Multi‐model ensemble prediction of terrestrial evapotranspiration across north China using Bayesian model averaging

Using high‐quality dataset from 12 flux towers in north China, the performance of four evapotranspiration (ET) models and the multi‐model ensemble approaches including the simple averaging (SA) and Bayesian model average (BMA) were systematically evaluated in this study. The four models were the single‐layer Penman–Monteith (P–M) model, the two‐layer Shuttleworthe–Wallace (S–W) model, the advection–aridity (A–A) model, and a modified Priestley–Taylor (PT‐JPL). Based on the mean value of Taylor skill (S) and the regression slope between measured and simulated ET values across all sites, the order of overall performance of the individual models from the best to the worst were: S–W (0.88, 0.87), PT‐JPL (0.80, 1.17), P–M (0.63, 1.73) and A–A (0.60, 1.68) [statistics stated as (Taylor skill, regression slope)]. Here, all models used the same values of parameters, LAI and fractional vegetation cover as well as the forcing meteorological data. Thus, the differences in model performance were mainly attributed to errors in model structure. To the ensemble approach, the BMA method has the advantage of generating more skillful and reliable predictions than the SA scheme. However, successful implementation of BMA requires accurate estimates of its parameters, and some degradation in performance were observed when the BMA parameters generated from the training period were used for the validation period. Thus, it is necessary to explore the seasonal variations of the BMA parameters according the different growth stages. Finally, the optimal conditional density function of half‐hourly ET approximated well by the double‐exponential distribution. Copyright © 2016 John Wiley & Sons, Ltd.     

Evaluating best evaporation estimate model for water surface evaporation in semi‐arid region,India

Evaluating performances of four commonly used evaporation estimate methods, namely; Bowen ratio energy balance (BREB), mass transfer (MT), Priestley–Taylor (PT) and pan evaporation (PE), based on 4 years experimental data, the most effective and the reliable evaporation estimates model for the semi‐arid region of India has been derived. The various goodness‐of‐fit measures, such as; coefficient of determination (R²), index of agreement (D), root mean square error (RMSE), and relative bias (RB) have been chosen for the performance evaluation. Of these models, the PT model has been found most promising when the Bowen ratio, β is known a priori, and based on its limited data requirement. The responses of the BREB, the PT, and the PE models were found comparable to each other, while the response of the MT model differed to match with the responses of the other three models. The coefficients, β of the BREB, µ of the MT, α of the PT and K_P of the PE model were estimated as 0·07, 2·35, 1·31 and 0·65, respectively. The PT model can successfully be extended for free water surface evaporation estimates in semi‐arid India. A linear regression model depicting relationship between daily air and water temperature has been developed using the observed water temperatures and the corresponding air temperatures. The model helped to generate unrecorded water temperatures for the corresponding ambient air temperatures. Copyright © 2007 John Wiley & Sons, Ltd.     

Predicting catchment flow in a semi‐arid region via an artificial neural network technique

A model of rainfall–runoff relationships is an essential tool in the process of evaluation of water resources projects. In this paper, we applied an artificial neural network (ANN) based model for flow prediction using the data for a catchment in a semi‐arid region in Morocco. Use of this method for non‐linear modelling has been demonstrated in several scientific fields such as biology, geology, chemistry and physics. The performance of the developed neural network‐based model was compared against multiple linear regression‐based model using the same observed data. It was found that the neural network model consistently gives superior predictions. Based on the results of this study, artificial neural network modelling appears to be a promising technique for the prediction of flow for catchments in semi‐arid regions. Accordingly, the neural network method can be applied to various hydrological systems where other models may be inappropriate. Copyright © 2004 John Wiley & Sons, Ltd.     

Centralized semi‐active control of post‐tensioned steel frames

Centralized semi‐active control is a technique for controlling the whole structure using one main computer. Centralized control systems introduce better control for relatively short to medium high structures where the response of any story cannot be separated from the adjacent ones. In this paper, two centralized control approaches are proposed for controlling the seismic response of post‐tensioned (PT) steel frames. The first approach, the stiffness control approach, aims to alter the stiffness of the PT frame so that it avoids large dynamic amplifications due to earthquake excitations. The second approach, deformation regulation control approach, aims at redistributing the demand/strength ratio in order to provide a more uniform distribution of deformations over the height of the structure. The two control approaches were assessed through simulations of the earthquake response of semi‐actively and passively controlled six‐story post‐tensioned steel frames. The results showed that the stiffness control approach is efficient in reducing the frame deformations and internal forces. The deformation regulation control approach was found to be efficient in reducing the frame displacements and generating a more uniform distribution of the inter‐story drifts. These results indicate that centralized semi‐active control can be used to improve the seismic performance of post‐tensioned steel frames. Copyright © 2014 John Wiley & Sons, Ltd.     

