Propagation of uncertainty from rainfall to runoff: A case study with a stochastic rainfall generator

We explore the impact of uncertainties in the spatial–temporal distribution of rainfall on the prediction of peak discharge in a typical mountain basin. To this end, we use a stochastic generator previously developed for rainfall downscaling, and we estimate the basin response by adopting a semi-distributed hydrological model. The results of the analysis provide information on the minimum rainfall resolution needed for operational flood forecasting, and confirm the sensitivity of peak discharge estimates to errors in the determination of the power spectrum of the precipitation field.     

Cross-spectral analysis of rainfall and runoff for Raritan and Mullica river basins in New Jersey

Cross-correlation and cross-spectral analysis were employed in the analysis of rainfall and runoff in two river basins: the Raritan and Mullica River basins in New Jersey. Cross-covariance and coherence were studied in the correlograms for the following correlation cases: (a) rainfall-runoff for each one basin separately; (b) rainfall-rainfall analysis for two main meteorological stations in each one of the basins; (c) runoff-runoff for two main gaging stations in each one of the basins. From the estimates of the coherence at various frequencies the cross-spectral analysis shows a highly nonlinear relationship between rainfall and runoff in Raritan and Mullica River basins. A poor coherence observed at the annual cycles for each basin makes it difficult to predict the annual oscillations of runoff from those of rain-fall by a linear regression model. The high coherence between rainfall (or runoff) at the first station and rainfall (or runoff) at the second station within the same basin at almost all frequencies establishes an accurate prediction on a linear basis.     

Modeling scale-dependent runoff generation in a small semi-arid watershed accounting for rainfall intensity and water depth

Observed scale effects of runoff on hillslopes and small watersheds derive from complex interactions of time-varying rainfall rates with runoff, infiltration and macro- and microtopographic structures. A little studied aspect of scale effects is the concept of water depth-dependent infiltration. For semi-arid rangeland it has been demonstrated that mounds underneath shrubs have a high infiltrability and lower lying compacted or stony inter-shrub areas have a lower infiltrability. It is hypothesized that runoff accumulation further downslope leads to increased water depth, inundating high infiltrability areas, which increases the area-averaged infiltration rate. A model was developed that combines the concepts of water depth-dependent infiltration, partial contributing area under variable rainfall intensity, and the Green–Ampt theory for point-scale infiltration. The model was applied to rainfall simulation data and natural rainfall–runoff data from a small sub-watershed (0.4 ha) of the Walnut Gulch Experimental Watershed in the semi-arid US Southwest. Its performance to reproduce observed hydrographs was compared to that of a conventional Green–Ampt model assuming complete inundation sheet flow, with runon infiltration, which is infiltration of runoff onto pervious downstream areas. Parameters were derived from rainfall simulations and from watershed-scale calibration directly from the rainfall–runoff events. The performance of the water depth-dependent model was better than that of the conventional model on the scale of a rainfall simulator plot, but on the scale of a small watershed the performance of both model types was similar. We believe that the proposed model contributes to a less scale-dependent way of modeling runoff and erosion on the hillslope-scale.     

A stochastic model for tropical rainfall at a single location

Modelling data that correspond to rainfall accumulated over fixed periods of time presents the challenging problem of dealing with a random variable that has a point mass at zero which corresponds to dry periods that occur with positive probability. One way to overcome this difficulty is to assume that the data correspond to a normal variate w, that has been truncated and transformed. The dry periods correspond to the (unobserved) negative values and the wet periods correspond to some power of the positive ones. The serial structure that is present in rainfall can be modelled by imposing a serial structure to w. We use a dynamic linear model on w using a Fourier representation to allow for the seasonality of the data, which in the case of tropical rainfall is very marked. The model is fitted using a Markov chain Monte Carlo method that uses latent variables to handle both dry periods and missing values. We use the model to estimate and predict both the amount of rainfall and the probability of a dry period. The method is illustrated with data collected in the Venezuelan state of Guárico.     

