At-site Flood Frequency Analysis Research Articles

Goodness-of-fit, identifiability and extrapolation: Can the Two-Component Extreme Value distribution be used in at-site flood frequency analysis?

When fitting three-parameter flood frequency models to annual maximum (AM) flood series, the lack-of-fit can be mitigated by censoring potentially influential low flows (PILFs). An alternative and less-studied approach is to apply mixture probability models with four or more parameters, which trade off greater flexibility to fit AM series against the need to deal with degeneracy which arises when there is insufficient information to identify mixture components. However, the issue of degeneracy and the lack of a robust inference framework present a significant barrier to adoption. This study investigated the potential of the most parsimonious mixture model, the four-parameter Two-Component Extreme Value (TCEV) model. A Bayesian framework using Markov chain Monte Carlo sampling was developed to robustly characterize parameter uncertainty even in the presence of degeneracy. Two new posterior diagnostics based on the strength of the TCEV components were developed to aid identification of degeneracy. Armed with a robust inference framework, the study evaluated the potential of TCEV using a case study based on 31 catchments in eastern Australia with records exceeding 70 years. The evaluation used short to long records to compare TCEV and Log Pearson III fit and extrapolative uncertainty. The Log Pearson approach (LP3-PILF) censors PILFs following the approach described in Australian Rainfall and Runoff. With respect to goodness-of-fit, we found in most cases that TCEV fitted AM flood peaks well without the need to censor or stratify data, and overall, no clear difference emerged between TCEV and LP3-PILF. However, and contrary to expectation, TCEV produced high flow quantile confidence intervals consistently narrower than LP3-PILF even in the presence of degeneracy. While more case studies in different regions are needed to confirm the potential of TCEV, this study is a reminder that goodness-of-fit is a necessary but not sufficient criterion for selecting the probability model that best represents flood frequency.

A novel explainable PSO-XGBoost model for regional flood frequency analysis at a national scale: Exploring spatial heterogeneity in flood drivers

Identifying flood drivers and accurately estimating design floods play a crucial role in fostering sustainable and effective planning and management strategies for mitigating flood risks. Regional Flood frequency Analysis (RFFA) is one of the most commonly used approaches to estimate design floods in ungauged watersheds. This study used XGBoost coupled with Particle Swarm Optimization (PSO) to estimate different quantiles of the design floods with return periods (from 2-year to 1000-year). After a preliminary assessment, 373 nationwide hydrometric stations were selected to conduct at-site flood frequency analysis by identifying the best-fitting distribution. Using the capabilities of GIS and Google Earth Engine (GEE), 83 independent features including different physiographical, geomorphological, land-use, soil types, and long-term hydro-climatic and environmental variables were extracted for the upstream watersheds. After fine-tuning the hyper-parameters of the XGBoost method for each flood quantile, the feature importance values were used to eliminate the insignificant features and refine the developed models. Additionally, classical methods such as Support Vector Regression (SVR) and Random Forest (RF) were implemented, to evaluate the XGBoost models efficiency. Different statistics demonstrated that the models effectively estimated flood quantiles, with the Nash-Sutcliffe Efficiency (NSE) varying from 0.709 to 0.840 across all models. A comparison of model performance reveals that the XGBoost method outperformed RF and SVR across all flood quantiles. Based on the developed models, design floods have been estimated for 949 stations across Iran. Furthermore, the Shapley additive explanation (SHAP) values were used to identify the main contributing features to model outputs and investigate the spatial heterogeneity of main flood drivers. According to the results, the perimeter and length of the watershed and heavy rainfall exhibit notably high importance compared to other features for all models. Based on the local SHAP values, in Northern, Northwestern, and Western basins, features associated with watershed sizes, such as perimeter, area and length exhibit the highest levels of importance. Moreover, the Southwest basins are more influenced by “heavy rainfall”. These findings demonstrate the promise of the developed models for estimating flood quantiles across diverse environmental, geomorphological, and hydro-climatic conditions. This capability is valuable for sustainable watershed management, especially in environments with limited maximum discharge data.

