Hydrological field data from a modeller's perspective: Part 1. Diagnostic tests for model structure

Hydrological scientists develop perceptual models of the catchments they study, using field measurements and observations to build an understanding of the dominant processes controlling the hydrological response. However, conceptual and numerical models used to simulate catchment behaviour often fail to take advantage of this knowledge. It is common instead to use a pre‐defined model structure which can only be fitted to the catchment via parameter calibration. In this article, we suggest an alternative approach where different sources of field data are used to build a synthesis of dominant hydrological processes and hence provide recommendations for representing those processes in a time‐stepping simulation model. Using analysis of precipitation, flow and soil moisture data, recommendations are made for a comprehensive set of modelling decisions, including Evapotranspiration (ET) parameterization, vertical drainage threshold and behaviour, depth and water holding capacity of the active soil zone, unsaturated and saturated zone model architecture and deep groundwater flow behaviour. The second article in this two‐part series implements those recommendations and tests the capability of different model sub‐components to represent the observed hydrological processes. Copyright © 2010 John Wiley & Sons, Ltd.     

Understanding coastal wetland hydrology with a new regional‐scale,process‐based hydrological model

Coastal wetlands represent an ecotone between ocean and terrestrial ecosystems, providing important services, including flood mitigation, fresh water supply, erosion control, carbon sequestration, and wildlife habitat. The environmental setting of a wetland and the hydrological connectivity between a wetland and adjacent terrestrial and aquatic systems together determine wetland hydrology. Yet little is known about regional‐scale hydrological interactions among uplands, coastal wetlands, and coastal processes, such as tides, sea level rise, and saltwater intrusion, which together control the dynamics of wetland hydrology. This study presents a new regional‐scale, physically based, distributed wetland hydrological model, PIHM‐Wetland, which integrates the surface and subsurface hydrology with coastal processes and accounts for the influence of wetland inundation on energy budgets and evapotranspiration (ET). The model was validated using in situ hydro‐meteorological measurements and Moderate Resolution Imaging Spectroradiometer (MODIS) ET data for a forested and herbaceous wetland in North Carolina, USA, which confirmed that the model accurately represents the major wetland hydrological behaviours. Modelling results indicate that topographic gradient is a primary control of groundwater flow direction in adjacent uplands. However, seasonal climate patterns become the dominant control of groundwater flow at lower coastal plain and land–ocean interface. We found that coastal processes largely influence groundwater table (GWT) dynamics in the coastal zone, 300 to 800 m from the coastline in our study area. Among all the coastal processes, tides are the dominant control on GWT variation. Because of inundation, forested and herbaceous wetlands absorb an additional 6% and 10%, respectively, of shortwave radiation annually, resulting in a significant increase in ET. Inundation alters ET partitioning through canopy evaporation, transpiration, and soil evaporation, the effect of which is stronger in cool seasons than in warm seasons. The PIHM‐Wetland model provides a new tool that improves the understanding of wetland hydrological processes on a regional scale. Insights from this modelling study provide benchmarks for future research on the effects of sea level rise and climate change on coastal wetland functions and services.     

On the role of spatial resolution on snow estimates using a process‐based snow model across a range of climatology and elevation

