 0.59 20050902 11192.11 − 17.65 666 2.81 0.57 Abs. average 17.39 17.78 Average − 3.79 12.21 0.67 Qualified rate 69.23% 69.23% 38.46% 2500 0 100 200 300 400 500 600 700 800 900 1000 1100 Simulation steps Observation Simulation Precipitation (a) 2000 0 100 200 300 400 500 600 700 800 900 1000 1100 Simulation steps Observation Simulation Precipitation (b) Figure 8: Continued. 3 3 Discharge (m /s) Discharge (m /s) Precipitation (mm) Precipitation (mm) Advances in Meteorology 13 1000 0 100 200 300 400 500 600 700 800 900 1000 Simulation steps Observation Simulation Precipitation (c) Figure 8: Hydrograph of simulation in ungauged basins in 2003, 2004, and 2005. –40 (a) (b) (c) (d) Figure 9: Box plot error percentage of (a) calibration flood volume, (b) calibration flood peak, (c) ungauged flood volume, and (d) ungauged flood peak. about inconsistencies with the actual conditions, which is analysis and comparison of calibration simulation results also the most important problem facing the prediction of with measured data and prediction results of the model in ungauged basins. adjacent similar ungauged basins, parameter transplantation It was found that the flood process of TOPKAPI model was tested. Based on the results, the following conclusions had steep rise and fall just, as shown in the hydrographs. were drawn: Accordingly, following the flood peak priority calibration (1) *e description of basin characteristics in TOPKAPI strategy, the systematic error of overall smaller flood volume model can be characterized by underlying surface appeared. *e water confluence process was faster than the data from land use and soil type obtained using actual situation, which is also one of the parts of the model to remote sensing technology, which is easy to obtain be further optimized. Nevertheless, TOPKAPI model still and use. In this study, it was assumed that reflects good transplantability in flood peak prediction in Huangchuan basin is an ungauged basin without any ungauged basins. Since remote sensing data could reflect hydrological stations, whereas underlying surface actual topography, vegetation, and soil types as much as data can still be acquired from satellite or remote possible, the eligibility rate of flood peak prediction reached sensing equipment, making it likely to describe the about 70%. *is is the advantage brought by detailed hy- basin in hydrological model. drological process based on the physical basis of the dis- (2) According to the results of the sensitivity analysis of tributed model, fully displaying the physical conditions of the model parameters, the soil thickness and surface underlying surface. Manning coefficient, with a large proportion of coverage, significantly affected the prediction results 6. Discussion and Conclusion of the model. *e horizontal hydraulic conductivity In this study, a physically based distributed hydrological has the least sensitivity. Since the classification of model TOPKAPI was applied in upper Huaihe, Xixian, and river depends on the area of the whole river basin adjacent Huangchuan basin. *e model uncertainty caused and other factors, some differences would exist by spatial and temporal generalization of model parameters among different river basins for the same order. In was evaluated using GLUE method. *rough uncertainty this study, this is also a reason for the reduction of Error percentage (%) Discharge (m /s) Precipitation (mm) 14 Advances in Meteorology prediction accuracy of ungauged river basins. *e References rules of parameter transplantation in prediction [1] R. Johnston and V. Smakhtin, “Hydrological modeling of research should be elucidated further. large river basins: how much is enough?,” Water Resources (3) *e model exhibits good applicability in the upper Management, vol. 28, no. 10, pp. 2695–2730, 2014. Huaihe basin. *e base assumptions of TOPKAPI [2] M. Cai, S. Yang, H. Zeng, C. Zhao, and S. Wang, “A dis- model made it feasible in humid semihumid area, tributed hydrological model driven by multi-source spatial and thus the flood volume flood peak as long as NSE data and its application in the Ili river basin of Central Asia,” all achieved high accuracy in calibration. Water Resources Management, vol. 28, no. 10, pp. 2851–2866, (4) In the prediction of ungauged basins, it was found [3] R. J. Zhao and X. R. Liu, “*e Xinanjiang model,” Computer that under premise of sufficient remote sensing data Models of Watershed Hydrology, vol. 135, pp. 371–381, 1995. of underlying surface, the transplantation of pa- [4] Z. Ren-Jun, “*e