Flood Flow Frequency Research Articles

Stochastic Watershed Model Ensembles for Long‐Range Planning: Verification and Validation

AbstractDeterministic watershed models (DWMs) are used in nearly all hydrologic planning, design, and management activities, yet they cannot generate streamflow ensembles needed for hydrologic risk management (HRM). The stochastic component of DWMs is often ignored in practice, leading to a systematic bias in extreme events. Since traditional stochastic streamflow models used in HRM struggle to account for anthropogenic change, there is a need to convert DWMs into stochastic watershed models (SWMs) to generate ensembles for use in HRM. A DWM can be converted to an SWM using a post‐processing (pp) approach to add error to the DWM predictions. Many pp methods advanced in the area of flood forecasting are useful in HRM and for correcting extreme event biases. Selecting a suitable error model for pp is challenging due to nonnormality, skewness, heteroscedasticity, and autocorrelation. We develop a parsimonious pp method based on an autoregressive (AR) model of the logarithm of the ratio of the observations and simulations, which leads to AR model residuals that are approximately symmetric and independent. We document the value of pp for improving flood and low flow frequency analysis and we reintroduce the concepts of verification and validation of stochastic streamflow ensembles to ensure that the SWM can reproduce both statistics it was and was not designed to reproduce, respectively. These concepts are illustrated on a Massachusetts basin using the USGS Precipitation Runoff Modeling System, with an additional analysis indicating the approach may be applicable to 1,225 other sites across the United States.

Nonstationary flood and low flow frequency analysis in the upper reaches of Huaihe River Basin, China, using climatic variables and reservoir index as covariates

Conventional water infrastructure designs for flood and low flows are usually based on the assumption of stationarity of extreme events. However, recent evidence suggests that the influences of climate variability and human activities have made the hypothesis of stationarity questionable. In this study, we used the generalized additive models for location, scale, and shape (GAMLSS) to construct a nonstationary model in which the parameters of the selected distributions were modelled as a function of climatic variables (i.e., climate indices and precipitation) and/or the reservoir index (RI). The nonstationary models were then used to analyse annual flood and low flow frequency at four hydrological stations in the upper reaches of the Huaihe River Basin, including Dapoling (DPL), Changtaiguan (CTG), Zhuganpu (ZGP), and Xixian (XX) stations. Annual floods were represented by the maximum daily streamflow in each year, and low flows were represented by the 95th quantile of the daily streamflow (Q95) in each year. The change point and trend analysis revealed that the flood series of the ZGP station and the low flow series of the DPL and XX stations exhibited significant downward and upward trends (p < 0.1)), respectively. The low flow series of the ZGP station showed a significant change point in 1980 (p < 0.1). GAMLSS modelling results showed that, in comparison with stationary models, nonstationary models that included precipitation and the Arctic Oscillation climate index as covariates for the gamma distribution location parameter provided a superior description of the flood series at the four stations. Nonstationary models that incorporated precipitation and/or RI as covariates for the Weibull distribution parameters fit the low flow series better than stationary models at all stations. Furthermore, we found that nonstationary models outperformed stationary models in terms of flood frequency analysis, covering all flood observation points and capturing the generally decreasing trend in flood series, as well as a decrease in the scatter of estimated flood value magnitudes. For the low flow frequency analysis, the comparison results showed that the nonstationary and stationary models performed identically for the DPL, CTG, and XX stations, where no significant change point was detected. However, for the ZGP station, where a significant change point was detected, the nonstationary models performed better than the stationary models and could accurately capture the changes in the magnitude of the estimated low flow values before and after the change point. Overall, the proposed nonstationary model can serve as a tool for nonstationary frequency analysis of flood and low flow series under the influence of climate variability and reservoir regulations, thus providing a reference for regional water infrastructure design.

On the need for streamflow drought frequency guidelines in the U.S.

