赣江流域多种数据同化方案的径流模拟比较

为了探讨水文模型在不同水文数据同化方案下的径流模拟差异,本文采用集合卡尔曼滤波算法,以遥感蒸散发产品、实测径流为观测数据,构建了基于新安江模型的数据同化框架。基于此框架设计了4种不同同化方案(DA-ET、DAET(K)、DA-ET-Q、DA-ET-Q(K))以及1种对照方案OL,以赣江流域开展实例研究,评估了水文数据同化中遥感蒸散发产品的时间分辨率、模型蒸散发相关参数时变与否以及多源数据同化对径流模拟的影响。结果表明:在DA-ET方案下,同化两种不同时间分辨率的蒸散发产品均能提高模型整体的径流模拟精度,且时间分辨率更高的产品的同化效果更好;在DA-ET方案的基础上,考虑加入实测径流进行同化能够提升模型径流模拟精度,且DA-ET(K)与DA-ET-Q(K)方案所得径流相对误差的减幅均超过了20%,说明在蒸散发同化过程中同时考虑蒸散发参数动态变化的结果更优;相较于OL方案,4种同化方案均能不同程度地提高模型对径流高水部分的模拟能力,但DA-ET-Q(K)方案表现最差,而其余方案差异并不显著。本研究有助于进一步了解不同数据同化方案在径流模拟中的差异,从而为水资源高效利用与科学管理提供科学依据...     

Impact of the spatial variability of daily precipitation on hydrological projections: A comparison of GCM‐ and RCM‐driven cases in the Han River basin,Korea

In this study, we investigate the impact of the spatial variability of daily precipitation on hydrological projections based on a comparative assessment of streamflow simulations driven by a global climate model (GCM) and two regional climate models (RCMs). A total of 12 different climate input datasets, that is, the raw and bias‐corrected GCM and raw and bias‐corrected two RCMs for the reference and future periods, are fed to a semidistributed hydrological model to assess whether the bias correction using quantile mapping and dynamical downscaling using RCMs can improve streamflow simulation in the Han River basin, Korea. A statistical analysis of the daily precipitation demonstrates that the precipitation simulated by the GCM fails to capture the large variability of the observed daily precipitation, in which the spatial autocorrelation decreases sharply within a relatively short distance. However, the spatial variability of precipitation simulated by the two RCMs shows better agreement with the observations. After applying bias correction to the raw GCM and raw RCMs outputs, only a slight change is observed in the spatial variability, whereas an improvement is observed in the precipitation intensity. Intensified precipitation but with the same spatial variability of the raw output from the bias‐corrected GCM does not improve the heterogeneous runoff distributions, which in turn regulate unrealistically high peak downstream streamflow. GCM‐simulated precipitation with a large bias correction that is necessary to compensate for the poor performance in present climate simulation appears to distort streamflow patterns in the future projection, which leads to misleading projections of climate change impacts on hydrological extremes.     

Evaluation and hydrological application of precipitation estimates derived from PERSIANN‐CDR,TRMM 3B42V7, and NCEP‐CFSR over humid regions in China

Satellite‐based and reanalysis quantitative precipitation estimates are attractive for hydrologic prediction or forecasting and reliable water resources management, especially for ungauged regions. This study evaluates three widely used global high‐resolution precipitation products [Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks‐Climate Data Record (PERSIANN‐CDR), Tropical Rainfall Measuring Mission 3B42 Version 7 (TRMM 3B42V7), and National Centers for Environment Prediction‐Climate Forecast System Reanalysis (NCEP‐CFSR)] against gauge observations with seven statistical indices over two humid regions in China. Furthermore, the study investigates whether the three precipitation products can be reliably utilized as inputs in Soil and Water Assessment Tool, a semi‐distributed hydrological model, to simulate streamflows. Results show that the precipitation estimates derived from TRMM 3B42V7 outperform the other two products with the smallest errors and bias, and highest correlation at monthly scale, which is followed by PERSIANN‐CDR and NCEP‐CFSR in this rank. However, the superiority of TRMM 3B42V7 in errors, bias, and correlations is not warranted at daily scale. PERSIANN‐CDR and 3B42V7 present encouraging potential for streamflow prediction at daily and monthly scale respectively over the two humid regions, whilst the performance of NCEP‐CFSR for hydrological applications varies from basin to basin. Simulations forced with 3B42V7 are the best among the three precipitation products in capturing daily measured streamflows, whilst PERSIANN‐CDR‐driven simulations underestimate high streamflows and high streamflow simulations driven by NCEP‐CFSR mostly are overestimated significantly. In terms of extreme events analysis, PERSIANN‐CDR often underestimates the extreme precipitation, so do extreme streamflow simulations forced with it. NCEP‐CFSR performs just the reverse, compared with PERSIANN‐CDR. The performance pattern of TRMM 3B42V7 on extremes is not certain, with coexisting underestimation and overestimation. Copyright © 2016 John Wiley & Sons, Ltd.     