Xinanjiang model applied in China,” Journal rameters among similar basins could achieve good of Hydrology, vol. 135, no. 1–4, pp. 371–381, 1992. forecasting results. [5] A. H. A. Suliman, M. Jajarmizadeh, S. Harun, and I. Z. Mat Darus, “Comparison of semi-distributed, GIS-based hydro- Besides, the high prediction accuracy of TOPKAPI logical models for the prediction of streamflow in a large model should be based on detailed meshing of computing catchment,” Water Resources Management, vol. 29, no. 9, units. For instance, the grid resolution adopted in this study pp. 3095–3110, 2015. was 200 m, and the prediction results were relatively good. [6] A. Kushwaha and M. K. Jain, “Hydrological simulation in a However, the calculation period took a long time. If the forest dominated watershed in Himalayan Region using model is applied in a larger watershed, the problem will be SWAT model,” Water Resources Management, vol. 27, no. 8, more prominent. *us, how to make an optimal trade-off pp. 3005–3023, 2013. between the calculation time of the model and the prediction [7] A. C. Guzha and T. B. Hardy, “Application of the distributed accuracy needs further studies. hydrological model, TOPNET, to the big darby creek wa- tershed, Ohio, USA,” Water Resources Management, vol. 24, no. 5, pp. 979–1003, 2010. Data Availability [8] W. *orsten and M. Alberto, “Convergence of approaches *e digital elevation model data can be acquired from toward reducing uncertainty in predictions in ungauged basins,” Water Resources Research, vol. 47, no. 6, pp. 453–460, CGIAR (http://srtm.csi.cgiar.org/). *e soil data are derived from the FAO soil map of the world, Global soil profile [9] G. L. Vandewiele, C.-Y. Xu, and W. Huybrechts, “Region- databases, which is available online at http://www.ngdc. alisation of physically-based water balance models in Belgium. noaa.gov/seg/eco/cdroms/reynolds.htm. *e land cover Application to ungauged catchments,” Water Resources map is derived from National Administration of Surveying Management, vol. 5, no. 3-4, pp. 199–208, 1991. Mapping and Geoinformation of China (NASG), which is [10] D. H. Burn and D. B. Boorman, “Estimation of hydrological freely downloadable for all Internet at http://www. parameters at ungauged catchments,” Journal of Hydrology, globallandcover.com/GLC30Download/index.aspx and the vol. 143, no. 3-4, pp. 429–454, 1993. characteristic data for each type of land cover are derived [11] L. Oudin, V. Andreassian, ´ C. Perrin, C. Michel, and from Corine Land Cover 2006 raster data by European N. L. Moine, “Spatial proximity, physical similarity, regression Environment Agency, which is accessible at https://www. and ungaged catchments: a comparison of regionalization eea.europa.eu/data-and-maps/data/clc-2006-raster. approaches based on 913 French catchments,” Water Re- sources Research, vol. 44, no. 3, pp. 893–897, 2008. [12] L. Oudin, A. Kay, V. Andreassian, ´ and C. Perrin, “Are Conflicts of Interest seemingly physically similar catchments truly hydrologically similar?,” Water Resources Research, vol. 46, no. 11, 2010. *e authors declare that they have no conflicts of interest. [13] R. Singh, S. A. Archfield, and T. Wagener, “Identifying dominant controls on hydrologic parameter transfer from gauged to ungauged catchments—a comparative hydrology Acknowledgments approach,” Journal of Hydrology, vol. 517, pp. 985–996, 2014. *is work was supported by the National Key Research and [14] S. A. Archfield and R. M. Vogel, “Map correlation method: Development Program of China (grant no. 2018YFC1508103) selection of a reference streamgage to estimate daily streamflow at ungaged catchments,” Water Resources Re- and the Fundamental Research Funds for the Central Uni- search, vol. 46, no. 10, pp. 5613–5618, 2010. versities (grant no. 2016B04714). [15] B. Gu¨nter, “Rainfall-runoff modeling of ungauged catchments,” in Encyclopedia of Hydrological Sciences, M. G. Anderson and Supplementary Materials J. J. McDonnell, Eds., John Wiley & Sons, Ltd., Hoboken, NJ, USA, 2006. *e following data are available in supplementary materials: [16] S. Murugesu, “Prediction in ungauged basins: a grand chal- the precipitation data, temperature, and discharge data used lenge for theoretical hydrology,” Hydrological Processes, for parameter calibration in Xixian basin and the coordinate vol. 17, no. 15, pp. 3163–3170, 2010. of meteorological stations and precipitation gauges used in [17] E. Ellouze-Gargouri and Z. Bargaoui, “Runoff estimation for an this research. (Supplementary Materials) ungauged