Abstract Extreme drought and resulting low streamflows occur throughout the U.S., causing billions of dollars in annual losses, detrimentally impacting ecosystems, as well as agricultural, hydropower, navigation, water supply, recreation, and a myriad of other water resource systems, leading to reductions in both the effectiveness and resiliency of our water resource infrastructure. Since 1966, with the introduction of Bulletin 13 titled ‘Methods of Flow Frequency Analysis’, the U.S. adopted uniform guidelines for performing flood flow frequency analysis to ensure and enable all federal agencies concerned with water resource design, planning, and management under flood conditions to obtain sensible, consistent, and reproducible estimators of flood flow statistics. Remarkably, over one-half century later, no uniform national U.S. guidelines for hydrologic drought streamflow frequency analysis exist, and the various assorted guidelines that do exist are not reliable because (1) they are based on methods developed for floods, which are distinctly different than low streamflows and (2) the methods do not take advantage of the myriad of advances in flood and low streamflow frequency analyses over the last 50 years. We provide a justification for the need for developing national guidelines for streamflow drought frequency analysis as an analog to the existing national guidelines for flood frequency analysis. Those guidelines should result in improved water resources design, planning, operations, and management under low streamflow conditions throughout the U.S. and could prove useful elsewhere.

Missing data representation by perception thresholds in flood flow frequency assessment

Flood flow frequency analysis (FFA) plays one of the key roles in many fields of hydraulic engineering and water resources management. The reliability of FFA results depends on many factors, an obvious one being the reliability of the input data - datasets of the annual peak flow. In practice, however, engineers often encounter the problem of incomplete datasets (missing data, data gaps and/or broken records) which increases the uncertainty of FFA results. In this paper, we perform at-site focused analysis, and we use a complete dataset of annual peak flows from 1931 to 2016 at the hydrologic station Senta of the Tisa (Tisza) river as the reference dataset. From this original dataset we remove some data and thus we obtain 15 new datasets with one continuous gap of different length and/or location. Each dataset we further subject to FFA by using the USACE HEC-SSP Bulletin 17C analysis, where we apply perception thresholds for missing data representation. We vary perception threshold lower bound for all missing flows in one dataset, so that we create 56 variants of the input HEC-SSP datasets. The flood flow quantiles assessed from the datasets with missing data and different perception thresholds we evaluate by two uncertainty measures. The results indicate acceptable flood quantile estimates are obtained, even for larger return periods, by setting a lower perception threshold bound at the value of the highest peak flow in the available - incomplete dataset.

Flood Flow Frequency Analysis to Estimate Potential Floodplain Nitrogen Treatment during Overbank Flow Events in Urban Stream Restoration Projects

Stream restoration for mitigation purposes has grown rapidly since the 1980s. As the science advances, some organizations (Chesapeake Bay Program, North Carolina Department of Environmental Quality) have approved or are considering providing nutrient credits for stream restoration projects. Nutrient treatment on floodplains during overbank events is one of the least understood processes that have been considered as part of the Chesapeake Bay Program’s Stream Restoration Nutrient Crediting program. This study analyzed ten years of streamflow and water quality data from five stations in the Piedmont of North Carolina to evaluate proposed procedures for estimating nitrogen removal on the floodplain during overbank flow events. The volume of floodplain flow, the volume of floodplain flow potentially treated, and the nitrogen load retained on the floodplain were calculated for each overbank event, and a sensitivity analysis was completed. On average, 9% to 15% of the total annual streamflow volume accessed the floodplain. The percentage of the average annual volume of streamflow potentially treated ranged from 1.0% to 5.1%. Annually, this equates to 0.2% to 1.0% of the total N load retained/removed on the floodplain following restoration. The relatively low nitrogen retention/removal rates were due to a majority of floodplain flow occurring during a few large events each year that exceeded the treatment capacity of the floodplain. On an annual basis, 90% of total floodplain flow occurred during half of all overbank events and 50% of total floodplain flow occurred during two to three events each year. Findings suggest that evaluating only overbank events may lead to undervaluing stream restoration because treatment is limited by hydrologic controls that restrict floodplain retention time. Treatment is further governed by floodplain and channel size.

Automatic Web Procedure for Calculating Flood Flow Frequency

The estimated flood flow frequency in a particular cross-section of a riverbed for a given return period is a topic of great interest for its application in hydrological, geomorphological and hydrogeological fields. Nevertheless, to establish a one-to-one relationship between rainfall and peak flow is a difficult problem to solve, due to the great number of factors involved (intensity and distribution of rainfall, hydromorphological characteristics of the watershed, type and distribution of vegetation, soil saturation conditions, etc.). In Italy, the Tiber River Basin Authority has developed a method to evaluate peak flows in the watersheds within the Tiber Basin. The relationship between rainfall depth with an assigned return period (RP) and the duration of the event was determined using data from 165 gauging stations throughout the Basin and in the neighbourhoods with respect to rainfall from 1 to 24 h and/or from 1 to 5 days. To calculate the peak flow with an assigned RP in small watersheds (area &lt; 100 km 2 ), the Tiber River Basin Authority proposed a methodology that combines the results of regional precipitation analysis of a duration from 1 to 24 h with the Curve Number method, which allows the volume of net rainfall (i.e., the rainfall that contributes to producing the peak flow) to be quantified. Such procedure includes the calculation of various parameters (run-off time, local rainfall and areal rainfall, net rainfall) in order to obtain the value of peak flow. To facilitate the use of this procedure, a WebGIS system has been developed, based on a series of scripts that calculate the values for the above parameters. The user only has to choose the point corresponding to the section of the channel in order to determine the peak flow and the return period. The computational procedure is performed using GRASS GIS that interfaces with the system using the standard WPS; the system returns to output a report with details of the various calculations of parameters and, as a final result, the value of requested peak flow.