Streamflow simulations for continental-scale river basins in a global atmospheric general circulation model

Streamflow simulations for 23 major river basins from the third-generation general circulation model (GCM) of the Canadian Centre for Climate Modelling and Analysis are assessed. Precipitation and runoff data are used from the AMIP II simulation in which the GCM is integrated for a 17-yr period with specific sea surface temperatures and sea-ice concentrations. Compared to the observations, the components of the global hydrological cycle and, the globally averaged precipitation and runoff over land, are well simulated. There remain, however, discrepancies in the simulation of regional precipitation and consequently runoff amounts, which lead to differences in basin-wide averaged quantities. Mean annual model precipitation is within 20% of the observed estimates for 13 out of 23 river basins considered. Model mean annual runoff is within 20% of the observed estimates for only 4 out of these 13 river basins. Analysis of basin-wide averaged monthly precipitation and streamflow data, and the errors associated with the mean, and amplitude and phase of the annual cycles, indicate that model streamflow simulations improve with improvement in GCM precipitation.     

Sensitivity analysis and uncertainty assessment in water budgets simulated by the variable infiltration capacity model for Canadian subarctic watersheds

In this study, we evaluate uncertainties propagated through different climate data sets in seasonal and annual hydrological simulations over 10 subarctic watersheds of northern Manitoba, Canada, using the variable infiltration capacity (VIC) model. Further, we perform a comprehensive sensitivity and uncertainty analysis of the VIC model using a robust and state-of-the-art approach. The VIC model simulations utilize the recently developed variogram analysis of response surfaces (VARS) technique that requires in this application more than 6,000 model simulations for a 30-year (1981–2010) study period. The method seeks parameter sensitivity, identifies influential parameters, and showcases streamflow sensitivity to parameter uncertainty at seasonal and annual timescales. Results suggest that the Ensemble VIC simulations match observed streamflow closest, whereas global reanalysis products yield high flows (0.5–3.0 mm day⁻¹) against observations and an overestimation (10–60%) in seasonal and annual water balance terms. VIC parameters exhibit seasonal importance in VARS, and the choice of input data and performance metrics substantially affect sensitivity analysis. Uncertainty propagation due to input forcing selection in each water balance term (i.e., total runoff, soil moisture, and evapotranspiration) is examined separately to show both time and space dimensionality in available forcing data at seasonal and annual timescales. Reliable input forcing, the most influential model parameters, and the uncertainty envelope in streamflow prediction are presented for the VIC model. These results, along with some specific recommendations, are expected to assist the broader VIC modelling community and other users of VARS and land surface schemes, to enhance their modelling applications.     

Statistical downscaling of GCM simulations to streamflow using relevance vector machine

General circulation models (GCMs), the climate models often used in assessing the impact of climate change, operate on a coarse scale and thus the simulation results obtained from GCMs are not particularly useful in a comparatively smaller river basin scale hydrology. The article presents a methodology of statistical downscaling based on sparse Bayesian learning and Relevance Vector Machine (RVM) to model streamflow at river basin scale for monsoon period (June, July, August, September) using GCM simulated climatic variables. NCEP/NCAR reanalysis data have been used for training the model to establish a statistical relationship between streamflow and climatic variables. The relationship thus obtained is used to project the future streamflow from GCM simulations. The statistical methodology involves principal component analysis, fuzzy clustering and RVM. Different kernel functions are used for comparison purpose. The model is applied to Mahanadi river basin in India. The results obtained using RVM are compared with those of state-of-the-art Support Vector Machine (SVM) to present the advantages of RVMs over SVMs. A decreasing trend is observed for monsoon streamflow of Mahanadi due to high surface warming in future, with the CCSR/NIES GCM and B2 scenario.     