Hydrological processes in mountainous settings depend on snow distribution, whose prediction accuracy is a function of model spatial scale. Although model accuracy is expected to improve with finer spatial resolution, an increase in resolution comes with modelling costs related to increased computational time and greater input data and parameter information. This computational and data collection expense is still a limiting factor for many large watersheds. Thus, this work's main objective is to question which physical processes lead to loss in model accuracy with regard to input spatial resolution under different climatic conditions and elevation ranges. To address this objective, a spatially distributed snow model, iSnobal, was run with inputs distributed at 50‐m—our benchmark for comparison—and 100‐m resolutions and with aggregated (averaged from the fine to the large resolution) inputs from the 50‐m model to 100‐, 250‐, 500‐, and 750‐m resolution for wet, average, and dry years over the Upper Boise River Basin (6,963 km²), which spans four elevation bands: rain dominated, rain–snow transition, and snow dominated below treeline and above treeline. Residuals, defined as differences between values quantified with high resolution (>50 m) models minus the benchmark model (50 m), of simulated snow‐covered area (SCA) and snow water equivalent (SWE) were generally slight in the aggregated scenarios. This was due to transferring the effects of topography on meteorological variables from the 50‐m model to the coarser scales through aggregation. Residuals in SCA and SWE in the distributed 100‐m simulation were greater than those of the aggregated 750 m. Topographic features such as slope and aspect were simplified, and their gradient was reduced due to coarsening the topography from the 50‐ to 100‐m resolution. Therefore, solar radiation was overestimated, and snow drifting was modified and caused substantial SCA and SWE underestimation in the distributed 100‐m model relative to the 50‐m model. Large residuals were observed in the wet year and at the highest elevation band when and where snow mass was large. These results support that model accuracy is substantially reduced with model scales coarser than 50 m.     

Role of temporal resolution of meteorological inputs for process‐based snow modelling

Accurate snow accumulation and melt simulations are crucial for understanding and predicting hydrological dynamics in mountainous settings. As snow models require temporally varying meteorological inputs, time resolution of these inputs is likely to play an important role on the model accuracy. Because meteorological data at a fine temporal resolution (~1 hr) are generally not available in many snow‐dominated settings, it is important to evaluate the role of meteorological inputs temporal resolution on the performance of process‐based snow models. The objective of this work is to assess the loss in model accuracy with temporal resolution of meteorological inputs, for a range of climatic conditions and topographic elevations. To this end, a process‐based snow model was run using 1‐, 3‐, and 6‐hourly inputs for wet, average, and dry years over Boise River Basin (6,963 km²), which spans rain dominated (≤1,400 m), rain–snow transition (>1,400 and ≤1,900 m), snow dominated below tree line (>1,900 and ≤2,400 m), and above tree line (>2,400 m) elevations. The results show that sensitivity of the model accuracy to the inputs time step generally decreases with increasing elevation from rain dominated to snow dominated above tree line. Using longer than hourly inputs causes substantial underestimation of snow cover area (SCA) and snow water equivalent (SWE) in rain‐dominated and rain–snow transition elevations, due to the precipitation phase mischaracterization. In snow‐dominated elevations, the melt rate is underestimated due to errors in estimation of net snow cover energy input. In addition, the errors in SCA and SWE estimates generally decrease toward years with low snow mass, that is, dry years. The results indicate significant increases in errors in estimates of SCA and SWE as the temporal resolution of meteorological inputs becomes coarser than an hour. However, use of 3‐hourly inputs can provide accurate estimates at snow‐dominated elevations. The study underscores the need to record meteorological variables at an hourly time step for accurate process‐based snow modelling.     

Quality assessment from a hydrological perspective of a digital elevation model derived from WorldView-2 remote sensing data

Abstract

A digital elevation model (DEM) derived from a stereo pair of WorldView-2 (WV-2) images was assessed against ground-truth GPS point datasets. Two assessment methods were used: (a) vertical accuracy assessment and (b) hydrological assessment of surface runoff variables. Three agricultural plots with different topographic slopes were selected to perform a vertical accuracy assessment, followed by a comparative assessment of a set of hydrological variables. The results show an overall vertical accuracy of 0.45 m, confirming the potential of WV-2 stereo images to extract elevation information at high spatial resolution. Concerning plot-scale micro-topographic features, the WV-2 DEM performed better on the plot with rolling slopes (5–10%), extracting variables such as the total length and drainage area of flow paths with relative errors lower than 20%. However, some limitations were detected in the extraction of variables such as terrain slope, drainage points of flow paths and terrain depressions in areas of flatter slopes (<5%).