catchment using geomorphological instantaneous Advances in Meteorology 15 unit hydrograph (GIUH) and copulas,” Water Resources Digital Computers, U.S. Department of Commerce, National Management, vol. 26, no. 6, pp. 1615–1638, 2012. Weather Service, and State of California, Department of [18] R. D. D. Moore, J. W. Trubilowicz, and J. M. Trubilowicz, Water Resources, Sacramento, CA, USA, 1973. “Prediction of streamflow regime and annual runoff for [33] P. Deng, Z. Li, and Z. Liu, “Numerical algorithm of distributed ungauged basins using a distributed monthly water balance TOPKAPI model and its application,” Water Science & En- model,” Jawra Journal of the American Water Resources As- gineering, vol. 1, no. 4, pp. 14–21, 2008. sociation, vol. 48, no. 1, pp. 32–42, 2012. [34] Z. Liu, “Application of GIS-based distributed hydrological [19] S. Patil and M. Stieglitz, “Modelling daily streamflow at model to flood forecasting,” Journal of Hydraulic Engineering, ungauged catchments: what information is necessary?,” Hy- vol. 35, pp. 70–75, 2004. drological Processes, vol. 28, no. 3, pp. 1159–1169, 2014. [35] G. Coccia and E. Todini, “Recent developments in predictive [20] P. Athira, K. P. Sudheer, R. Cibin, and I. Chaubey, “Pre- uncertainty assessment based on the model conditional dictions in ungauged basins: an approach for regionalization processor approach,” Hydrology and Earth System Sciences, of hydrological models considering the probability distribu- vol. 15, no. 10, pp. 3253–3274, 2011. tion of model parameters,” Stochastic Environmental Research [36] S. Sinclair and G. G. S. Pegram, “A comparison of ASCATand and Risk Assessment, vol. 30, no. 4, pp. 1131–1149, 2016. modelled soil moisture over South Africa, using TOPKAPI in [21] M. Waseem, J.-y. Shin, and T.-W. Kim, “Comparing spatial land surface mode,” Hydrology and Earth System Sciences, interpolation schemes for constructing a flow duration curve vol. 14, no. 4, pp. 613–626, 2010. in an ungauged basin,” Water Resources Management, vol. 29, [37] L. I. Zhijia, X. Wang, L. Yanxiang, L. Chen, and L. I. Lanru, no. 7, pp. 2249–2265, 2015. “Application of TOPKAPI model and comparison with [22] A. J. Hyun and H. I. Choi, “A new flood index for use in Xin’anjiang model,” Water Power, vol. 11, 2013. evaluation of local flood severity: a case study of small [38] S. Yang, G. Dong, D. Zheng, H. Xiao, Y. Gao, and Y. Lang, ungauged catchments in Korea,” Jawra Journal of the “Coupling Xinanjiang model and SWAT to simulate agri- American Water Resources Association, vol. 49, no. 1, pp. 1–14, cultural non-point source pollution in Songtao watershed of Hainan, China,” Ecological Modelling, vol. 222, no. 20–22, [23] B. Sahoo, M. Perumal, T. Moramarco, and S. Barbetta, “Rating pp. 3701–3717, 2011. curve development at ungauged river sites using variable [39] H. Lu, ¨ T. Hou, R. Horton et al., “*e streamflow estimation parameter muskingum discharge routing method,” Water using the Xinanjiang rainfall runoff model and dual state- Resources Management, vol. 28, no. 11, pp. 3783–3800, 2014. parameter estimation method,” Journal of Hydrology, vol. 480, ´ ´ [24] J. A. B. Canovas, M. Eguibar, J. M. Bodoque, A. Dıez-Herrero, pp. 102–114, 2013. M. Stoffel, and I. Gutierrez-P ´ erez, ´ “Estimating flash flood [40] S. Sinclair and G. G. S. Pegram, “A sensitivity assessment of discharge in an ungauged mountain catchment with 2d hy- the TOPKAPI model with an added infiltration module,” draulic models and dendrogeomorphic paleostage in- Journal of Hydrology, vol. 479, pp. 100–112, 2013. dicators,” Hydrological Processes, vol. 25, no. 6, pp. 970–979, [41] P. Zhang, S. Jiang, H. Chen, M. Zhao, and M. Li, “Hydro- logical simulation capability of TRMM satellite precipitation [25] R. Xu and X. Huang, L. Li and S. Cai, A new grid-associated data in Xixian catchment, Huai River basin,” Journal of Water algorithm in the distributed hydrological model simulations,” Resources Research, vol. 6, no. 2, pp. 148–155, 2017. Science in China Series E: Technological Sciences, vol. 53, no. 1, [42] L. Liang, L. Zhao, D. Qi, C. Wang, H. Bao, and Y. Zhang, “*e pp. 235–241, 2010. experiment of hydrologic probabilistic forecast based on the [26] E. Sisay, A. Halefom, D. Khare, L. Singh, and T. Worku, precipitation forecast calibrated by bayesian model averag- “Hydrological modelling of ungauged urban watershed using ing,” Journal of Applied Meteorological Science, vol. 4, 2013. swat model,” Modeling Earth Systems and Environment, vol. 3, [43] P. Deng and L. I. Zhijia, “Comparison of three