Turkey Creek—A Case Study of Ecohydrology and Integrated Watershed Management in the Low-Gradient Atlantic Coastal Plain, USA

Water yield, water supply and quality, wildlife habitat, and ecosystem productivity and services are important societal concerns for natural resource management in the 21st century. Watershed-scale ecohydrologic studies can provide needed context for addressing complex spatial and temporal dynamics of these functions and services. This study was conducted on the 5240 ha Turkey Creek watershed (WS 78) draining a 3rd order stream on the Santee Experimental Forest within the South Carolina Atlantic Coastal Plain, USA. The study objectives were to present the hydrologic characteristics of this relatively undisturbed, except by a hurricane (Hugo, 1989), forested water-shed and to discuss key elements for watershed management, including water resource assessment (WRM), modeling integrated water resources management, environmental assessment, land use planning, social impact assessment, and information management. Runoff coefficients, flow duration curves, flood and low flow frequency curves, surface and ground water yields were assessed as elements of the WRM. Results from the last 10 years of interdisciplinary studies have also advanced the understanding of coastal ecohydrologic characteristics and processes, water balance, and their modeling including the need of high resolution LiDAR data. For example, surface water dynamics were shown to be regulated primarily by the water table, dependent upon pre- cipitation and evapotranspiration (ET). Analysis of pre- and post-Hugo streamflow data showed somewhat lower but insignificant (α = 0.05) mean annual flow but increased frequency of larger flows for the post-Hugo compared with the pre-Hugo level. However, there was no significant difference in mean annual ET, potentially indicating the resiliency of this coastal forest. Although the information from this study may be useful for comparison of coastal ecohydrologic issues, it is becoming increasingly clear that multi-site studies may be warranted to understand these complex systems in the face of climate change, sea level rise, and increasing development in coastal regions.

Implications of using on-farm flood flow capture to recharge groundwater and mitigate flood risks along the Kings River, CA.

The agriculturally productive San Joaquin Valley faces two severe hydrologic issues: persistent groundwater overdraft and flooding risks. Capturing flood flows for groundwater recharge could help address both of these issues, yet flood flow frequency, duration, and magnitude vary greatly as upstream reservoir releases are affected by snowpack, precipitation type, reservoir volume, and flood risks. This variability makes dedicated, engineered recharge approaches expensive. Our work evaluates leveraging private farmlands in the Kings River Basin to capture flood flows for direct and in lieu recharge, calculates on-farm infiltration rates, assesses logistics, and considers potential water quality issues. The Natural Resources Conservation Service (NRCS) soil series suggested that a cementing layer would hinder recharge. The standard practice of deep ripping fractured the layer, resulting in infiltration rates averaging 2.5 in d(-1) (6 cm d(-1)) throughout the farm. Based on these rates 10 acres are needed to infiltrate 1 cfs (100 m(3) h(-1)) of flood flows. Our conceptual model predicts that salinity and nitrate pulses flush initially to the groundwater but that groundwater quality improves in the long term due to pristine flood flows low in salts or nitrate. Flood flow capture, when integrated with irrigation, is more cost-effective than groundwater pumping.