Soil moisture and rainfall intensity thresholds for runoff generation in southwestern Wisconsin agricultural watersheds

The goal of this study was to improve understanding of the factors that influence runoff generation during non‐frozen ground periods in small agricultural watersheds in southwestern Wisconsin where the landscapes are controlled by dolostone bedrock in order to provide agricultural producers with a manure management tool. Six small watersheds (ranging from 6 to 17 ha) within two southwestern Wisconsin farm sites (Discovery Farms Program (DFP) and Pioneer Farm (PF)) were instrumented, and surface runoff was continuously monitored from 2004 to 2007. The soils in all watersheds were formed in deep (~1 m) loessial sites. A direct‐plant management strategy and corn‐soybean crop rotation were utilized within watersheds at DFP. A conventional tillage system (chisel plow in the fall followed by soil finisher in the spring) and a corn‐oat‐alfalfa crop rotation were utilized within watersheds at PF. At PF, the amount of precipitation leaving the landscape as surface runoff (1.8%) was two times greater compared to DFP (0.9%), indicating that the direct‐plant management system was better at retaining precipitation than the chisel plow/soil finisher system. Using breakpoint regression analysis, a non‐linear response in runoff generation with antecedent soil moisture (ASM) was observed with a threshold ASM of 0.39 cm³cm^?3 (approximately 80% of total porosity) for all six watersheds. Below this threshold, runoff coefficients were near zero. Above this threshold, runoff coefficients increased with ASM. A non‐linear response in runoff generation with maximum 30 min rainfall intensity (I30) was also observed, and threshold I30 values increased as ASM decreased and as crop cover increased. Copyright © 2012 John Wiley & Sons, Ltd.     

Combining a rainfall–runoff model and a regionalization approach for flood and water resource assessment in the western Po Valley,Italy

ABSTRACT

Selecting the best structure and parameterization of rainfall–runoff models is not straightforward and depends on a broad number of factors. In this study, the “Modello Idrologico Semi-Distribuito in continuo” (MISDc) was tested on 63 mountainous catchments in the western Po Valley (Italy) and the optimal model parameters were regionalized using different strategies. The model performance was evaluated through several indexes analysing hydrological regime, high-flow condition and flow–duration curve (FDC). In general, MISDc provides a good fit behaviour with a Kling-Gupta Efficiency index greater than 0.5 for 100% and 84% of cases for calibration and validation, respectively. Concerning the regionalization, spatial proximity approach is the most accurate solution obtaining satisfactory performance. Lastly, the predicted FDCs showed an excellent similarity with the observed ones. Results encourage to apply MISDc over the study area for flood forecasting and for assessing water resources availability thanks to the modest computational efforts and data requirements.     

Identification and stochastic generation of representative rainfall temporal patterns in Hong Kong territory

In hydrosystem engineering design and analysis, temporal pattern for rainfall events of interest is often required. In this paper, statistical cluster analysis of dimensionless rainfall pattern is applied to identify representative temporal rainfall patterns typically occurred in Hong Kong Territory. For purpose of selecting an appropriate rainfall pattern in engineering applications, factors affecting the occurrence of different rainfall patterns are examined by statistical contingency tables analysis through which the inter-dependence of the occurrence frequency of rainfall patterns with respect to geographical location, rainfall duration and depth, and seasonality is investigated. Furthermore, due to inherent variability of rainfall mass curves or hyetographs within each classified rainfall pattern, a practical procedure to probabilistically generate plausible rainfall patterns is described. The procedure preserves the inherent stochastic features of random dimensionless rainfall hyetograph ordinates, which in general are correlated non-normal multivariate compositional variables.     

The relationship between rainfall inputs and flood generation in south–east Spain

Rainfall and flood data are relatively sparse in semi‐arid areas; hence there have been relatively few investigations into the relationships between rainfall inputs and flood generation in these environments. Previous work has shown that flood properties are influenced by a combination of precipitation characteristics including amount, intensity, duration and spatial distribution. Therefore floods may be produced by high intensity, short duration storms, or longer duration, low intensity rainfall. Most of this research has been undertaken in small catchments in either hyper‐arid or relatively high rainfall Mediterranean climates. This paper presents results from a 6 year data record in south‐east Spain from research conducted in two basins, the Rambla Nogalte (171 km²) and the Rambla de Torrealvilla (200 km²). Data cover an area of approximately 500 km² and an annual average rainfall of 300 mm. At coarse temporal resolutions gauges spread over large areas record similar patterns of rainfall, although spells of rain show much more complexity; pulses of rain within storms can vary considerably in total rainfall, intensity and duration over the same area. The analysis for south‐east Spain shows that most storms occur over a period of less than 24 h, but that the number of rainfall events declines as the duration exceeds 8 h. This is at odds with data on floods for the study area suggesting that they are produced by storms lasting longer than 18 h. However, one flood event was produced by a very short (15 min) storm with high intensity rainfall. Most floods tended to occur in May/June or September, which coincides with wetter months of the year (September, October, December and May). Floods are also more highly related to the total rainfall occurring in a spell of rain, than to intensity. The complexity of storm rainfall increases with the storm total, which makes it difficult to generalize on the importance of rainfall intensity for flood generation. Copyright © 2007 John Wiley & Sons, Ltd.     