Comparative Analysis of Flood Estimation using Log-Pearson Type III and Gumbel Max Models in the Cauvery River, India

Flooding is one of the most destructive global disasters in scale, geographical extent, property and life loss, and population displacement. The Cauvery River is one of the flood vulnerable rivers in the Peninsular region of India. At-site flood frequency analysis is performed using flow data obtained at the Kodumudi gauged site in the Cauvery River. Log Pearson Type III and Gumbel Max distribution models are used in the present study to estimate peak floods for different return periods. The Central Water Commission provides the annual maximum discharge for the Kodumudi gauged site over 39 years (1980-2018). The goodness of fit test employing the Kolmogorov-Smirnov and Anderson- Darling tests, reveals that Log-Pearson Type III best estimates peak floods in the study area. The peak floods predicted by Log-Pearson Type III for return periods 2, 5, 10, 25, 50, 100, 200, and 500 years are approximately 929, 1886, 2998, 5303, 8002, 11929, 17633, and 29228 cumecs. Hydraulic structures can be designed in the region based on 100-year flood. The present research could help with flooding management approaches, vulnerability analyses, and hydraulic structure design in the study region.

Alternate pathway for regional flood frequency analysis in data-sparse region

Accurately analyzing flood frequency is crucial for developing effective flood management strategies and designing flood protection infrastructure, but the complex and nonlinear nature of the hydrological system poses significant challenges. The use of advanced statistical techniques, hydrological modeling, and computational tools has facilitated the implementation of several regional flood frequency analysis (RFFA) methods, including machine learning (ML) approaches. However, such regional methods, being statistical and data-intensive with limited hydrologic inferences, can be challenging for regions with limited data. In this study, we evaluated an alternate approach or path for RFFA in two contrasting regions: a data-sparse region in India and a data-dense region in the USA. The alternate approach involves using deep learning (DL)-based model for streamflow prediction, followed by at-site flood frequency analysis to estimate flood quantiles. The proposed alternate approach was compared with the classical RFFA method, which utilized the Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) algorithms to establish the intricate relationship between different flood quantiles and predictor variables related to physio-meteorological conditions. The findings showed that the alternate approach outperformed the classical RFFA method in the data-sparse region of India, reducing the mean absolute error (MAE) and root mean squared error (RMSE) by approximately 50 %, and achieving a higher coefficient of determination (R2 = 0.85 to 0.96). In the data-dense region, the classical and alternate approaches produced comparable results. However, the alternate approach has the benefit of flexibility and offers a complete time series of daily flow at the ungauged watersheds, enabling the estimation of other flow attributes or magnitudes for any return period without needing a separate classical RFFA model. The study demonstrates that the alternate approach can produce reliable flood quantile estimates in regions with limited data, offering a promising solution for managing floods in these areas.

At-site flood frequency analysis in Brazil

At-site flood frequency analysis in Brazil

Regional Flood Frequency Analysis: A Bibliometric Overview

In water resources management, environmental and ecological studies, estimation of design streamflow is often needed. For gauged catchments, at-site flood frequency analysis is used for this purpose; however, for ungauged catchments, regional flood frequency analysis (RFFA) is the preferred method. RFFA attempts to transfer flood characteristics from gauged to ungauged catchments based on the assumption of regional homogeneity. A bibliometric analysis on RFFA is presented here using Web of Science (WoS) and Scopus databases. A total of 626 articles were selected from these databases. From the bibliometric analysis, it has been found that Journal of Hydrology and Water Resources Research are the two leading journals reporting RFFA research. In RFFA research, leading countries include Canada, USA, UK, Italy and Australia. In terms of citations, the top performing researchers are Ouarda T, Burn D, Rahman A, Haddad K and Chebana F. Future research should be directed towards the identification of homogeneous regions, application of efficient artificial intelligence (AI)-based RFFA models, incorporation of climate change impacts and uncertainty analysis.

Modelling accuracy for urban design flood estimation

ABSTRACT Management of flood risk remains a major problem in many urban environments. To generate the data needed for estimation of the flood risk, catchment models have been used with the reliability of the predicted catchment response for design flood estimation dependent upon the model calibration. However, the level of calibration required to achieve reliable design flood estimation remains unspecified. The purpose of this paper is to assess the event modelling accuracy needed if data from the calibrated model are to be used for continuous simulation of data for flood frequency analysis. For this purpose, a SWMM-based catchment model was investigated using 25 monitored events, while the assessment of the calibration was based on a normalised peak flow error. Alternative sets of parameter values were used to obtain estimates of the peak flow for each of the selected events. The best performing sets of these sets of parameter values were used with SWMM in a continuous simulation mode to predict flow sequences for extraction of Annual Maxima Series for an At-Site Flood Frequency Analysis. From the analysis of these At-Site Flood Frequency Analyses, it was concluded that the normalised peak flow error needed to be less than 10% if reliable design flood quantile estimates were to be obtained.