Choosing an arbitrary calibration period for hydrologic models: How much does it influence water balance simulations?

The selection of calibration and validation time periods in hydrologic modelling is often done arbitrarily. Nonstationarity can lead to an optimal parameter set for one period which may not accurately simulate another. However, there is still much to be learned about the responses of hydrologic models to nonstationary conditions. We investigated how the selection of calibration and validation periods can influence water balance simulations. We calibrated Soil and Water Assessment Tool hydrologic models with observed streamflow for three United States watersheds (St. Joseph River of Indiana/Michigan, Escambia River of Florida/Alabama, and Cottonwood Creek of California), using time period splits for calibration/validation. We found that the choice of calibration period (with different patterns of observed streamflow, precipitation, and air temperature) influenced the parameter sets, leading to dissimilar simulations of water balance components. In the Cottonwood Creek watershed, simulations of 50-year mean January streamflow varied by 32%, because of lower winter precipitation and air temperature in earlier calibration periods on calibrated parameters, which impaired the ability for models calibrated to earlier periods to simulate later periods. Peaks of actual evapotranspiration for this watershed also shifted from April to May due to different parameter values depending on the calibration period's winter air temperatures. In the St. Joseph and Escambia River watersheds, adjustments of the runoff curve number parameter could vary by 10.7% and 20.8%, respectively, while 50-year mean monthly surface runoff simulations could vary by 23%–37% and 169%–209%, depending on the observed streamflow and precipitation of the chosen calibration period. It is imperative that calibration and validation time periods are chosen selectively instead of arbitrarily, for instance using change point detection methods, and that the calibration periods are appropriate for the goals of the study, considering possible broad effects of nonstationary time series on water balance simulations. It is also crucial that the hydrologic modelling community improves existing calibration and validation practices to better include nonstationary processes.     

Improving WRF-Hydro runoff simulations of heavy floods through the sea surface temperature fields with higher spatio-temporal resolution

This study investigates the impact of the spatio-temporal accuracy of four different sea surface temperature (SST) datasets on the accuracy of the Weather Research and Forecasting (WRF)-Hydro system to simulate hydrological response during two catastrophic flood events over the Eastern Black Sea (EBS) and the Mediterranean (MED) regions of Turkey. Three time-variant and high spatial resolution external SST products (GHRSST, Medspiration and NCEP-SST) and one coarse-resolution and time-invariant SST product (ERA5- and GFS-SST for EBS and MED regions, respectively) already embedded in the initial and the boundary conditions datasets of WRF model are used in deriving near-surface atmospheric variables through WRF. After the proper event-based calibration is performed to the WRF-Hydro system using hourly and daily streamflow data in both regions, uncoupled model simulations for independent SST events are conducted to assess the impact of SST-triggered precipitation on simulated extreme runoff. Some localized and temporal differences in the occurrence of the flood events with respect to observations depending on the SST representation are noticeable. SST products represented with higher cross-correlations (GHRSST and Medspiration) revealed significant improvement in flood hydrographs for both regions. The GHRSST dataset shows a substantial improvement in NSE (~70%), RMSE reduction up to 20%, and an increase in correlation from 0.3 to 0.8 with respect to the invariable SST (ERA5) in simulated runoffs over the EBS region. The use of both GHRSST and Medspiration SST data characterized with high spatio-temporal correlation resulted in runoff simulations exactly matching the observed runoff peak of 300 m³/s by reducing the overestimation seen in invariable SST (GFS) in the MED region. Improved precipitation simulation skills of the WRF model with the detailed SST representation show that the hydrographs of GHRSST and Medspiration simulations show better performance compared to the simulated hydrographs by observed precipitation.     