Editor Z.W. Kundzewicz; Associate editor D. Gerten     

Parameterization and multi‐criteria calibration of a distributed storm flow model applied to a Mediterranean agricultural catchment

The principal challenge in the parameterization of storm flow models for agricultural catchments with an artificial drainage network and fields with different degrees of tillage lies in the parsimonious definition of distributed model parameters in a way that reduces the number of calibration parameters to a justifiable minimum. This paper presents a comprehensive case study for the parameter estimation of a distributed storm flow model applied to an agricultural catchment (0.91 km²) in the Mediterranean region. Model parameterization was combined with procedures for multi‐criteria, multi‐storm calibration, where we automatically calibrated three parameters related to flow velocity and infiltration, and compared single and multi‐storm criteria that are based on discharge volume, peak flow, and the Nash–Sutcliffe coefficient. Multi‐storm calibration yielded a set of parameter values for the simulation batch with best multi‐storm overall performance, which are close to the median values in the pre‐calibration of individual storms. Our results suggest that flow velocities and proportionality of the channel infiltration rate do not vary significantly over the course of 11 years. Copyright © 2012 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.     

Estimating suspended sediment concentrations in areas with limited hydrological data using a mixed‐effects model

Sediment rating curves are commonly used to estimate the suspended sediment load in rivers and streams under the assumption of a constant relation between discharge (Q) and suspended sediment concentrations (SSC) over time. However, temporal variation in the sediment supply of a watershed results in shifts in this relation by increasing variability and by introducing nonlinearities in the form of hysteresis or a path‐dependent relation. In this study, we used a mixed‐effects linear model to estimate an average SSC–Q relation for different periods of time within the hydrologic cycle while accounting for seasonality and hysteresis. We tested the performance of the mixed‐effects model against the standard rating curve, represented by a generalized least squares regression, by comparing observed and predicted sediment loads for a test case on the Chilliwack River, British Columbia, Canada. In our analyses, the mixed‐effects model reflected more accurate patterns of interpolated SSC from Q data than the rating curve, especially for the low‐flow summer months when the SSC–Q relation is less clear. Akaike information criterion scores were lower for the mixed‐effects model than for the standard model, and the mixed‐effects model explained nearly twice as much variance as the standard model (52% vs 27%). The improved performance was achieved by accounting for variability in the SSC–Q relation within each month and across years for the same month using fixed and random effects, respectively, a characteristic disregarded in the sediment rating curve. Copyright © 2012 John Wiley & Sons, Ltd.     

Scale effects of Hortonian overland flow and rainfall‐runoff dynamics: laboratory validation of a process‐based model

Hortonian runoff was measured in the laboratory from uniform slopes of lengths of 1·5, 3·0, and 6·0 m for steady, high‐intensity rainstorms with durations of 1·0 to 7·5 min. A clear reduction in runoff per unit slope length was found as slope lengths were increased. This effect becomes more pronounced with decreasing storm duration. The runoff data were used to validate a simple process‐based model that combines the Philip‐two‐term infiltration equation with the kinematic wave overland flow principle. The predicted and experimental results agreed well. Laboratory findings were extrapolated with the aid of the model to slopes and rainfall durations similar to those found under West African conditions. The calculated reduction of runoff per unit length is similar to reported observations. Thus, this process‐based model can largely explain the phenomenon of runoff reduction with increasing slope length. Copyright © 2002 John Wiley & Sons, Ltd.     

Conditional spectrum‐based ground motion selection. Part II: Intensity‐based assessments and evaluation of alternative target spectra