hydrological no. 2, pp. 693–702, 2017. models in flood simulation for Xixian basin of Huaihe River,” [27] M. Saber, T. Hamaguchi, T. Kojiri, K. Tanaka, and T. Sumi, “A Journal of Hohai University, vol. 41, pp. 377–382, 2013. physically based distributed hydrological model of wadi [44] Y. Y. Han and T. Cai, “*e impacts of different land use system to simulate flash floods in arid regions,” Arabian patterns on water volume in the Xixian watershed, China,” Journal of Geosciences, vol. 8, no. 1, pp. 143–160, 2015. Applied Mechanics and Materials, vol. 737, pp. 728–731, 2015. [28] E. Todini and L. Ciarapica, “*e TOPKAPI model,” in [45] P. Shi, C. Chen, R. Srinivasan et al., “Evaluating the SWAT Mathematical Models of Large Watershed Hydrology, model for hydrological modeling in the Xixian watershed and V. P. Singh, D. K. Frevert, and S. P. Meyer, Eds., pp. 471–550, a comparison with the XAJ model,” Water Resources Man- Water Resources Publications, Littleton, CO, USA, 2002. agement, vol. 25, no. 10, pp. 2595–2612, 2011. [29] L. Ciarapica and E. Todini, “TOPKAPI: a model for the [46] P. Shi, Y. Hou, Y. Xie et al., “Application of a SWATmodel for representation of the rainfall-runoff process at different hydrological modeling in the xixian watershed, China,” scales,” Hydrological Processes, vol. 16, no. 2, pp. 207–229, Journal of Hydrologic Engineering, vol. 18, no. 11, pp. 1522– 1529, 2013. [30] Z. Liu, M. L. V. Martina, and E. Todini, “Flood forecasting [47] Y. Y. Han and T. Cai, “*e impacts of land-use change using a fully distributed model: application of the TOPKAPI patterns on soil erosion in the Xixian basin, China,” Applied model to the upper Xixian catchment,” Hydrology and Earth Mechanics and Materials, vol. 737, pp. 762–765, 2015. System Sciences, vol. 9, no. 4, pp. 347–364, 2005. [48] G. M. Hornberger and R. C. Spear, “An approach to the [31] Z. Liu and T. Ezio, “Assessing the TOPKAPI non-linear preliminary analysis of environmental systems,” Journal of reservoir cascade approximation by means of a characteristic lines solution,” Hydrological Processes, vol. 19, no. 10, Environmental Management, vol. 12, no. 1, pp. 7–18, 1981. [49] R.-S. Blasone, J. A. Vrugt, H. Madsen, D. Rosbjerg, pp. 1983–2006, 2005. [32] R. J. C. Burnash, R. L. Ferral, and R. McGuire, A Generalized B. A. Robinson, and G. A. Zyvoloski, “Generalized likelihood Streamflow Simulation System; Conceptual Modeling for uncertainty estimation (GLUE) using adaptive Markov chain 16 Advances in Meteorology Monte Carlo sampling,” Advances in Water Resources, vol. 31, no. 4, pp. 630–648, 2008. [50] J. E. Nash and J. V. Sutcliffe, “River flow forecasting through conceptual models part I—A discussion of principles,” Journal of Hydrology, vol. 10, no. 3, pp. 282–290, 1970. [51] X. Jin, C.-Y. Xu, Q. Zhang, and V. P. Singh, “Parameter and modeling uncertainty simulated by glue and a formal bayesian method for a conceptual hydrological model,” Journal of Hydrology, vol. 383, no. 3-4, pp. 147–155, 2010. [52] Standards Press of China, “Standard for hydrological in- formation and hydrological forecasting (Gb/T 22482-2008),” in Ministry of Water Resources of the People’s Republic of China, Standards Press of China, Beijing, China, 2008. [53] C. Gabriele, M. Cinzia, E. Ortiz, and T. Ezio, “Application of the TOPKAPI model within the DMIP 2 Project,” in Pro- ceedings of the 23rd Conference on Hydrology, Phoenix, AZ, USA, January 2009. International Journal of The Scientific Advances in Advances in Geophysics Chemistry Scientica World Journal Public Health Hindawi Hindawi Hindawi Hindawi Publishing Corporation Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www www.hindawi.com .hindawi.com V Volume 2018 olume 2013 www.hindawi.com Volume 2018 Journal of Environmental and Public Health Advances in Meteorology Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com Applied & Environmental Journal of Soil Science Geological Research Hindawi Volume 2018 Hindawi www.hindawi.com www.hindawi.com Volume 2018 International Journal of International Journal of Agronomy Ecology International Journal of Advances in International Journal of Forestry Research Microbiology Agriculture Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 International Journal of Journal of Journal of International Journal of Biodiversity Archaea Analytical Chemistry Chemistry Marine Biology Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal

Advances in Meteorology – Hindawi Publishing Corporation

Published: Sep 5, 2019

There are no references for this article.