Spatial interpolation schemes of daily precipitation for hydrologic modeling

Distributed hydrologic models typically require spatial estimates of precipitation interpolated from sparsely located observational points to the specific grid points. We compare and contrast the performance of regression-based statistical methods for the spatial estimation of precipitation in two hydrologically different basins and confirmed that widely used regression-based estimation schemes fail to describe the realistic spatial variability of daily precipitation field. The methods assessed are: (1) inverse distance weighted average; (2) multiple linear regression (MLR); (3) climatological MLR; and (4) locally weighted polynomial regression (LWP). In order to improve the performance of the interpolations, the authors propose a two-step regression technique for effective daily precipitation estimation. In this simple two-step estimation process, precipitation occurrence is first generated via a logistic regression model before estimate the amount of precipitation separately on wet days. This process generated the precipitation occurrence, amount, and spatial correlation effectively. A distributed hydrologic model (PRMS) was used for the impact analysis in daily time step simulation. Multiple simulations suggested noticeable differences between the input alternatives generated by three different interpolation schemes. Differences are shown in overall simulation error against the observations, degree of explained variability, and seasonal volumes. Simulated streamflows also showed different characteristics in mean, maximum, minimum, and peak flows. Given the same parameter optimization technique, LWP input showed least streamflow error in Alapaha basin and CMLR input showed least error (still very close to LWP) in Animas basin. All of the two-step interpolation inputs resulted in lower streamflow error compared to the directly interpolated inputs.     

Hydrological modelling using proxies for gauged precipitation and temperature

Precipitation and temperature time series suffer from many problems, such as short time, inadequate spatial coverage, missing data, and biases from various causes, which are particularly critical in remote areas such as Northern Canada. The development of alternative datasets for using as proxies for inadequate/missing weather data represents a key research area. In this paper, the performance of 6 alternative datasets is evaluated for hydrological modelling over 12 watersheds located across Canada and the contiguous United States. The datasets can be classified into 3 distinct categories: (a) interpolated gridded data, (b) reanalysis data, and (c) climate model outputs. Hydrological simulations were carried out using a lumped conceptual hydrological model calibrated using standard weather data and compared against results using a calibration specific to each alternative dataset. Prior to the hydrological simulations, the alternative datasets were all evaluated with respect to their ability to reproduce gridded daily precipitation and temperature characteristics over North America. The results show that both the reanalysis data and climate model data adequately represent the spatial pattern of daily precipitation and temperature over North America. The North American Regional Reanalysis (NARR) dataset consistently shows the best performance. With respect to hydrological modelling, the observed discharges are accurately represented by both the gridded and NARR datasets, and more so for the NARR data. The National Centers for Environmental Prediction dataset consistently performs worst as it is unable to even capture the seasonal pattern of observed streamflow for 3 out of the 12 watersheds. These results indicate that the NARR dataset could be used as a proxy for gauged precipitation and temperature for hydrological modelling over watersheds where observational datasets are deficient. The results also illustrate the ability of climate model data to be used for performing hydrological modelling when driven by reanalysis data at their boundaries, and especially so for high‐resolution models.     

Temporal analysis of Soil and Water Assessment Tool (SWAT) performance based on remotely sensed precipitation products

No study has systematically evaluated streamflow modelling between monthly and daily time scales. This study examines streamflow from seven watersheds across the USA where five different precipitation products were used as primary input into the Soil and Water Assessment Tool (SWAT) to generate simulated streamflow. Time scales examined include monthly, dekad (10 days), pentad (5 days), triad (3 days), and daily. The seven basins studied are the San Pedro (Arizona), Cimarron (north‐central Oklahoma), mid‐Nueces (south Texas), mid‐Rio Grande (south Texas and northern Mexico), Yocano (northern Mississippi), Alapaha (south Georgia), and mid‐St. Francis (eastern Arkansas). The precipitation products used to drive simulations include rain gauge, NWS Multisensor Precipitation Estimator, Tropical Rainfall Measurement Mission (TRMM), Multi‐Satellite Precipitation Analysis, TRMM 3B42‐V6, and Climate Prediction Center Morphing Method (CMORPH). Understanding how streamflow varies at sub‐monthly time scales is important because there are a host of hydrological applications such a flood forecast guidance and reservoir inflow forecasts that reside in a temporal domain between monthly and daily time scales. The major finding of this study is the quantification of a strong positive correlation between performance metrics and time step at which model performance deteriorates. Better performing simulations, with higher Nash–Sutcliffe values of 0.80 and above can support modeling at finer time scales to at least daily and perhaps beyond into the sub‐daily realm. These findings are significant in that they clearly document the ability of SWAT to support modeling at sub‐monthly time steps, which is beyond the capability for which SWAT was initially designed. Copyright © 2012 John Wiley & Sons, Ltd.     