In a companion paper, an overview and problem definition was presented for ground motion selection on the basis of the conditional spectrum (CS), to perform risk‐based assessments (which estimate the annual rate of exceeding a specified structural response amplitude) for a 20‐story reinforced concrete frame structure. Here, the methodology is repeated for intensity‐based assessments (which estimate structural response for ground motions with a specified intensity level) to determine the effect of conditioning period. Additionally, intensity‐based and risk‐based assessments are evaluated for two other possible target spectra, specifically the uniform hazard spectrum (UHS) and the conditional mean spectrum (CMS, without variability).It is demonstrated for the structure considered that the choice of conditioning period in the CS can substantially impact structural response estimates in an intensity‐based assessment. When used for intensity‐based assessments, the UHS typically results in equal or higher median estimates of structural response than the CS; the CMS results in similar median estimates of structural response compared with the CS but exhibits lower dispersion because of the omission of variability. The choice of target spectrum is then evaluated for risk‐based assessments, showing that the UHS results in overestimation of structural response hazard, whereas the CMS results in underestimation. Additional analyses are completed for other structures to confirm the generality of the conclusions here. These findings have potentially important implications both for the intensity‐based seismic assessments using the CS in future building codes and the risk‐based seismic assessments typically used in performance‐based earthquake engineering applications. Copyright © 2013 John Wiley & Sons, Ltd.     

Simulating pan‐Arctic runoff with a macro‐scale terrestrial water balance model

A terrestrial hydrological model, developed to simulate the high‐latitude water cycle, is described, along with comparisons with observed data across the pan‐Arctic drainage basin. Gridded fields of plant rooting depth, soil characteristics (texture, organic content), vegetation, and daily time series of precipitation and air temperature provide the primary inputs used to derive simulated runoff at a grid resolution of 25 km across the pan‐Arctic. The pan‐Arctic water balance model (P/WBM) includes a simple scheme for simulating daily changes in soil frozen and liquid water amounts, with the thaw–freeze model (TFM) driven by air temperature, modelled soil moisture content, and physiographic data. Climate time series (precipitation and air temperature) are from the National Centers for Environmental Prediction (NCEP) reanalysis project for the period 1980–2001. P/WBM‐generated maximum summer active‐layer thickness estimates differ from a set of observed data by an average of 12 cm at 27 sites in Alaska, with many of the differences within the variability (1σ) seen in field samples. Simulated long‐term annual runoffs are in the range 100 to 400 mm year^?1. The highest runoffs are found across northeastern Canada, southern Alaska, and Norway, and lower estimates are noted along the highest latitudes of the terrestrial Arctic in North America and Asia. Good agreement exists between simulated and observed long‐term seasonal (winter, spring, summer–fall) runoff to the ten Arctic sea basins (r = 0·84). Model water budgets are most sensitive to changes in precipitation and air temperature, whereas less affect is noted when other model parameters are altered. Increasing daily precipitation by 25% amplifies annual runoff by 50 to 80% for the largest Arctic drainage basins. Ignoring soil ice by eliminating the TFM sub‐model leads to runoffs that are 7 to 27% lower than the control run. The results of these model sensitivity experiments, along with other uncertainties in both observed validation data and model inputs, emphasize the need to develop improved spatial data sets of key geophysical quantities (particularly climate time series) to estimate terrestrial Arctic hydrological budgets better. Copyright © 2003 John Wiley & Sons, Ltd.     

Residual structural capacity evaluation of steel moment‐resisting frames with dynamic‐strain‐based model updating method

This paper presents a method for evaluating the residual structural capacity of earthquake‐affected steel structures. The method first quantifies the damage severity of a beam by computing the dynamic‐strain‐based damage index. Next, the model used to analyze the structure is updated based on the damage index, to reflect the observed damage conditions. The residual structural capacity is then estimated in terms of changes in stiffness and strength, which can be applied by structural engineers, via a nonlinear static analysis of the updated model. The main contributions of this paper are in performance evaluation of the dynamic‐strain‐based damage index for seismically induced damage using a newly developed substructure testing environment, consideration of various damage patterns in composite beams, and extension of a local damage evaluation technique to a residual capacity estimation procedure by incorporating the model‐updating technique. In laboratory testing, the specimens were damaged quasi‐statically, and vibration tests were conducted as the damage proceeded. First, a bare steel beam–column connection was tested, and then a similar one with a floor slab was used for a more realistic case. The estimated residual structural capacities for these specimens were compared with the static test results. The results verified that the proposed method can provide fine estimates of the stiffness and strength deteriorations within 10% for the specimen without the floor slab and within 30% for that with the floor slab. Copyright © 2017 John Wiley & Sons, Ltd.     