Selecting the Best Flood Flow Frequency Model Using Multi-Criteria Group Decision-Making

Several criteria should be considered when selecting a probability distribution to describe hydrological data. This study examines how multiple criteria can be combined to make the best selection. Selection becomes more difficult and subjective when more than two criteria are used to determine the best distribution. Under these conditions, multi-criteria decision-making is necessary. In this study, 12 distributions were tested and compared for flood frequency analysis based on five selection criteria: root mean square error, Kolmogorov-Smirnov test statistic, relative average bias, modified Anderson-Darling test, and deviation in skewness and kurtosis. The comparisons are made using a multi-criteria group decision-making (MCGDM) based on ordered weighted averaging (OWA). Two preemptive goal programming models based on variance and entropy methods for a predetermined level of optimism by a group of decision-makers, determined the weight of the OWA operator. The model was applied to a case study of Mahabad River, a major river flowing into Lake Urmia in northwestern Iran. A sensitivity analysis of the results of MCGDM was done by changing the degree of optimism of the decision-makers. The results of sensitivity analysis showed the dependence of the rankings to the optimism degrees of the decision-makers. The 3-parameter Weibull (WBL3) and Pearson type 3 (PE3) distributions had the best results for Mahabad River flood data. A Monte Carlo simulation was conducted to test the ability of the models. It was concluded from the findings of the simulation that the PE3 distribution is the most appropriate for short sample sizes, while WBL3 is preferable for larger sample sizes.

A generalized Grubbs‐Beck test statistic for detecting multiple potentially influential low outliers in flood series

The Grubbs‐Beck test is recommended by the federal guidelines for detection of low outliers in flood flow frequency computation in the United States. This paper presents a generalization of the Grubbs‐Beck test for normal data (similar to the Rosner (1983) test; see also Spencer and McCuen (1996)) that can provide a consistent standard for identifying multiple potentially influential low flows. In cases where low outliers have been identified, they can be represented as “less‐than” values, and a frequency distribution can be developed using censored‐data statistical techniques, such as the Expected Moments Algorithm. This approach can improve the fit of the right‐hand tail of a frequency distribution and provide protection from lack‐of‐fit due to unimportant but potentially influential low flows (PILFs) in a flood series, thus making the flood frequency analysis procedure more robust.

Flood Flow Frequency Model Selection Using L-moment Method in Arid and Semi Arid Regions of Iran

Statistical frequency analysis is the most common procedure for the analysis of flood data at a gauged location thatin first step it is needed to select a model to represent the population. Among them, the central moment has been themost common and widely used, and with the using of computers, the application of the maximum likelihood hasincreased. This research was carried out in order to recognition of suitable probability distributions with perviouscommon methods. In order to investigate of suitable probability distribution for flood flow, using L-moment methodthrough the existing hydrometric stations in the region, 17 hydrometric stations were selected for peak dischargesdata studies. According to results of this research for peak discharge, LP3 distribution and ordinary moment method,P3 distribution and L-moment method, LN2 and LN3 distributions and ordinary method have been suitabledistinguished for 53%, 35%, 6% and 6% of stations, respectively. We concluded that L-moment method is suitable todetermine peak series probability distributions in the Iran central plateau and P3 is the best probability distribution formodeling peak series in this region.

Reconstructing extreme flows using an airflow index-based stochastic weather generator and a hydrological simulation model

This paper offers an alternative to the use of geomorphological and sedimentological evidence for the reconstruction of flood and low flow frequencies. It is based on a technique developed to estimate the hydrological impact of future climate change and it uses either observed or calculated meteorological parameters. It is possible to use this method directly without modification to ‘hindcast’ events within the period of regional meteorological records, which in areas like the UK extend back over the last 150 years. It should also be possible to extend the approach to earlier periods using reconstructed meteorological parameters based on surrogate data, such as ships' logs. The resulting extreme flow sequences may be used to study past hydrological regimes or to improve estimates of present-day risks by extending the flow records. The technique uses an airflow index-based stochastic weather generator to create hydrometeorological parameters to input into a physically based hydrological simulation model. The method is illustrated here in a reconstruction of daily flow series for the River Wye catchment above Rhayader, mid-Wales, for the period 1889–1998. The method makes use of observed climatic variables for the entire period, with the aim of capturing actual climate variability occurring over the 110-year period. Changes in high flow characteristics are assessed using the mean annual flood (MAF), Q 5 flow and peaks-over-threshold (POT) calculated from the 110-year simulated daily flow series. This particular application shows evidence of a possible trend towards increasing magnitude and frequency of high flow events, which, if continued, would have implications for flood and water resource management. Looking at evidence from the last 110 years helps to place possible future trends within the context of past variations in high flow extremes due to both natural and anthropogenically influenced fluctuations in climate.