An evaluation of a multi-day rainfall – runoff volume – peak discharge transform for flood frequency estimation

ABSTRACT A new method to estimate design discharge quantiles is described based on converting multi-day rainfall P to flood event runoff RO, factored to generate discharge Q. The so-called PROQ transfer function is founded on simple flood volume-peak and GRADEX rainfall-runoff tanh relationships. Performance testing of PROQ, in both at-site and regional design flood contexts up to 1 in 100 annual exceedance probability, was made using south east Queensland streamgauge data. A statistical comparison against proven methods showed that the PROQ transform has significant potential as an alternative for design flood estimation. An example of how PROQ can be used within a design flood framework and recommendations for further enhancement are provided. Abbreviations: AEP: Annual exceedance probability; AMS: Annual maximum series, extracted from the flood record at a gauge site; ANOVA: Analysis of variance; ARR: Australian Rainfall and Runoff guidelines; A-S: At-site. Describes a set of methods to estimate design flood quantiles by statistical analysis of the flood record at an individual gauge site; E: Nash–Sutcliffe efficiency; FFA: Flood frequency analysis; G-B: Multiple Grubbs-Beck test recommended by ARR 2019 for low flow censoring. Used for at-site flood frequency analysis; GEV: General extreme value probability distribution; GRADEX: Gradient of extreme values. Design flood probability concept originating in France based on parallelism of rainfall and runoff quantile curves; L: Retention of rainfall within the catchment during flood event, expressed as a depth; LP3: Log Pearson 3 probability distribution; P: Rainfall depth; PRM: Probabilistic Rational Method. An ARR method for ungauged, undeveloped Australian catchments superseded in 2016; PROQ: Transfer function based on converting P to RO and then factoring RO to estimate Q; PW: Palmen and Weeks. Regional method for ungauged, undeveloped Queensland catchments developed by Palmen and Weeks (2011); R: Retention curve number. Used in probabilistic charting of design floods based on PROQ; RE: Absolute relative error; REG: Regional. Describes a set of methods to estimate design flood quantiles using information obtained from at-site analyses of several representative catchments within a region; RO: Flood event runoff depth; SR30: Strike rate of estimates within ±30% tolerance.

Identification of a preferred statistical distribution for at-site flood frequency analysis in Canada

Floods are the most frequent natural disaster in Canada, putting Canadian lives and property at risk. Projected variations in precipitation and temperature are expected to further intensify extreme events, necessitating improved flood planning and water resource management. The Natural Sciences and Engineering Research Council funded FloodNet project is developing a standardized flood estimation manual for Canada; such a nation-wide manual would make flood frequency application more effective, consistent, and reproducible, reducing the need for subjective judgement. This research investigates a preferred at-site flood frequency distribution for Canadian annual peak flow dataset. Four frequently used distributions: Generalized Logistic distribution, Generalized Extreme Value distribution, and Pearson Type III distributions with and without log transformation, are assessed using two robust goodness-of-fit tests (i.e., Modified Anderson-Darling test and L-Moment test) across a wide range of Canadian annual peak flood samples. Goodness-of-fit assessments are analysed using different configurations of flood samples, including annual maximum flow and instantaneous peak flow samples, stationary samples and de-trended non-stationary samples, samples with varied record length, and samples from various geographical regions of Canada. Estimated flood quantiles are compared with estimates of hypothetical true quantiles to assess predictive performance of the considered distribution. Results show the Generalized Extreme Value distribution is better than the other considered distributions because of its acceptance for most station samples and the closeness to the estimates of hypothetical true quantiles. This study provides a basis for recommending the Generalized Extreme Value distribution for at-site flood frequency analysis in Canada.

Selecting the best probability distribution for at-site flood frequency analysis; a study of Torne River

At-site flood frequency analysis is a direct method of estimation of flood frequency at a particular site. The appropriate selection of probability distribution and a parameter estimation method are important for at-site flood frequency analysis. Generalized extreme value, three-parameter log-normal, generalized logistic, Pearson type-III and Gumbel distributions have been considered to describe the annual maximum steam flow at five gauging sites of Torne River in Sweden. To estimate the parameters of distributions, maximum likelihood estimation and L-moments methods are used. The performance of these distributions is assessed based on goodness-of-fit tests and accuracy measures. At most sites, the best-fitted distributions are with LM estimation method. Finally, the most suitable distribution at each site is used to predict the maximum flood magnitude for different return periods.