Parameter identification for on‐line model updating in hybrid simulations using a gradient‐based method

To improve the efficiency of model fitting, parameter identification techniques have been actively investigated. Recently, the applications of parameter identification migrated from off‐line model fitting to on‐line model updating. The objective of this study is to develop a gradient‐based method for model updating to advance hybrid simulation also called hybrid test. A novel modification of the proposed method, which can reduce the number of design variables to improve the identification efficiency, is illustrated in detail. To investigate the model updating, simulated hybrid tests were conducted with a 5‐story steel frame equipped with buckling‐restrained braces (BRBs) utilized in the shaking table tests conducted in E‐Defense in Japan in 2009. The calibrated analytical model that was verified with the test results can serve as the reference model. In the simulated hybrid tests, the physical BRB substructure is numerically simulated by utilizing a truss element with the 2‐surface model identical to the part of the reference model. Such numerical verification allows simulation of measurement errors for investigation on the performance of the proposed method. Moreover, the feasibility of sharing the identified parameter values, which were obtained from the physical substructure responses, with the relevant numerical models is also verified with the artificial component responses derived from the physical experiments.     

Merging energy‐based design criteria and reliability‐based methods: exploring a new concept

This paper examines the potential development of a probabilistic design methodology, considering hysteretic energy demand, within the framework of performance‐based seismic design of buildings. This article does not propose specific energy‐based criteria for design guidelines, but explores how such criteria can be treated from a probabilistic design perspective. Uniform hazard spectra for normalized hysteretic energy are constructed to characterize seismic demand at a specific site. These spectra, in combination with an equivalent systems methodology, are used to estimate hysteretic energy demand on real building structures. A design checking equation for a (hypothetical) probabilistic energy‐based performance criterion is developed by accounting for the randomness of the earthquake phenomenon, the uncertainties associated with the equivalent system analysis technique, and with the site soil factor. The developed design checking equation itself is deterministic, and requires no probabilistic analysis for use. The application of the proposed equation is demonstrated by applying it to a trial design of a three‐storey steel moment frame. The design checking equation represents a first step toward the development of a performance‐based seismic design procedure based on energy criterion, and additional works needed to fully implement this are discussed in brief at the end of the paper. Copyright © 2006 John Wiley & Sons, Ltd.     

A coupled hydrology–biogeochemistry model to simulate dissolved organic carbon exports from a permafrost‐influenced catchment

We outline the development of a simple, coupled hydrology–biogeochemistry model for simulating stream discharge and dissolved organic carbon (DOC) dynamics in data sparse, permafrost‐influenced catchments with large stores of soil organic carbon. The model incorporates the influence of active layer dynamics and slope aspect on hydrological flowpaths and resulting DOC mobilization. Calibration and evaluation of the model was undertaken using observations from Granger Basin within the Wolf Creek research basin, Yukon, northern Canada. Results show that the model was able to capture the dominant hydrological response and DOC dynamics of the catchment reasonably well. Simulated DOC was highly correlated with observed DOC (r² = 0.65) for the study period. During the snowmelt period, the model adequately captured the observed dynamics, with simulations generally reflecting the timing and magnitude of the observed DOC and stream discharge. The model was less successful over the later summer period although this partly reflected a lack of DOC observations for calibration. The developed model offers a valuable framework for investigating the interactions between hydrological and DOC processes in these highly dynamic systems, where data acquisition is often very difficult. © 2015 The Authors Hydrological Processes Published by John Wiley & Sons, Ltd.     