Calibration and uncertainty analysis of the SWAT model using Genetic Algorithms and Bayesian Model Averaging

In this paper, the Genetic Algorithms (GA) and Bayesian Model Averaging (BMA) were used to simultaneously conduct calibration and uncertainty analysis for the Soil and Water Assessment Tool (SWAT). In this combined method, several SWAT models with different structures are first selected; next GA is used to calibrate each model using observed streamflow data; finally, BMA is applied to combine the ensemble predictions and provide uncertainty interval estimation. This method was tested in two contrasting basins, the Little River Experimental Basin in Georgia, USA, and the Yellow River Headwater Basin in China. The results obtained in the two case studies show that this combined method can provide deterministic predictions better than or comparable to the best calibrated model using GA. The 66.7% and 90% uncertainty intervals estimated by this method were analyzed. The differences between the percentage of coverage of observations and the corresponding expected coverage percentage are within 10% for both calibration and validation periods in these two test basins. This combined methodology provides a practical and flexible tool to attain reliable deterministic simulation and uncertainty analysis of SWAT.     

Comparison of radar and gauge precipitation data in watershed models across varying spatial and temporal scales

Precipitation is a key control on watershed hydrologic modelling output, with errors in rainfall propagating through subsequent stages of water quantity and quality analysis. Most watershed models incorporate precipitation data from rain gauges; higher‐resolution data sources are available, but they are associated with greater computational requirements and expertise. Here, we investigate whether the Multisensor Precipitation Estimator (MPE or Stage IV Next‐Generation Radar) data improve the accuracy of streamflow simulations using the Soil and Water Assessment Tool (SWAT), compared with rain gauge data. Simulated flows from 2002 to 2010 at five timesteps were compared with observed flows for four nested subwatersheds of the Neuse River basin in North Carolina (21‐, 203‐, 2979‐, and 10 100‐km² watershed area), using a multi‐objective function, informal likelihood‐weighted calibration approach. Across watersheds and timesteps, total gauge precipitation was greater than radar precipitation, but radar data showed a conditional bias of higher rainfall estimates during large events (>25–50 mm/day). Model parameterization differed between calibrations with the two datasets, despite the fact that all watershed characteristics were the same across simulation scenarios. This underscores the importance of linking calibration parameters to realistic processes. SWAT simulations with both datasets underestimated median and low flows, whereas radar‐based simulations were more accurate than gauge‐based simulations for high flows. At coarser timesteps, differences were less pronounced. Our results suggest that modelling efforts in watersheds with poor rain gauge coverage can be improved with MPE radar data, especially at short timesteps. Published 2013. This article is a U.S. Government work and is in the public domain in the USA.     