A geomorphology‐based integrated stream–aquifer interaction model for semi‐gauged catchments

Many of the existing stream–aquifer interaction models available in the literature are very complex with limited applicability in semi‐gauged and ungauged catchments. In this study, to estimate the influent and effluent subsurface water fluxes under limited geo‐hydrometeorological data availability conditions, a simple stream–aquifer interaction model, namely, the variable parameter McCarthy–Muskingum (VPMM) hillslope‐storage Boussinesq (hsB) model, has been developed. This novel model couples the VPMM streamflow transport with the hsB groundwater flow transport modules in online mode. In this integrated model, the surface water–groundwater flux exchange process is modelled by the Darcian approach with the variable hydraulic heads between the river stage and groundwater table accounting for the rainfall forcing. Considering the exchange fluxes in the hyporheic zone and lateral overland flow contribution, this approach is field tested in a typical 48‐km stretch of the Brahmani River in eastern India to simulate the streamflow and its depth with the minimum Nash–Sutcliffe efficiency of 94% and 88%; the maximum root mean square error of 134 m³/s and 0.35 m; and the minimum index of agreement of 98% and 97%, respectively. This modelling approach could be very well utilized in data‐scarce world‐river basins to estimate the stream–aquifer exchange flux due to rainfall forcings.     

Thermodynamic consistency of potential evapotranspiration estimates in Florida

This study investigates the validity of fundamental assumptions regarding the partitioning of energy that are implicit in the Penman–Monteith (PM) and Priestly–Taylor (PT) models of potential evapotranspiration (PET). Both these models require energy conservation, but differ in the partitioning of sensible and latent heat fluxes as measured by the Bowen ratio. Application of these models at four research sites in Florida collectively show that these models can be calibrated to provide adequate PET prediction at all sites. The prediction of Bowen ratios using the calibrated PET models showed that the PT model is preferable for lakes and is acceptable for marsh and forest; the crop‐coefficient application of the PM model is preferable for grass and is acceptable for marsh; and the standard PM model is acceptable for marsh. For the marsh and forest covers, the margins of preference between the three models are not very strong. This investigation was limited to climatic conditions in Florida and caution should be exercised in extending these results to other sites and climates. Copyright © 2010 John Wiley & Sons, Ltd.     

The assessment of surface water resources for the semi‐arid Yongding River Basin from 1956 to 2000 and the impact of land use change

The assessment of surface water resources (SWRs) in the semi‐arid Yongding River Basin is vital as the basin has been in a continuous state of serious water shortage over the last 20 years. In this study, the first version of the geomorphology‐based hydrological model (GBHM) has been applied to the basin over a long period of time (1956–2000) as part of an SWR assessment. This was done by simulating the natural hydrological processes in the basin. The model was first evaluated at 18 stream gauges during the period from 1990 to 1992 to evaluate both the daily streamflows and the annual SWRs using the land use data for 1990. The model was further validated in 2000 with the annual SWRs at seven major stream gauges. Second, the verified model was used in a 45‐year simulation to estimate the annual SWRs for the basin from 1956 to 2000 using the 1990 land use data. An empirical correlation between the annual precipitation and the annual SWRs was developed for the basin. Spatial distribution of the long‐term mean runoff coefficients for all 177 sub‐basins was also achieved. Third, an additional 10‐year (1991–2000) simulation was performed with the 2000 land use data to investigate the impact of land use changes from 1990 to 2000 on the long‐term annual SWRs. The results suggest that the 10‐year land use changes have led to a decrease of 8·3 × 10⁷ m³ (7·9% of total) for the 10‐year mean annual SWRs in the simulation. To our knowledge, this work is the first attempt to assess the long‐term SWRs and the impact of land use change in the semi‐arid Yongding River Basin using a semi‐distributed hillslope hydrological model. Copyright © 2010 John Wiley & Sons, Ltd.     