Overtopping risk assessment in river diversion facility design

Diversion canals supplemented with upstream and downstream cofferdams are constructed to divert river flow for the construction of diversion weirs. Insufficient canal capacity leads to overtopping of water from the crest of the upstream cofferdam and sides of the diversion canal, which is regarded as the dominant risk mode in the system. A dynamic reliability model, which is based on a resistance-loading methodology with random independent loading following a Poisson process and random fixed resistance, may be used to assess the risk levels for various return periods and construction duration of a diversion weir. The system resistance is considered as the maximum canal capacity described by a lognormal probability density function. The river flow rate corresponding to a certain return period is considered as the external loading on the system. In a case study, a dynamic reliability analysis is performed for a diversion facility. From a flood flow frequency analysis, log-Pearson type 3 distribution is selected to describe the loading. The relationships between the total cost of the diversion facility, its reliability, safety, and duration of construction are examined in a decision making framework.Key words: dynamic reliability, diversion facility, resistance, loading.

Floodflow frequency model selection in Australia

Uniform flood frequency guidelines in Australia and the United States recommend the use of the log Pearson type 3 (LP3) distribution in flood frequency investigations. Many investigators have suggested alternate models such as the Generalized Extreme Value (GEV) distribution as an improvement over the LP3 distribution. Using floodflow data at 61 sites across Australia, we explore the suitability of various flood frequency models using L-moment diagrams. We also repeat the experiment performed in the original US Water Resource Council report (Bulletin 17B) which led to the LP3 mandate in the United States. Our evaluations reveal that among the models tested, the GEV and Wakeby distributions provide the best approximation to floodflow data in the regions of Australia that are dominated by rainfall during the winter months, such as southwest Western Australia and Tasmania. For the remainder of the continent, the Generalized Pareto (GPA) and Wakeby distributions provide the best approximation to floodflow data. The two- and three-parameter log-normal models and the LP3 distribution performed satisfactorily, yet not as well as either the GEV or GPA distributions. Other models such as the Gumbel, log-normal, normal, Pearson, exponential, and uniform distributions are shown to perform poorly. Recent research indicates that regional index-flood type procedures should be more accurate and more robust than the type of at-site procedures evaluated here. Nevertheless, this study reveals that index-flood procedures should not be restricted to a single distribution such as the GEV distribution because other distributions such as the GPA distribution perform significantly better in the most densely populated regions of Australia.

Thresholds and the spatial variability of flood power during extreme floods

Analysis of the effects of extreme events on overall landscape form has shifted geomorphic attention away from a focus on the magnitude and frequency relationship of discharge to a focus on flood power. This paper analyzes the variation in flood power both at-a-station and downstream in a basin of differing lithologies, and discusses the geologic and geomorphic controls on flood power. Design floods ranging from the 2-year discharge tothe 500-year discharge are routed trough a drainage basin, and channel boundary shear stress and unit stream power are calculated using HEC-2 along 21 cross valley transects. Because of the strong lithologic control on valley width and channel slope, an irregular downstream pattern of flood power emerges which reflects primarily the pattern in valley width. Narrow valleys correspond to a greater percent of channel conveyance during large magnitude floods and increasing rates of depth increases - thus maximizing flood power in these zones. The opposite occurs in anomalously wide valleys where flood power is minimized. A downstream spatial pattern emerges reflecting both local and general controls. For a given probability event, flood power varies at least threefold within the basin with a maximum occuring in a specific downstream section. Holding flood power constant produces an extremely wide range in flow probabilities. In order to establish a link between critical flood power magnitude and flow frequency, minimum flood powers are developed from an analysis of previous work on “catastrophic” flooding for humid, alluvial channels. These minimum thresholds tentatively correspond to 100 N/m 2 or 300 W/m 2 for shear stress and unit stream power, respectively. Discharges required to attain these critical flood powers vary from events roughly twice the 100-year flood to a maximum of roughly 18 times the discharge of the 100 year flood.

The value of streamflow record augmentation procedures in low-flow and flood-flow frequency analysis

Streamflow record augmentation procedures exploit the cross-correlation among streamflows at two or more streamgages to obtain improved estimates of the mean and variance of the flows at a short-record gage. Recent improvements in these procedures provide unbiased estimates of the mean and variance of the flows at the short-record site which always have equal or lower variance than simple at-site sample estimates. Essentially, record augmentation procedures increase the effective record length at a short-record site in proportion to the additional length of the nearby longer record and the cross-correlation of the concurrent streamflows at the two sites. An experiment documents the effective increase in record lengths on a site-by-site basis and on a regional basis using a network of 23 streamgages in or near Massachusetts. The increases in effective record lengths owing to the use of streamflow record augmentation procedures are substantial for the very-short-record sites for both flood-flow and low-flow statistics. However, the serial correlation associated with both flood-flow and low-flow sequences reduces those gains considerably.