Hydrologic response of a Hawaiian watershed to future climate change scenarios

The impact of potential future climate change scenarios on streamflow and evapotranspiration (ET) in a mountainous Hawaii watershed was studied using the distributed hydrology soil vegetation model (DHSVM). The hydrologic response of the watershed was simulated for 43 years for different levels of atmospheric CO₂ (330, 550, 710 and 970 ppm), temperature (+1.1 and + 6.4 °C) and precipitation (±5%, ±10% and ±20%) on the basis of the Intergovernmental Panel on Climate Change (IPCC) AR4 projections under current, B1, A1B1 and A1F1 emission scenarios. Vegetation leaf conductance and leaf area index were modified to reflect the increase in CO₂ concentration. The relative departure of streamflow and ET from their levels during the reference scenarios was calculated on a monthly and annual basis. Results of this study indicate that the streamflow and ET are less sensitive to changes in temperature compared with changes in precipitation. However, temperature increase coupled with precipitation showed significant effect on ET and streamflow. Changes in leaf conductance and leaf area index with increasing CO₂ concentration under A1F1 scenario had a significant effect on ET and subsequently on streamflow. Evapotranspiration is less sensitive than streamflow for a similar level of change in precipitation. On the basis of a range of climate change scenarios, DHSVM predicted a change in ET by ±10% and streamflow between ?51% and 90%. From the six ensemble mean scenarios for AR4 A1B, simulations suggest reduction in streamflow by 6.7% to 17.2%. These reductions would produce severe impact on water availability in the region. Copyright © 2011 John Wiley & Sons, Ltd.     

Evaluation of dynamically downscaled reanalysis precipitation data for hydrological application

Skilful and reliable precipitation data are essential for seasonal hydrologic forecasting and generation of hydrological data. Although output from dynamic downscaling methods is used for hydrological application, the existence of systematic errors in dynamically downscaled data adversely affects the skill of hydrologic forecasting. This study evaluates the precipitation data derived by dynamically downscaling the global atmospheric reanalysis data by propagating them through three hydrological models. Hydrological models are calibrated for 28 watersheds located across the southeastern United States that is minimally affected by human intervention. Calibrated hydrological models are forced with five different types of datasets: global atmospheric reanalysis (National Centers for Environmental Prediction/Department of Energy Global Reanalysis and European Centre for Medium‐Range Weather Forecasts 40‐year Reanalysis) at their native resolution; dynamically downscaled global atmospheric reanalysis at 10‐km grid resolution; stochastically generated data from weather generator; bias‐corrected dynamically downscaled; and bias‐corrected global reanalysis. The reanalysis products are considered as surrogates for large‐scale observations. Our study indicates that over the 28 watersheds in the southeastern United States, the simulated hydrological response to the bias‐corrected dynamically downscaled data is superior to the other four meteorological datasets. In comparison with synthetically generated meteorological forcing (from weather generator), the dynamically downscaled data from global atmospheric reanalysis result in more realistic hydrological simulations. Therefore, we conclude that dynamical downscaling of global reanalysis, which offers data for sufficient number of years (in this case 22 years), although resource intensive, is relatively more useful than other sources of meteorological data with comparable period in simulating realistic hydrological response at watershed scales. Copyright © 2013 John Wiley & Sons, Ltd.     

Space–time changes in hydrological processes in response to human activities and climatic change in the south China

Due to the influence of climate change and human activities, more and more regions around the world are nowadays facing serious water shortages. This is particularly so with the Guangdong province, an economically prosperous region in China. This study aims at understanding the abrupt behavior of hydrological processes by analyzing monthly precipitation series from 257 rain gauging stations and monthly streamflow series from 25 hydrological stations using the likelihood ratio statistic and schwarz information criterion (SIC). The underlying causes of the changing properties of hydrological processes are investigated by analyzing precipitation changes and information of water reservoirs. It is found that (1) streamflow series in dry season seems to exhibit abrupt changes when compared to that in the flood season; (2) abrupt changes in the values of mean and variance of hydrological variables in the dry season are more common than those in the streamflow series in the flood season, which implies that streamflow in the dry season is more sensitive to human activities and climate change than that in the flood season; (3) no change points are identified in the annual precipitation and precipitation series in the flood season. Annual streamflow and streamflow in the flood season exhibit no abrupt changes, showing the influence of precipitation on streamflow changes in the flood season. However, streamflow changes in the dry season seem to be heavily influenced by hydrological regulations of water reservoirs. The results of this study are of practical importance for regional water resource management in the Guangdong province.     