Spatial and temporal variability of annual and seasonal precipitation over the desert region of China during 1951–2005

The spatial and temporal variations of precipitation in the desert region of China (DRC) from 1951 to 2005 were investigated using a rotated empirical orthogonal function (REOF), the precipitation concentration index (PCI) and the Mann–Kendall trend test method (M‐K method). In addition, the association between variation patterns of precipitation and large‐scale circulation were also explored using the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis data. The results indicated that the spatial pattern of precipitation was primarily the local climate effect significant type, with the first three EOFs explaining a total of 55·3% of the variance, and the large‐scale climate system effect type, which explained 9·8% of the variance. Prior to the 1970s, the East Asian summer monsoon was stronger, which resulted in abundant precipitation in the Inner Mongolia region. Conversely, the climate of the Xinjiang region was controlled by westerly circulation and had lower precipitation. However, this situation has been reversed since the 1980s. It is predicted that precipitation will decrease by 15–40 and 0–10 mm/year in the Inner Mongolia plateau and southern Xinjiang, respectively, whereas it will likely increase by 10–40 mm/year in northern Xinjiang. Additionally, 58–62% of the annual rainfall occurred during summer in the DRC, with precipitation increasing during spring and summer and decreasing in winter. The intra‐annual precipitation is becoming uniform, but the inter‐annual variability in precipitation has been increasing in the western portions of the DRC. The probability of precipitation during the study period increased by 30% and 22·2% in the extreme‐arid zones and arid zones, respectively. Conversely, the probability of precipitation during the study period decreased by 18·5% and 37·5% in the semi‐arid zones and semi‐wet zones, respectively. It is predicted that the northwest portion of the DRC will become warmer and wetter, while the central portion will become warmer and drier and the northeast portion will be subjected to drought. Copyright © 2010 John Wiley & Sons, Ltd.     

Modelling surface runoff to evaluate the effects of wildfires in multiple semi‐arid,shrubland‐dominated catchments

Wildfires change the infiltration properties of soil, reduce the amount of interception and result in increased runoff. A wildfire at Northeast Attica, Central Greece, in August 2009, destroyed approximately one third of a study area consisting of a mixture of shrublands, pastures and pines. The present study simultaneously models multiple semi‐arid, shrubland‐dominated Mediterranean catchments and assesses the hydrological response (mean annual and monthly runoff and runoff coefficients) during the first few years following wildfires. A physically based, hydrological model (MIKE SHE) was chosen. Calibration and validation results of mean monthly discharge presented very good agreement with the observed data for the pre‐wildfire and post‐wildfire period for two subcatchments (Nash–Sutcliffe Efficiency coefficient of 79.7%). The model was then used to assess the pre‐wildfire and post‐wildfire runoff responses for each of seven catchments in the study area. Mean annual surface runoff increased for the first year and after the second year following the wildfires increased by 112% and 166%, respectively. These values are within the range observed in similar cases of monitored sites. This modelling approach may provide a way of prioritizing catchment selection with respect to post‐fire remediation activities. Additionally, this modelling assessment methodology would be valuable to other semi‐arid areas because it provides an important means for comprehensively assessing post‐wildfire response over large regions and therefore attempts to address some of the scaled issues in the specific literature field of research. Copyright © 2015 John Wiley & Sons, Ltd.     

Spatial distribution and temporal variation of reference evapotranspiration in arid and semi‐arid regions of Iran

Analysis of spatial and temporal variations of reference evapotranspiration (ET_o) is important in arid and semi‐arid regions where water resources are limited. The main aim of this study was to analyse the spatial distribution and the annual, seasonal and monthly trends of the Penman–Monteith ET_o for 21 stations in the arid and semi‐arid regions of Iran. Three statistical tests the Mann‐Kendall, Sen's slope estimator and linear regression were used for the analysis. The analysis revealed that ET_o increased from January to July and deceased from July to December at almost all stations. Additionally, higher annual ET_o values were found in the southeast of the study region and lower values in the northwest of the region. Although the results showed both positive and negative trends in annual ET_o series, ET_o generally increased, significantly so in six (～30%) of the stations. Analysis of the impacts of meteorological variables on the temporal trends of ET_o indicated that the increasing trend of ET_o was most likely due to a significant increase in minimum air temperature, while decreasing trend of ET_o was mainly caused by a significant decrease in wind speed. At the sites where increasing ET_o trends were statistically significant, the rate of increase varied from (+)8·36 mm/year at Mashhad station to (+)31·68 mm/year at Iranshahr station. On average, an increasing trend of (+)4·42 mm/year was obtained for the whole study area during the last four decades. Seasonal and monthly ET_o have also tended to increase at the majority of the stations. The greatest numbers of significant trends were observed in winter on the seasonal time‐scale and in September on the monthly time‐scale. Copyright © 2011 John Wiley & Sons, Ltd.     