Coastal profile response to sea level rise: a process‐based approach

Prevailing ideas and calculations of coastal response to sea level rise (SLR) are often based on the Bruun model (Bruun P., Sea‐level rise as a cause of shore erosion, Journal Waterways Harbors Division, ASCE 88 : 117–130, 1962) that predicts upward and landward transfer of an equilibrium profile during SLR through offshore sediment transport on the shoreface. The model is based on a number of assumptions of questionable validity as well as outdated concepts on how sediment is transported across the shoreface. This contribution takes a numerical modelling approach that is based on first‐order processes contributing to the movement of sediment across the shoreface. Using a wave transformation model that predicts hydrodynamic processes driving cross‐shore sediment transport and an energetics‐based model for the coupling between hydrodynamics and sediment transport, we show that cross‐shore sediment transport is mainly onshore directed at the boundary between the lower and the upper shoreface, in agreement with the model proposed by Davidson‐Arnott (Conceptual model of the effects of sea level rise on sandy coasts, Journal of Coastal Research 21 : 1166–1172, 2005). The transition from onshore to offshore directed transport is located well within the surf zone and with a rising sea level this transition point becomes displaced landward and upward. Tests also show that substrate slope is of fundamental importance to the manner in which beaches react to rising sea level. Copyright © 2012 John Wiley & Sons, Ltd.     

Application of 1D paleo‐fluvial process modelling at a basin scale to augment sparse borehole data: example of a Permian formation in the Galilee Basin,Australia

The heterogeneous hydraulic conductivity (K) in water‐bearing formations controls subsurface flow and solute transport processes. Geostatistical techniques are often employed to characterize the K distribution in space based on the correlation between K measurements. However, at the basin scale, there are often insufficient measurements for inferring the spatial correlation. This is a widespread problem that we address in this study using the example of the Betts Creek Beds (BCB) in the Galilee Basin, Australia. To address the lack of data, we use a 1D stochastic fluvial process‐based model (SFPM) to quantify the total sediment thickness, Z( x ), and the sandstone proportion over the total thickness, P_s( x ), in the BCB. The semivariograms of Z( x ) and P_s( x ) are then extracted and used in sequential Gaussian simulation to construct the 2D spatial distribution of Z( x ) and P_s( x ). P_s( x ) can be converted to a K distribution based on classical averaging methods. The results demonstrate that the combination of SFPM and geostatistical simulation allows for the evaluation of upscaled K distribution with a limited number of K measurements. Copyright © 2015 John Wiley & Sons, Ltd.     

On the use of multi‐algorithm,genetically adaptive multi‐objective method for multi‐site calibration of the SWAT model

With the availability of spatially distributed data, distributed hydrologic models are increasingly used for simulation of spatially varied hydrologic processes to understand and manage natural and human activities that affect watershed systems. Multi‐objective optimization methods have been applied to calibrate distributed hydrologic models using observed data from multiple sites. As the time consumed by running these complex models is increasing substantially, selecting efficient and effective multi‐objective optimization algorithms is becoming a nontrivial issue. In this study, we evaluated a multi‐algorithm, genetically adaptive multi‐objective method (AMALGAM) for multi‐site calibration of a distributed hydrologic model—Soil and Water Assessment Tool (SWAT), and compared its performance with two widely used evolutionary multi‐objective optimization (EMO) algorithms (i.e. Strength Pareto Evolutionary Algorithm 2 (SPEA2) and Non‐dominated Sorted Genetic Algorithm II (NSGA‐II)). In order to provide insights into each method's overall performance, these three methods were tested in four watersheds with various characteristics. The test results indicate that the AMALGAM can consistently provide competitive or superior results compared with the other two methods. The multi‐method search framework of AMALGAM, which can flexibly and adaptively utilize multiple optimization algorithms, makes it a promising tool for multi‐site calibration of the distributed SWAT. For practical use of AMALGAM, it is suggested to implement this method in multiple trials with relatively small number of model runs rather than run it once with long iterations. In addition, incorporating different multi‐objective optimization algorithms and multi‐mode search operators into AMALGAM deserves further research. Copyright © 2009 John Wiley & Sons, Ltd.