At-site Data Research Articles

Statistical and hydrodynamic numerical modeling to quantify storm surge hazard: Comparison of approaches applied to U.S. North Atlantic coast

Estimating the storm surge magnitude and annual exceedance probability is a key element in the siting and design of coastal nuclear power plants in both the U.S. and France. However, differences in storm climatology, specifically the relative importance of tropical cyclones (TCs) versus extratropical storms (XTCs), have driven differences in estimation method development. This work compares purely statistical modeling with combined statistical and numerical simulation modeling approaches for extreme storm surge applied to the U.S. North Atlantic coast which is subject to both tropical and extratropical storms. Two frequency analysis methods are applied to observed water levels and compared to a copula-based joint probability analysis of TCs and automated frequency analysis of XTCs that is enriched with numerically simulated storms. One frequency analysis method is applied using (1) hourly at-site data and (2) hourly at-site data enriched with additional data from a homogeneous region. The other frequency analysis method is applied using (1) hourly at-site data and (2) hourly at-site data enriched with monthly water level maxima. Variables of interest used in the comparison are skew storm surge, maximum instantaneous storm surge, non-tidal residual and maximum seal level. The performance of the methods (mean surge and water level estimates and confidence intervals) depend on the variable of interest and, to some extent, on return period. Inclusion of additional information (e.g., regional water levels, and monthly maxima) in the frequency analysis methods does not have a large impact on estimated mean surge and water levels, but significantly reduces resulting confidence intervals (over 40% reduction in some cases). However, the confidence intervals still grow with increasing return period. Inclusion of simulated storms in the joint probability analysis results in significantly different mean surge and water level estimates (up to 25% higher than the frequency analysis in some cases). The joint probability analysis confidence intervals are wider than those for the frequency analysis methods lower return periods (e.g., 60%–80% wider at 100 years), but they grow much more slowly and are significantly narrower for higher return periods (e.g., 40%–60% narrower at 1 000 years). Although there are appreciable differences between the results documented in this paper, these are reasonable due to differences in the data and methods used in this comparison.

Research on Nonpoint Source Pollution Assessment Method in Data Sparse Regions: A Case Study of Xichong River Basin, China

The NPS pollution is difficult to manage and control due to its complicated generation and formation mechanism, especially in the data sparse area. Thus the ECM and BTOPMC were, respectively, adopted to develop an easy and practical assessment method, and a comparison between the outputs of them is then conducted in this paper. The literature survey and field data were acquired to confirm the export coefficients of the ECM, and the loads of TN and TP were statistically analyzed in the study area. Based on hydrological similarity, runoff data from nearby gauged sites were pooled to compensate for the lack of at-site data and the water quality submodel of BTOPMC was then applied to simulate the monthly pollutant fluxes in the two sections from 2010 to 2012. The results showed agricultural fertilizer, rural sewage, and livestock and poultry sewage were the main pollution sources, and under the consideration of self-purification capacity of river, the outputs of the two models were almost identical. The proposed method with a main thought of combining and comparing an empirical model and a mechanistic model can assess the water quality conditions in the study area scientifically, which indicated it has a good potential for popularization in other regions.

A nonstationary index-flood technique for estimating extreme quantiles for annual maximum streamflow

Summary The magnitude and timing of peak streamflow events may be affected by land-use changes along with climate change, thus leading to nonstationarity in the records. Temporal trend, along with change-points, in peak flow records can affect the accuracy of quantile estimates; therefore, these issues should not be disregarded. Commonly used techniques for pooled flood frequency analysis do not account for nonstationarity found in the data recorded for members of a region. To overcome this shortcoming, the objective of this research is to introduce a trend centered pooling approach for regionalization in which pooling groups are created based on the form of trend found in the at-site data. The approach involves the formation of regions comprised entirely of sites exhibiting either statistically significant increasing or decreasing trends. Regional parameter estimates are determined using a maximum likelihood approach, which is carried out with the assumption of second-order nonstationarity. The technique was applied to four homogenous regions all located in differing hydroclimatological Canadian regions. The uncertainty of quantile estimates calculated through the implementation of this technique was established using a balanced regional vector resampling approach. The results indicate that there is less uncertainty in quantile estimates found through the application of the trend centered pooling approach when compared to a regional stationary analysis of the same regions. The potential for overestimation/underestimation of design quantiles in the presence of significant regional nonstationarity (i.e. decreasing/increasing trends) was elucidated.

A novel procedure for delineation of hydrologically homogeneous regions and the classification of ungauged sites for design flood estimation

Summary Regional flood frequency techniques are widely used to estimate flood quantiles when flow data is unavailable for the basin under study or the record length is insufficient for reliable analyses. Data from nearby gauged sites are pooled to compensate for the lack of at-site data. This requires the delineation of hydrologically homogeneous regions in which the flood regime is sufficiently similar to allow the spatial transfer of information. It is generally accepted that hydrologic similarity results from similarity among basins’ physiographic characteristics, and thus these characteristics can be used to delineate regions and classify ungauged sites. However, as currently practiced, the delineation is highly subjective and dependent on the similarity measures and classification techniques employed. Herein, a novel procedure for region delineation is proposed and evaluated using data for sites across the Southeastern United States. Key components of this procedure are a new statistical metric to identify physically discordant sites and a new methodology to identify the physical attributes that are the most indicative of extreme hydrologic response. The novel approach for region delineation is shown to produce regions which are more homogeneous and more efficient for quantile estimation at ungauged sites than those delineated using alternative physically-based procedures typically employed in practice. In addition, the identified physical attributes can be used to infer the flood regime and estimate quantiles at sites outside the extent of the area used for model development.