Improving the representation of soil moisture by using a semi‐analytical infiltration model

Soil moisture is widely recognized as a fundamental variable governing the mass and energy fluxes between the land surface and the atmosphere. In this study, the soil moisture modelling at sub‐daily timescale is addressed by using an accurate representation of the infiltration component. For that, the semi‐analytical infiltration model proposed by Corradini et al. (1997) has been incorporated into a soil water balance model to simulate the evolution in time of surface and profile soil moisture. The performances of this new soil moisture model [soil water balance module‐semi‐analytical (SWBM‐SA)] are compared with those of a precedent version [SWBM‐Green–Ampt (GA)] where the GA approach was employed. Their capability to reproduce in situ soil moisture observations at three sites in Italy, Spain and France is analysed. Hourly observations of quality‐checked rainfall, temperature and soil moisture data for a 2‐year period are used for testing the modelling approaches. Specifically, different configurations for the calibration and validation of the models are adopted by varying a single parameter, that is, the saturated hydraulic conductivity. Results indicate that both SWBMs are able to reproduce satisfactorily the hourly soil moisture temporal pattern for the three sites with root mean square errors lower than 0.024 m³/m³ both in the calibration and validation periods. For all sites, the SWBM‐SA model outperforms the SWBM‐GA with an average reduction of the root mean square error of ~20%. Specifically, the higher improvement is observed for the French site for which in situ observations are measured at 30 cm depth, and this is attributed to the capability of the SA infiltration model to simulate the time evolution of the whole soil moisture profile. The reasonable models performance coupled with the need to calibrate only a single parameter makes them useful tools for soil moisture simulation in different regions worldwide, also in scarcely gauged areas. Copyright © 2013 John Wiley & Sons, Ltd.     

Energy balance and evaporation loss of an agricultural reservoir in a semi‐arid climate (south‐eastern Spain)

A typical agricultural water reservoir (AWR) of 2400 m² area and 5 m depth, located in a semi‐arid area (southern Spain), was surveyed on a daily basis for 1 year. The annual evaporation flux was 102·7 W m^?2, equivalent to an evaporated water depth of 1310 mm year^?1. The heat storage rate G exhibited a clear annual cycle with a peak gain in April (G ～ 45 W m^?2) and a peak loss in November (G ～ 40 W m^?2), leading to a marked annual hysteretic trend when evaporation (λE) was related to net radiation (R_n). λE was strongly correlated with the available energy A, representing 91% of the annual AWR energy loss. The sensible heat flux H accounted for the remaining 9%, leading to an annual Bowen ratio in the order of 0·10. The equilibrium and advective evaporation terms of the Penman formula represented 76 and 24%, respectively, of the total evaporation, corresponding to a annual value of the Priestley–Taylor (P–T) coefficient (α) of 1·32. The P–T coefficient presented a clear seasonal pattern, with a minimum of 1·23 (July) and a maximum of 1·65 (December), indicating that, during periods of limited available energy, AWR evaporation increased above the potential evaporation as a result of the advection process. Overall, the results stressed that accurate prediction of monthly evaporation by means of the P–T formula requires accounting for both the annual cycle of storage and the advective component. Some alternative approaches to estimating R_n, G and α are proposed and discussed. Copyright © 2009 John Wiley & Sons, Ltd.     

Performance of a conceptual and physically based model in simulating the response of a semi‐urbanized watershed in San Antonio,Texas