Flood Frequency Analysis Using L Moments: a Comparison between At-Site and Regional Approach

Regional flood frequency analysis has been carried out for estimating peak discharge at regional level over the Kerala State, India, along with at-site flood frequency analysis. For the study, the annual peak discharges of 43 gauging stations having length of data from 14 to 47 years spread over the Kerala State were used. Using L moments and L moment ratio, best fit distribution was identified among the five distributions; Generalised Extreme Value (GEV), Generalised Pareto Distribution (GPA), Generalised Logistic (GLO), Generalised Normal (LN3) and Pearson Type III (PE3) distribution for both at-site and regional flood frequency analysis. Chi-square test, ranking method using statistical indicators and L moment ratio diagram were used for identifying the best fit distribution for at-site flood frequency analysis. It was found that GPA was the best fit distribution for 27 stations, GLO for 14, LN3 for 1 station and GEV for 2 station. After discordancy test, five different homogeneous regions were identified through heterogeneity test to carry out the regional flood frequency analysis (RFFA). After identifying best-fit distributions for each zone, flood growth curves were derived by incorporating the catchment characteristics of the basins. The distribution that was best fit in case of at-site analysis for a gauging site is found to be entirely different than the best fit distribution resulted from RFFA. In RFFA, growth curves that provide the flood magnitude for various return periods can be used to estimate the flood magnitude and frequency at ungauged sites in each region of the Kerala State in India.

At-Site Flood Frequency Analysis Coupled with Multiparameter Probability Distributions

Abstract The proper design of hydraulic structures as well as river basin management are directly dependent on adequate estimates of maximum streamflow, preferably obtained from long historical series. However, the scarce hydrological monitoring, recurrent in developing countries and the need for estimates associated with high return periods (RPs) have led to the use of estimation methods based statistical procedures, such as at-site flood frequency analysis. This study presents a framework for at-site flood frequency analysis coupled with multiparameter probability distribution functions (PDFs) (GEV, LN3, PE3, GLO, GPA, KAP and WAK), in which all the statistical procedures are derived from L-moments, in order to investigate the applicability of these PDFs in comparison to those of 2-parameters (EV1, LN2 and Gamma). The modeling framework was evaluated considering 106 maximum annual streamflow (MAS) series for the Rio Grande do Sul State - Brazil. PDFs’ goodness-of-fit was studied in accordance with the Anderson-Darling test. It can be concluded that: i) the multiparameter distributions, especially KAP and WAK, had performance superior to the traditional 2-parameter distributions, providing a greater number of historical series better adjusted by such multiparameter PDFs; ii) shorter series were usually better represented by GEV when compared to the other PDFs, which is an important characteristic when long historical series are not frequently available; and iii) the quantile estimates derived from multiparameter PDFs presented lower Relative Absolute Error, thus emphasizing the importance of using such PDFs in water resources management and engineering projects.

Flood Frequency Analysis at Ungauged Sites Based on Regionally Estimated Streamflows

Abstract Estimation of flood events at ungauged sites is often performed through regional flood frequency analysis (RFFA). RFFA uses the available information at gauged sites to estimate the desired design events at the ungauged site. These regional methods are based on a prior aggregation of the hydrological information at the gauged sites, which implies loss of information. In the present study, a different approach or path for conducting RFFA is presented. First, the daily streamflow series at the ungauged site is regionally estimated from daily information at the gauged sites through a regional flow duration curve approach. Then, a local flood frequency analysis is performed on the extracted maximum peak flow series. The proposed approach, referred to as regional streamflow-based frequency analysis (RSBFA), is applied to a case study in the province of Quebec, Canada. Results indicate that the performance of the RSBFA approach is comparable to traditional methods. However, the proposed method has the advantages of being simple, flexible, and of providing the whole daily streamflow series at the ungauged site, which allows the direct estimation of a large number of other flow characteristics, such as low-flow features. The RSBFA approach also avoids performing a complete at-site flood frequency analysis at each gauged site. The fact that all the regional information is included in the regionally estimated daily streamflow series implies a number of benefits: annual or seasonal, absolute or specific, stationary or nonstationary, and univariate or multivariate flood quantiles corresponding to any return period may then be obtained through the estimated series without reconducting a regional analysis.