中国区域夏季再分析资料高空变量可信度的检验

利用全球探空资料(IGRA)对1989—2008年美国国家环境预报中心(NCEP)和大气研究中心(NCAR)再分析资料、NCEP和美国能源部(DOE)再分析资料、NCEP气候预测系统再分析资料(CFSR)、日本气象厅25年再分析资料(JRA-25)、欧洲数值预报中心再分析资料(ERA-Interim)和美国国家航空航天局(NASA)现代回顾性再分析资料(MERRA)的高空变量在中国地区对流层中高层的可信度进行了初步的检验.分析结果表明:再分析资料对中高层位势高度和温度的夏季平均气候态具有较好的再现能力,其EOF的时空变化特征与观测吻合也较好;再分析资料的绝对湿度值较观测结果要偏大,其中MERRA与观测最为接近.再分析资料不能很好地反映经向风的夏季平均气候态及年际变化特征,EOF的时空模态和观测偏离也较大.总体而言,NCEP/NCAR、NCEP/DOE及NCEP/CFSR对这些变量的再现能力较JRA-25、ERA-Interim和MERRA弱.     

Incorporating the spatio-temporal distribution of rainfall and basin geomorphology into nonlinear analyses of streamflow dynamics

Many recent studies have been devoted to the investigation of the nonlinear dynamics of rainfall or streamflow series based on methods of dynamical systems theory. Although finding evidence for the existence of a low-dimensional deterministic component in rainfall or streamflow is of much interest, not much attention has been given to the nonlinear dependencies of the two and especially on how the spatio-temporal distribution of rainfall affects the nonlinear dynamics of streamflow at flood time scales. In this paper, a methodology is presented which simultaneously considers streamflow series, spatio-temporal structure of precipitation and catchment geomorphology into a nonlinear analysis of streamflow dynamics. The proposed framework is based on “hydrologically-relevant” rainfall-runoff phase-space reconstruction acknowledging the fact that rainfall-runoff is a stochastic spatially extended system rather than a deterministic multivariate one. The methodology is applied to two basins in Central North America using 6-hour streamflow data and radar images for a period of 5 years. The proposed methodology is used to: (a) quantify the nonlinear dependencies between streamflow dynamics and the spatio-temporal dynamics of precipitation; (b) study how streamflow predictability is affected by the trade-offs between the level of detail necessary to explain the spatial variability of rainfall and the reduction of complexity due to the smoothing effect of the basin; and (c) explore the possibility of incorporating process-specific information (in terms of catchment geomorphology and an a priori chosen uncertainty model) into nonlinear prediction. Preliminary results are encouraging and indicate the potential of using the proposed methodology to understand via nonlinear analysis of observations (i.e., not based on a particular rainfall-runoff model) streamflow predictability and limits to prediction as a function of the complexity of spatio-temporal forcing relative to basin geomorphology.     

A hybrid factorial stepwise-cluster analysis method for streamflow simulation – a case study in northwestern China

ABSTRACT

In this study, a hybrid factorial stepwise-cluster analysis (HFSA) method is developed for modelling hydrological processes. The HFSA method employs a cluster tree to represent the complex nonlinear relationship between inputs (predictors) and outputs (predictands) in hydrological processes. A real case of streamflow simulation for the Kaidu River basin is applied to demonstrate the efficiency of the HFSA method. After training a total of 24?108 calibration samples, the cluster tree for daily streamflow is generated based on a stepwise-cluster analysis (SCA) approach and is then used to reproduce the daily streamflows for calibration (1995–2005) and validation (2008–2010) periods. The Nash-Sutcliffe coefficients for calibration and validation are 0.68 and 0.65, respectively, and the deviations of volume are 1.68% and 4.11%, respectively. Results show that: (i) the HFSA method can formulate a SCA-based hydrological modelling system for streamflow simulation with a satisfactory fitting; (ii) the variability and peak value of streamflow in the Kaidu River basin can be effectively captured by the SCA-based hydrological modelling system; (iii) results from 2⁶ factorial experiments indicate that not only are minimum temperature and precipitation key drivers of system performance, but also the interaction between precipitation and minimum temperature significantly impacts on the streamflow. The findings are useful in indicating that the streamflow of the study basin is a mixture of snowmelt and rainfall water.

EDITOR D. Koutsoyiannis; ASSOCIATE EDITOR G. Thirel