Examination of homogeneity of selected Irish pooling groups

Abstract. Flood frequency analysis is a necessary and important part of flood risk assessment and management studies. Regional flood frequency methods, in which flood data from groups of catchments are pooled together in order to enhance the precision of flood estimates at project locations, is an accepted part of such studies. This enhancement of precision is based on the assumption that catchments so pooled together are homogeneous in their flood producing properties. If homogeneity is assured then a homogeneous pooling group of sites lead to a reduction in the error of quantile estimates, relative to estimators based on single at-site data series alone. Homogeneous pooling groups are selected by using a previously nominated rule and this paper examines how effective one such rule is in selecting homogeneous groups. In this paper a study, based on annual maximum series obtained from 85 Irish gauging stations, examines how successful a common method of identifying pooling group membership is in selecting groups that actually are homogeneous. Each station has its own unique pooling group selected by use of a Euclidean distance measure in catchment descriptor space, commonly denoted dij and with a minimum of 500 station years of data in the pooling group. It was found that dij could be effectively defined in terms of catchment area, mean rainfall and baseflow index. The study then investigated how effective this selected method is in selecting groups of catchments that are actually homogenous as indicated by their L-Cv values. The sampling distribution of L-CV (t2) in each pooling group and the 95% confidence limits about the pooled estimate of t2 are obtained by simulation. The t2 values of the selected group members are compared with these confidence limits both graphically and numerically. Of the 85 stations, only 1 station's pooling group members have all their t2 values within the confidence limits, while 7, 33 and 44 of them have 1, 2 or 3 or more, t2 values outside the confidence limits. The outcomes are also compared with the heterogeneity measures H1 and H2. The H1 values show an upward trend with the ranges of t2 values in the pooling group whereas the H2 values do not show any such dependency. A selection of 27 pooling groups, found to be heterogeneous, were further examined with the help of box-plots of catchment descriptor values and one particular case is considered in detail. Overall the results show that even with a carefully considered selection procedure, it is not certain that perfectly homogeneous pooling groups are identified.

Regional Dry Spells Frequency Analysis by L-Moment and Multivariate Analysis

The spatial variation of the statistical characteristics of the extreme dry events, such as the annual maximum dry spell length (AMDSL), is a key practice for regional drought analysis and mitigation management. For arid and semi arid regions, where the data set is short and insufficient, the regionalization methods are applied to transfer at-site data to a region. The spatial variation and regional frequency distribution of annual maximum dry spell length for Isfahan Province, located in the semi arid region of Iran, was investigated using a daily database compiled from 31 rain gauges and both L-moment and multivariate analysis. The use of L-moment method showed a homogeneous region over entire province with generalized logistic distribution (GLOG) as the regional frequency distribution. However, the cluster analysis performed two regions in west and east of the province where L-moment method demonstrated the homogeneity of the regions and GLOG and Pearson Type III (PIII) distributions as regional frequency distributions for each region, respectively. The principal component analysis was applied on at-site statistics of AMDSL and found the L-coefficient of skewness (LCs) and maximum AMDSL the main variables describing the spatial variation of AMDSL over the Isfahan Province. The comparison of two homogeneous regions also proved the difference between two regions. Therefore, this study indicates the advantage of the use of multivariate methods with L-moment method for hydrologic regionalization and spatial variation of drought statistical characteristics.

Comparison of the annual minimum flow and the deficit below threshold approaches: case study for the province of New Brunswick, Canada

A low-flow frequency analysis using the annual minimum flow (AMF) and the deficit below threshold (DBT) approaches was carried out for 30 hydrometric stations across the province of New Brunswick. The AMF method considers only the annual minimum events, and the DBT method considers all low flows below a certain threshold (or truncation level). In the present study, the DBT method characterizes low flow in terms of deficit intensity, which is becoming increasingly important in both water and aquatic resources management. The annual minimum series were fitted using the three-parameter Weibull distribution (AMF–WEI3), and the intensity data series were fitted using the two-parameter Weibull distribution (DBT–WEI2) and the generalized Pareto distribution (DBT–GP). All parameter estimates were obtained using the maximum likelihood method. The AMF–WEI3 and DBT–GP approaches provided a good fit to at-site data in terms of annual minimum flow and deficit intensity, respectively. However, the present study showed that the DBT–WEI2 approach underestimated the right tail of low-flow intensity. The Anderson–Darling statistic was also calculated for the data series and can provide insight into which distribution may adequately model the low-flow characteristics. A regionalization study was also performed using the AMF–WEI3 and DBT–GP methods.