The applicability of a parsimonious model for local and regional prediction of runoff

Abstract

The Hydrological Recursive Model (HRM), a conceptual rainfall-runoff model, was applied for local and regional simulation of hourly discharges in the transnational Alzette River basin (Luxembourg-France-Belgium). The model was calibrated for a range of various sub-basins with a view to analysing its ability to reproduce the variability of basin responses during flood generation. The regionalization of the model parameters was obtained by fitting simultaneously the runoff series of calibration sub-basins after their spatial discretization in lithological contrasting isochronal zones. The runoff simulations of the model agreed well with the recorded runoff series. Significant correlations with some basin characteristics and, noticeably, the permeability of geological formations, could be found for two of the four free model parameters. The goodness of fit for runoff predictions using the derived regional parameter set was generally satisfactory, particularly for the statistical characteristics of streamflow. A more physically-based modelling approach, or at least an explicit treatment of quick surface runoff, is expected to give better results for high peak discharge.     

Seasonal evolution of runoff generation on agricultural land in the Belgian loess belt and implications for muddy flood triggering

Muddy floods due to agricultural runoff are a widespread and frequent phenomenon in the European loess belt, and particularly in central Belgium. These floods are triggered when high quantities of runoff are generated on cropland and cause severe erosion. Three soil surface characteristics are relevant to determine the runoff potential of cultivated soils: soil cover by crops and residues, soil surface crusting and roughness. These characteristics have been observed on 65 cultivated fields throughout 2005. A heavy rainfall event representative for events triggering muddy floods in the region (60 mm h^?1 during 30 minutes) has been simulated using a 0·5 m² simulator on fields with the 17 most observed combinations of soil surface characteristics in central Belgium. Runoff is not observed in the case of (ploughed) bare uncrusted soils, nor in the case of soils covered by crops showing a transitional crust and a moderate roughness (1–2 cm). In the cases where runoff has been observed, mean runoff coefficients ranged from 13% (wheat in July) to 58% (sugar beet or maize in May and June). Grassed buffer strips (GBSs) and grassed waterways (GWWs) show a higher runoff coefficient (62% for GBSs and 73% for GWWs) than most cultivated soils (13–58%). Furthermore, it is demonstrated that small plot measurements can be used to estimate runoff generation at the field scale. A classification of runoff generation risk based on the surveys of soil surface characteristics has been applied to common crops of central Belgium. February as well as the period between May and September are the most critical for runoff at the field scale. However, it appears from monitoring of a 16 ha catchment that the highest runoff volumes and peak discharges are recorded between May and August after heavy rainfall, explaining why 85% of muddy floods are recorded during this period in central Belgium. Copyright © 2007 John Wiley & Sons, Ltd.     

Analysis of surface roughness in relation to soil loss and runoff at high rainfall intensities

The decay of roughness is an important factor governing surface processes such as infiltration and soil erosion. Thus the decay of surface roughness under different surface conditions was investigated and related to quantitative amounts of soil loss, runoff and sediment concentration in a laboratory experiment. Rainfall with an intensity of 128 mm/h was applied to a bare or mulched surfaces of a sandy loam soil with known surface roughness at specified time intervals. The decay of roughness as expressed by roughness ratio, in this experiment, was better predicted when related to an exponential function of the square root of cumulative kinetic energy of rainfall rather than with the cumulative rainfall. The roughness decay equations in literature did not predict breakdown under mulched surfaces accurately. Thus the exponent parameters of the roughness decay equations were adjusted to reflect the reduced decay occurring under mulched surfaces. In a bare soil, regression equations expressing the dependent variables as a function of initial roughness index were significant, but with low coefficients of determination, being 0·39 for soil loss, 0·12 for runoff and 0·36 for sediment concentration. In addition to initial roughness index, cumulative kinetic energy of rainfall was further included in the regressions. This led to an increase in coefficients of determination, which was 0·81 for soil loss, 0·74 for runoff and 0·49 for sediment concentration. The coefficients of determination (0·87 for soil loss, 0·85 for runoff and 0·51 for sediment concentration) were further increased when the final roughness index was included in addition to initial roughness index and cumulative kinetic energy in the regressions. This work shows that soil loss and runoff could be predicted from bare soil surface provided the initial roughness and the energy of rainfall is known. However, field verifications of these relationships are needed under different tillage tools and under natural rainfall. Copyright © 2002 John Wiley & Sons, Ltd.     