Compensation flows in the UK: A hydrological review

AbstractA survey of United Kingdom reservoirs identified the development of compensation flow policy from the early 19th Century to the present day. Details on the level and pattern of releases below 261 impounding reservoirs provided the basis for assessing the spatial and temporal variation in compensation flows and the impact of impoundment on downstream flow regimes. By standardizing compensation releases by the natural pre‐impoundment mean discharge, release patterns from reservoirs with a wide range of both catchment area and climate could be compared. Analysis of the change in flood and low flow frequency following impoundment enabled the impact of reservoir regulation to be evaluated at the regional scale. For individual reservoirs the natural downstream increase in river flow provides a basis for determining the length of river reach which is sensitive to regime changes. The majority of present day compensation discharges were set to satisfy river interests which no longer apply or were based on inadequate hydrological or biological information. This has led to a wide range of release patterns and it is therefore recommended that the re‐evaluation of existing water resource schemes should include an assessment of the suitability of current release patterns for present day downstream demands and water resource requirements.

Hydrologic Study of Large Basin in Eastern Wyoming

A case study of the Horse Creek basin in eastern Wyoming is presented. The study involved calculating the 100‐year frequency and the probable maximum flood at several potential dam locations using a deterministic hydrologic model. The basin drains 1,560 sq mi. The basin has a long‐record gage near the point of interest, but the record is suspect due to irrigation diversion and return flows. Therefore, a flood flow frequency analysis from a hydrologically similar basin was performed. This included peak discharge and high volume‐duration‐frequency derivations. The model allows centering of a storm over a critical portion of the basin, with residual rainfall over the remainder. The study concludes that generation of rare‐event and probable‐maximum‐flood (PMF) design discharges at gaged locations requires careful analysis of the basin and gaged data. This includes review of hydrologically similar basins to allow generation of appropriate statistical parameters for calibration of a deterministic model.

The effect of discharge measurement error in flood frequency analysis

Observed hydrological data are usually assumed to be accurate when performing a flood-flow frequency analysis. A great deal of research has been conducted to determine flood frequency under this assumption. However, all flood data are subject to some measurement error and the effect of this error on frequency analysis is not well documented. There are only a few studies which have dealt with this subject and their conclusions are generally inconsistent. Consequently, there is a need for further study in this area. The major emphasis of this study was to demonstrate the effects of measurement error on the results of a flood flow frequency analysis. The Pearson Type III distribution was mainly used in the analysis, although the log-Pearson Type III was also used. To estimate the distribution parameters, the moment method was used as well as the curve-fitting method, which is the prescribed technique for parameter estimation in hydrological frequency analysis in China. Unbiasedness and efficiency of the 100 yr and 1000 yr design event estimates are taken as the major criteria for comparing the goodness of fit in frequency analysis. Results from the Monte Carlo method of generating sequentially synthetic hydrological records indicate that in the absence of historical outliers, the effects of small measurement errors are themselves small and may be neglected. When a historical outlier whose error parameter is smaller than 0.3 is included in the analysis, the resulting frequency curve is improved.

A Uniform Technique for Flood Frequency Analysis

In 1967 the U.S. Water Resources Council (WRC) published Bulletin 15 recommending that a uniform technique be used by all Federal agencies in estimating floodflow frequencies for gaged watersheds. This uniform technique consisted of fitting the logarithms of annual peak discharges to a Pearson Type III distribution using the method of moments. The objective was to adopt a consistent approach for the estimation of floodflow frequencies that could be used in computing average annual flood losses for project evaluation. In addition, a consistent approach was needed for defining equitable flood‐hazard zones as part of the National Flood Insurance Program. In 1976 WRC published Bulletin 17 which extended and updated Bulletin 15 but still recommended the use of the “log‐Pearson Type III” method. Since 1976, two updates of Bulletin 17 (17A and 17B) have been published which clarify or improve on this base method, or do both. This paper gives a brief historical review of the development of these bulletins and the motivation and justification for the adoption of this uniform technique. Special emphasis is given to Bulletin 17B, the current guidelines used by Federal agencies. Specific techniques examined are the development of regional skew, weighting of regional and station skew, the basis for the low‐ and high‐outlier tests, and the basis for the adjustment of frequency curves using historical information.