The need for accurate hydrologic analysis and rainfall–runoff modelling tools has been rapidly increasing because of the growing complexity of operational hydrologic and hydraulic problems associated with population growth, rapid urbanization and expansion of agricultural activities. Given the recent advances in remote sensing of physiographic features and the availability of near real‐time precipitation products, rainfall–runoff models are expected to predict runoff more accurately. In this study, we compare the performance and implementation requirements of two rainfall–runoff models for a semi‐urbanized watershed. One is a semi‐distributed conceptual model, the Hydrologic Engineering Center‐Hydrologic Modelling System (HEC‐HMS). The other is a physically based, distributed‐parameter hydrologic model, the Gridded Surface Subsurface Hydrologic Analysis (GSSHA). Four flood events that took place on the Leon Creek watershed, a sub‐watershed of the San Antonio River basin in Texas, were used in this study. The two models were driven by the Multisensor Precipitation Estimator radar products. One event (in 2007) was used for HEC‐HMS and GSSHA calibrations. Two events (in 2004 and 2007) were used for further calibration of HEC‐HMS. Three events (in 2002, 2004 and 2010) were used for model validation. In general, the physically based, distributed‐parameter model performed better than the conceptual model and required less calibration. The two models were prepared with the same minimum required input data, and the effort required to build the two models did not differ substantially. Copyright © 2012 John Wiley & Sons, Ltd.     

Hydrological processes of a closed catchment‐lake system in a semi‐arid Mediterranean environment

Data collected in 4 years of field observations were used in conjunction with continuous simulation models to study, at the small‐basin scale, the water balance of a closed catchment‐lake system in a semi‐arid Mediterranean environment. The open water evaporation was computed with the Penman equation, using the data set collected in the middle of the lake. The surface runoff was partly measured at the main tributary and partly simulated using a distributed, catchment, hydrological model, calibrated with the observed discharge. The simplified structure of the developed modelling mainly concerns soil moisture dynamics and bedrock hydraulics, whereas the flow components are physically based. The calibration produced high efficiency coefficients and showed that surface runoff is greatly affected by soil water percolation into fractured bedrock. The bedrock reduces the storm‐flow peaks and the interflow and has important multi‐year effects on the annual runoff coefficients. The net subsurface outflow from the lake was calculated as the residual of the lake water balance. It was almost constant in the dry seasons and increased in the wet seasons, because of the moistening of the unsaturated soil. During the years of observation, rainfall 30% higher than average caused abundant runoff and a continuous rise in the lake water levels. The analysis allows to predict that, in years with lower than the average rainfall, runoff will be drastically reduced and will not be able to compensate for negative balance between precipitation and lake evaporation. Such highly unsteady situations, with great fluctuations in lake levels, are typical of closed catchment‐lake systems in the semi‐arid Mediterranean environment. Copyright © 2012 John Wiley & Sons, Ltd.     

Direct finite element method for nonlinear earthquake analysis of 3‐dimensional semi‐unbounded dam–water–foundation rock systems

A direct finite element method for nonlinear earthquake analysis of 2‐dimensional dam–water–foundation rock systems has recently been presented. The analysis procedure uses standard viscous‐damper absorbing boundaries to model the semi‐unbounded foundation‐rock and fluid domains and specifies the seismic input as effective earthquake forces at these boundaries. Presented in this paper is a generalization of the direct finite element method with viscous‐damper boundaries to 3‐dimensional dam–water–foundation rock systems. Step‐by‐step procedures for determining the effective earthquake forces starting from a ground motion specified at a control point on the foundation‐rock surface is developed, and several numerical examples are computed and compared with independent benchmark solutions to demonstrate the effectiveness of the analysis procedure for modeling 3‐dimensional systems.     

Precipitation stable isotope variability and subcloud evaporation processes in a semi‐arid region

The stable isotopic (²H/¹H and ¹⁸O/¹⁶O) composition of precipitation has been used for a variety of hydrological and paleoclimate studies, a starting point for which is the behaviour of stable isotopes in modern precipitation. To this end, daily precipitation samples were collected over a 7‐year period (2008–2014) at a semi‐arid site located at the Macquarie Marshes, New South Wales (Australia). The samples were analysed for stable isotope composition, and factors affecting the isotopic variability were investigated. The best correlation between δ ¹⁸O of precipitation was with local surface relative humidity. The reduced major axis precipitation weighted local meteoric water line was δ ²H = 7.20 δ ¹⁸O + 9.1. The lower slope and intercept (when compared with the Global Meteoric Water Line) are typical for a warm dry climate, where subcloud evaporation of raindrops is experienced. A previously published model to estimate the degree of subcloud evaporation and the subsequent isotopic modification of raindrops was enhanced to include the vertical temperature and humidity profile. The modelled results for raindrops of 1.0 mm radius showed that on average, the measured D‐excess (=δ ²H ? 8 δ ¹⁸O) was 19.8‰ lower than that at the base of the cloud, and 18% of the moisture was evaporated before ground level (smaller effects were modelled for larger raindrops). After estimating the isotopic signature at the base of the cloud, a number of data points still plotted below the global meteoric water line, suggesting that some of the moisture was sourced from previously evaporated water. Back trajectory analysis estimated that 38% of the moisture was sourced over land. Precipitation samples for which a larger proportion of the moisture was sourced over land were ¹⁸O and ²H‐enriched in comparison to samples for which the majority of the moisture was sourced over the ocean. The most common weather systems resulting in precipitation were inland trough systems; however, only East Coast Lows contributed to a significant difference in the isotopic values. Copyright © 2016 Australian Nuclear Science and Technology Organisation. Hydrological Processes. © 2016 John Wiley & Sons, Ltd.     