Relationship between point rainfall, average sampled rainfall and ground truth at the event scale in the Sahel

Geostatistical techniques are used to quantify the reference mean areal rainfall (ground truth) from sparse raingaugenetworks. Based on the EPSAT-Niger event cumulative rainfall, a linear relationship between the ground truth considered as the mean area rainfall estimated from the densely available raingauge network and the area rainfall estimated from sparse network are derived. Also, a linear relationship between the ground truth and point rainfall is established. As it was reported experimentally by some authors, the slope of these relationships is less than one. Based on the geostatistical framework, the slope and the ordinate at the origin can be estimated as a function of the spatial structure of the rainfall process. It is shown that the slope is smaller than one. For the special case of one gauge inside a fixed area or a Field Of View (FOV), an areal reduction factor is derived. It has a limit value which depends only on the size of the area and the spatial structure of the rainfall process. The relative variance error of estimating the FOV cumulative rainfall from point rainfall is also given.     

Geochemical modeling of groundwater in karst area and its application at Pingdingshan coalfield

Several approaches including hydrogeochemistry and isotope hydrogeology have been used to investigate the karst groundwater systems at Pingdingshan coalfield in recent years. The results of the modeling and evaluation of groundwater chemistry, as parts of recent research progress at the area, are presented. The characteristics of Cambrian karst groundwater flow was analyzed in terms of tritium distribution of groundwater based on the fact that the Guodishan fault, the largest one within the coalfield, is divided into permeable (southern and northern) and impermeable (middle) sections. The evolution of groundwater chemistry, the suitability of geothermometers and the feature of karst development were deduced and discussed using the speciation modeling and mass balance approach. Project partially supported by China Postdoctoral Science Foundation.     

Identification of best predictors for forecasting seasonal rainfall and runoff in Australia

The first step towards developing a reliable seasonal runoff forecast is identifying the key predictors that drive rainfall and runoff. This paper investigates the lag relationships between rainfall across Australia and runoff across southeast Australia versus 12 atmospheric‐oceanic predictors, and how the relationships change over time. The analysis of rainfall data indicates that the relationship is greatest in spring and summer in northeast Australia and in spring in southeast Australia. The best predictors for spring rainfall in eastern Australia are NINO4 [sea surface temperature (SST) in western Pacific] and thermocline (20 °C isotherm of the Pacific) and those for summer rainfall in northeast Australia are NINO4 and Southern Oscillation Index (SOI) (pressure difference between Tahiti and Darwin). The relationship in northern Australia is greatest in spring and autumn with NINO4 being the best predictor. In western Australia, the relationship is significant in summer, where SST2 (SST over the Indian Ocean) and II (SST over the Indonesian region) is the best predictor in the southwest and northwest, respectively. The analysis of runoff across southeast Australia indicates that the runoff predictability in the southern parts is greatest in winter and spring, with antecedent runoff being the best predictor. The relationship between spring runoff and NINO4, thermocline and SOI is also relatively high and can be used together with antecedent runoff to forecast spring runoff. In the northern parts of southeast Australia, the atmospheric‐oceanic variables are better predictors of runoff than antecedent runoff, and have significant correlation with winter, spring and summer runoff. For longer lead times, the runoff serial correlation is reduced, especially over the northern parts, and the atmospheric‐oceanic variables are likely to be better predictors for forecasting runoff. The correlations between runoff versus the predictors vary with time, and this has implications for the development of forecast relationship that assumes stationarity in the historical data. Copyright © 2010 John Wiley & Sons, Ltd.     

Relationship between point rainfall, average sampled rainfall and ground truth at the event scale in the Sahel

Geostatistical techniques are used to quantify the reference mean areal rainfall (ground truth) from sparse raingaugenetworks. Based on the EPSAT-Niger event cumulative rainfall, a linear relationship between the ground truth considered as the mean area rainfall estimated from the densely available raingauge network and the area rainfall estimated from sparse network are derived. Also, a linear relationship between the ground truth and point rainfall is established. As it was reported experimentally by some authors, the slope of these relationships is less than one. Based on the geostatistical framework, the slope and the ordinate at the origin can be estimated as a function of the spatial structure of the rainfall process. It is shown that the slope is smaller than one. For the special case of one gauge inside a fixed area or a Field Of View (FOV), an areal reduction factor is derived. It has a limit value which depends only on the size of the area and the spatial structure of the rainfall process. The relative variance error of estimating the FOV cumulative rainfall from point rainfall is also given.     