Regional flood frequency analysis and prediction in ungauged basins including estimation of major uncertainties for mid-Norway

Abstract Study region 26 boreal catchments (mid-Norway). Study focus We performed regional flood frequency analysis (RFFA) using the L -moments method and annual maximum series (AMS) of mean daily streamflow observations for reliable prediction of flood quantiles. We used similarity in at-site and regional parameters of distributions, high flow regime and seasonality, and runoff response from precipitation-runoff models to identify homogeneous catchments, bootstrap resampling for estimation of uncertainty and regression methods for prediction in ungauged basins (PUB). New hydrological insights for the region The rigorous similarity criteria are useful for identification of catchments. Similarity in runoff response has the least identification power. For the PUB, a linear regression between index-flood and catchment area (R 2 = 0.95) performed superior to a power-law (R 2 = 0.80) and a linear regression between at-site quantiles and catchment area (e.g. R 2 = 0.88 for a 200 year flood). There is considerable uncertainty in regional growth curves (e.g. −6.7% to −13.5% and +5.7% to +24.7% respectively for 95% lower and upper confidence limits (CL) for 2–1000 years return periods). The peaks of hourly AMS are 2–47% higher than that of the daily series. Quantile estimates from at-site flood frequency analysis (ASFFA) for some catchments are outside the 95% CL. Uncertainty estimation, sampling of flood events from instantaneous or high-resolution observations and comparative evaluation of RFFA with ASFFA are important.

Improving at-site flood frequency analysis with additional spatial information: a probabilistic regional envelope curve approach

Extreme flood events have detrimental effects on society, the economy and the environment. Widespread flooding across South East Queensland in 2011 and 2013 resulted in the loss of lives and significant cost to the economy. In this region, flood risk planning and the use of traditional flood frequency analysis (FFA) to estimate both the magnitude and frequency of the 1-in-100 year flood is severely limited by short gauging station records. On average, these records are 42 years in Eastern Australia and many have a poor representation of extreme flood events. The major aim of this study is to test the application of an alternative method to estimate flood frequency in the form of the Probabilistic Regional Envelope Curve (PREC) approach which integrates additional spatial information of extreme flood events. In order to better define and constrain a working definition of an extreme flood, an Australian Envelope Curve is also produced from available gauging station data. Results indicate that the PREC method shows significant changes to the larger recurrence intervals (≥100 years) in gauges with either too few, or too many, extreme flood events. A decision making process is provided to ascertain when this method is preferable for FFA.

Regional flood frequency analysis: comparison of L-moment and conventional approaches for an Indian catchment

A realistic estimation of flood magnitude for a given return period is essential for successful operation and economical design of pivotal hydraulic structures such as dams, reservoirs and spillways, bridges, small culverts, urban drainage systems, flood plain zoning, economic evaluation of flood protection projects, etc. The flood frequency analysis is generally used for flood estimation, entirely based on the assumption that the floods are random and the floods in future are supposed to bear similar statistical properties that have occurred in the past. If adequate data are available at-site flood frequency analysis is used, otherwise regional flood frequency analysis (RFFA) is preferred for flood estimation. The RFFA for various small and medium catchments of India with different approaches have been carried out and reported in the literature. The focus of the present study is to carry out the RFFA using different approaches i.e. conventional method and L-moment and comparing the results for the Krishna and Penner subzone 3(h) of India. It was found that the values of growth factors obtained by the conventional method are higher than those of L-moment methods, thus leading to conservative estimate of the peak floods.

Flood estimation in Mahanadi river system, India using partial duration series

ABSTRACTFlood frequency analysis is a pre-requisite for setting up and safeguarding of many hydraulic structures, such as dams, barrages, check-dams, culverts and urban drainage systems. In the flood frequency analysis, partial duration series (PDS) may be considered when dealing with values exceeding certain limits causing floods. In fact, the PDS is capable of getting more information about extreme events than the annual maximum series (AMS). Additionally, an assumption that, the magnitude of the extreme events of a PDS is best described by a generalized Pareto (GP) distribution. The present work investigates the at-site flood frequency analysis to find the average number of peaks (λ) for modelling the PDS on the basis of the PDS/GP assumption and variability in the GP parameters coupled with the quantile estimation with an increase in the value of average number of peaks (λ) each year in the Mahanadi river system, Odisha, India. Also, to verify the PDS/GP assumption we tested seven different frequency distributions (Exponential, Gumbel, logistics, generalized extreme value (GEV), Lognormal (LN), generalized logistics (GL) and Pearson Type 3). Extensive daily discharge data collected from 23 gauging sites were used for the analysis. The results indicate precision and stability of GP distribution parameters for λ = 4 for almost all the discharge sites. The peak flood estimated for various return periods in the Mahanadi river system using GP distribution is endowed with high correlation statistics for this λ value.