River–aquifer interactions in a semi‐arid environment stressed by groundwater abstraction

Rivers and aquifers are, in many cases, a connected resource and as such the interactions between them need to be understood and quantified for the resource to be managed appropriately. The objective of this paper is to advance the understanding of river–aquifer interactions processes in semi‐arid environments stressed by groundwater abstraction. This is performed using data from a specific catchment where records of precipitation, evapotranspiration, river flow, groundwater levels and groundwater abstraction are analysed using basic statistics, hydrograph analysis and a simple mathematical model to determine the processes causing the spatial and temporal changes in river–aquifer interactions. This combined approach provides a novel but simple methodology to analyse river–aquifer interactions, which can be applied to catchments worldwide. The analysis revealed that the groundwater levels have declined (~ 3 m) since the onset of groundwater abstraction. The decline is predominantly due to the abstraction rather than climatic changes (r = 0.84 for the relationship between groundwater abstraction and groundwater levels; r = 0.92 for the relationship between decline in groundwater levels and magnitude of seasonal drawdown). It is then demonstrated that, since the onset of abstraction, the river has changed from being gaining to losing during low‐flow periods, defined as periods with flow less than 0.5, 1.0 or 1.5 GL/day (1 GL/day = 1 × 10⁶ m³/day). If defined as < 1.0 GL/day, low‐flow periods constitute approximately 65% of the river flows; the periods where the river is losing at low‐flow conditions are thus significant. Importantly, there was a significant delay (> 10 years) between the onset of groundwater abstraction and the changeover from gaining to losing conditions. Finally, a relationship between the groundwater gradient towards the river and the river flow at low‐flow is demonstrated. The results have important implications for water management as well as water ecology and quality. Copyright © 2012 John Wiley & Sons, Ltd.     

Performance‐based design with semi‐active structural control technique

The recent spate of large earthquakes has triggered diverse performance requirements for structures. This has led to increasing worldwide interest in performance‐based design methods. To establish such methods, however, it is necessary to evaluate structure conditions after defining the loads, and this is difficult to accomplish. On the other hand, there has been steady progress on research and development of structural control techniques for improving structural performance. These technological innovations need to be rationally incorporated into structural design. In particular, semi‐active structural control techniques are effective in improving structural performance during large earthquakes. By effectively incorporating them into the design, it is possible to meet the various structural performance requirements. This paper first outlines the various structural control methods and focuses on the semi‐active structural control technique as the main topic. It then describes an example to verify the effectiveness of the semi‐active structural control technique in high‐rise buildings. Copyright © 2001 John Wiley & Sons, Ltd.     

Comparing the performance of empirical black‐box models for river flow forecasting in the Heihe River Basin,Northwestern China

For many practical reasons, the empirical black‐box models have become an increasingly popular modelling tool for river flow forecasting, especially in mountainous areas where very few meteorological observatories exist. In this article, precipitation data are used as the only input to estimate river flow. Using five empirical black‐box models—the simple linear model, the linear perturbation model, the linearly varying gain factor model, the constrained nonlinear system model and the nonlinear perturbation model–antecedent precipitation index—modelling results are compared with actual results in three catchments within the Heihe River Basin. The linearly varying gain factor model and the nonlinear perturbation model yielded excellent predictions. For better simulation accuracy, a commonly used multilayer feed‐forward neural network model (NNM) was applied to incorporate the outputs of the individual models. Comparing the performance of these models, it was found that the best results were obtained from the NNM model. The results also suggest that more reliable and precise predictions of river flow can be obtained by using the NNM model while also incorporating the combined outputs of different empirical black‐box models. Copyright © 2012 John Wiley & Sons, Ltd.