一种新的综合物理指数构建及其在强降水预报中的应用

强降水是洪灾及相关衍生灾害的最主要原因之一,而过去单靠某一种变量诊断预报强降水,具有较大难度.本文在已有研究的基础上,根据强降水发生发展的物理机制,将引起降水的热力、动力和水汽条件综合考虑,尝试性地构建了一个新的综合指数THP(Temperature, Helicity and Precipitable water).然后针对两次强降水过程,利用NCEP/NCAR 1°×1°的再分析资料和地面常规观测资料,对THP指数进行了诊断分析,并选用2012年7月1日-8月15日的降水实况,对该指数进行了普适性检验.结果表明:(1) THP指数的变化可以有效表征强降水过程的发展和移动.对于降水落区的预报,THP指数的大值区与未来6 h的降水中心基本对应;对于降水发生时刻的预报,THP指数的位相变化超前于地面降水的变化,具有较好的指示性;(2) 对于高空槽前型降水,THP指数对降水强度也有一定的诊断意义,且普适性检验表明,该指数在我国中东部地区的盛夏期间具有良好的适用性;(3) 基于配料法的思想,THP指数将有利于强降水出现的、具有清晰物理意义的信号进行了集成,相比于表征单一物理量的指数,其稳定性得到了增强.     

Development of a portable field rainfall simulator for use in hillside vineyard runoff and erosion studies

An inexpensive, mobile field rainfall simulator and runoff plot frame were developed for use on hillside vineyards. The simulator framework and components were lightweight, readily available and easily manageable such that they can be handled by one person during transport, set–up and operation. The vineyard rainfall simulator was simpler than many of the machines in recent use for similar studies, yet offered equal or improved performance for small‐plot studies. The system developed consistent sized 2·58 mm raindrops at intensities ranging from 20 to 90 mm/h. The average distribution uniformity coefficient at an intensity of 60 mm/h was 91·7%, with a deviation of only 2·2%. This coefficient was similar to the range reported for a more complex rotating disk simulator, and was notably greater than that obtained for other similar devices. The system water capacity of 40 l allowed for 1‐h storm durations at 60 mm/h, usually sufficient time for commencement of erosion and runoff. The runoff plot frame was designed to be quickly installed, and to discourage sediment deposition in the routing of runoff to collect containers. Copyright © 2000 John Wiley & Sons, Ltd.     

Dynamic risk assessment model for flood disaster on a projection pursuit cluster and its application

Aiming at reducing the losses from flood disaster, a dynamic risk assessment model for flood disaster is studied in this article. This model is built upon the projection pursuit cluster principle and risk indexes in the system, proceeding from the whole structure to its component parts. In this study, a fuzzy analytic hierarchy approach is employed to screen out the index system and determine the index weight, while the future value of each index is simulated by an improved back-propagation neural network algorithm. The proposed model adopts a dynamic evaluation method to analyze temporal data and assesses risk development by comprehensive analysis. The projection pursuit theory is used for clustering spatial data. The optimal projection vector is applied to calculate the risk cluster type. Therefore, the flood disaster risk level is confirmed and then the local conditions for presenting the control strategy. This study takes the Tunxi area, Huangshan city, as an example. After dynamic risk assessment model establishment, verification and application for flood disasters between the actual and simulated data from 2001 to 2013, the comprehensive risk assessment results show that the development trend for flood disaster risk is still in a decline on the whole, despite the rise in a few years. This is in accordance with the actual conditions. The proposed model is shown to be feasible for theory and application, providing a new way to assess flood disaster risk.     

Use of a derived distribution approach for flood prediction in poorly gauged basins: A case study in Italy

A derived distribution approach is developed for flood prediction in poorly gauged basins. This couples information on the expected storm scaling, condensed into Depth Duration Frequency curves, with soil abstractions modeled using Soil Conservation Service Curve Number method and hydrological response through Nash’s Instantaneous Unit Hydrograph. A simplified framework is given to evaluate critical duration for flood design. Antecedent moisture condition distribution is included. The method is tested on 16 poorly gauged Mediterranean watersheds in Tyrrhenian Liguria, North Western Italy, belonging to a homogeneous hydrological regions. The derived flood distribution is compared to the regional one, currently adopted for flood design. The evaluation of Curve Number is critical for peak flood evaluation and needs to be carefully carried out. This can be done including local Annual Flood Series data in the estimation of the derived distribution, so gathering the greatest available information. However, Curve Number influence decreases for the highest return periods. When considerable return periods are required for flood design and few years of data are available, the derived distribution provides more accurate estimates than the approach based on single site distribution fitting. A strategy based on data availability for application of the approach is then given. The proposed methodology contributes to the ongoing discussion concerning PUB (Prediction in Ungauged Basins) decade of the IAHS association and can be used by researchers and practitioners for those sites where no flood data, or only a few, are available, provided precipitation data and land use information are at hand.