Relevance of an at-site flood frequency analysis method for extreme events based on stochastic simulation of hourly rainfall

Extreme events are rarely observed, so their analysis is generally based on observations of more frequent values. The relevance of the flood frequency analysis (FFA) method depends on its capability to estimate the frequency of extreme values with reasonable accuracy using extrapolation. An FFA method based on stochastic simulation of flood event is assessed based on its reliability and stability. For such an assessment, different training/testing decompositions are performed for a set of data from more than 1000 gauging stations. We showed that the method enables relevant ‘predictive’ estimates, e.g. by assigning correct return periods to the record values that are systematically absent in calibration datasets. The model is also highly stable vis-a-vis the sampling. This characteristic is linked to the use of regional statistical rainfall data and a simple rainfall–runoff model that requires the calibration of only one parameter.Editor D. Koutsoyiannis Associate editor Q. Zhang

A framework for multivariate data-based at-site flood frequency analysis: Essentiality of the conjugal application of parametric and nonparametric approaches

Summary In watershed management, flood frequency analysis (FFA) is performed to quantify the risk of flooding at different spatial locations and also to provide guidelines for determining the design periods of flood control structures. The traditional FFA was extensively performed by considering univariate scenario for both at-site and regional estimation of return periods. However, due to inherent mutual dependence of the flood variables or characteristics [i.e., peak flow ( P ), flood volume ( V ) and flood duration ( D ), which are random in nature], analysis has been further extended to multivariate scenario, with some restrictive assumptions. To overcome the assumption of same family of marginal density function for all flood variables, the concept of copula has been introduced. Although, the advancement from univariate to multivariate analyses drew formidable attention to the FFA research community, the basic limitation was that the analyses were performed with the implementation of only parametric family of distributions. The aim of the current study is to emphasize the importance of nonparametric approaches in the field of multivariate FFA; however, the nonparametric distribution may not always be a good-fit and capable of replacing well-implemented multivariate parametric and multivariate copula-based applications. Nevertheless, the potential of obtaining best-fit using nonparametric distributions might be improved because such distributions reproduce the sample’s characteristics, resulting in more accurate estimations of the multivariate return period. Hence, the current study shows the importance of conjugating multivariate nonparametric approach with multivariate parametric and copula-based approaches, thereby results in a comprehensive framework for complete at-site FFA. Although the proposed framework is designed for at-site FFA, this approach can also be applied to regional FFA because regional estimations ideally include at-site estimations. The framework is based on the following steps: (i) comprehensive trend analysis to assess nonstationarity in the observed data; (ii) selection of the best-fit univariate marginal distribution with a comprehensive set of parametric and nonparametric distributions for the flood variables; (iii) multivariate frequency analyses with parametric, copula-based and nonparametric approaches; and (iv) estimation of joint and various conditional return periods. The proposed framework for frequency analysis is demonstrated using 110 years of observed data from Allegheny River at Salamanca, New York, USA. The results show that for both univariate and multivariate cases, the nonparametric Gaussian kernel provides the best estimate. Further, we perform FFA for twenty major rivers over continental USA, which shows for seven rivers, all the flood variables followed nonparametric Gaussian kernel; whereas for other rivers, parametric distributions provide the best-fit either for one or two flood variables. Thus the summary of results shows that the nonparametric method cannot substitute the parametric and copula-based approaches, but should be considered during any at-site FFA to provide the broadest choices for best estimation of the flood return periods.

Bias compensation in flood frequency analysis

Flood frequency analysis (FFA) is essential for water resources management. Long flow records improve the precision of estimated quantiles; however, in some cases, sample size in one location is not sufficient to achieve a reliable estimate of the statistical parameters and thus, regional FFA is commonly used to decrease the uncertainty in the prediction. In this paper, the bias of several commonly used parameter estimators, including L-moment, probability weighted moment and maximum likelihood estimation, applied to the general extreme value (GEV) distribution is evaluated using a Monte Carlo simulation. Two bias compensation approaches: compensation based on the shape parameter, and compensation using three GEV parameters, are proposed based on the analysis and the models are then applied to streamflow records in southern Alberta. Compensation efficiency varies among estimators and between compensation approaches. The results overall suggest that compensation of the bias due to the estimator and short sample size would significantly improve the accuracy of the quantile estimation. In addition, at-site FFA is able to provide reliable estimation based on short data, when accounting for the bias in the estimator appropriately.Editor D. Koutsoyiannis; Associate editor Sheng Yue