A methodology for the estimation of short duration design storms in South Africa using a regional approach based on L-moments

The network of recording raingauges in South Africa with long records is relatively sparse and only 55 out of a total of 412 raingauges currently have record lengths longer than 30 years. In addition, much of the digitised rainfall data from numerous stations are viewed as being unreliable, with important events missing and errors in the digitised data. In an attempt to overcome data limitations and to improve reliability, a regional, index storm approach based on L-moments has been applied to estimate short duration (≤24 h) design rainfalls in South Africa. The regionalisation was performed using only site characteristics (latitude, longitude, altitude, concentration of precipitation, mean annual precipitation (MAP), rainfall seasonality and distance from sea) and required a minimum of subjective relocation of stations in order to obtain relatively homogeneous clusters. The at-site data were used for independent testing of the clusters of the stations for homogeneity. Using information for 172 rainfall stations which have at least 10 years of short duration rainfall records, 15 relatively homogeneous clusters were identified in South Africa. The index storm used was the mean of the annual maximum series and a relationship was derived between the index storm and MAP. This methodology enables short duration design storms to be estimated at ungauged sites in South Africa. The accuracy of design storms estimated using this technique was assessed at a site not used in the regionalisation procedure and error bounds were computed for the quantile growth curves for each of the clusters.

Adjustment of Regional Regression Equations for Urban Storm-Runoff Quality Using At-Site Data

Regional regression equations have been developed to estimate urban storm-runoff loads and mean concentrations using a national data base. Four statistical methods using at-site data to adjust the regional equation predictions were developed to provide better local estimates. The four adjustment procedures are a single-factor adjustment, a regression of the observed data against the predicted values, a regression of the observed values against the predicted values and additional local independent variables, and a weighted combination of a local regression with the regional prediction. Data collected at five representative storm-runoff sites during 22 storms in Little Rock, Arkansas, were used to verify, and, when appropriate, adjust the regional regression equation predictions. Comparison of observed values of storm-runoff loads and mean concentrations to the predicted values from the regional regression equations for nine constituents (chemical oxygen demand, suspended solids, total nitrogen as N, total ...

Adjustment of Regional Regression Equations for Urban Storm-Runoff Quality Using At-Site Data

Regional regression equations have been developed to estimate urban storm-runoff loads and mean concentrations using a national data base. Four statistical methods using at-site data to adjust the regional equation predictions were developed to provide better local estimates. The four adjustment procedures are a single-factor adjustment, a regression of the observed data against the predicted values, a regression of the observed values against the predicted values and additional local independent variables, and a weighted combination of a local regression with the regional prediction. Data collected at five representative storm-runoff sites during 22 storms in Little Rock, Arkansas, were used to verify, and, when appropriate, adjust the regional regression equation predictions. Comparison of observed values of storm-runoff loads and mean concentrations to the predicted values from the regional regression equations for nine constituents (chemical oxygen demand, suspended solids, total nitrogen as N, total ammonia plus organic nitrogen as N, total phosphorus as P, dissolved phosphorus as P, total recoverable copper, total recoverable lead, and total recoverable zinc) showed large prediction errors ranging from 63 percent to more than several thousand percent. Prediction errors for 6 of the 18 regional regression equations were less than 100 percent and could be considered reasonable for water-quality prediction equations. The regression adjustment procedure was used to adjust five of the regional equation predictions to improve the predictive accuracy. For seven of the regional equations the observed and the predicted values are not significantly correlated. Thus neither the unadjusted regional equations nor any of the adjustments were appropriate. The mean of the observed values was used as a simple estimator when the regional equation predictions and adjusted predictions were not appropriate.

Statistical characteristics of U.S. Historical Climatology Network temperature distributions

The U.S. Historical Climatology Network (HCN) adjusted data are designed to aid in the investigation of climate change. The primary goal in the development of the adjustment procedures was to produce temporally homogeneous at-site data, with respect to averages, that could be used in the investigation of regional, long-term trends. The data, however, have been (and likely will continue to be) used for applications other than the detection of climate trends, and some of these other uses involve statistical properties of the data distributions. This study examines the change in distributional characteristics that result from the site location and urbanization adjustments to maximum and minimum temperature data in the HCN. It is shown that at most locations distributional shapes are changed when the data are adjusted. For climate change studies the differences may not be important, but for other studies involving distributional characteristics they could be important. In particular, probabilistic assessments of unusual climatic conditions that are used for design and planning purposes are highly dependent upon the shape of the frequency distributions. Changing the skewness and kurtosis, I.e. the shape of a frequency distribution, often changes the probability densities in the tails of the distribution, and therefore impacts the probabhstic assessments and the subsequent decisions that are based on these assessments.

Application of the Bayesian approach in regional analysis of extreme rainfalls

Based on the Partial Duration Series model a regional Bayesian approach is introduced in the modelling of extreme rainfalls from a country-wide system of recording raingauges in Denmark. The application of the Bayesian principles is derived in case of both exponential and generalized Pareto-distributed exceedances. The method is applied to, respectively, the total precipitation depth and the maximum 10 minutes rain intensity of individual storms from 41 stations. By means of the regional analysis prior distributions of the parameters in the Partial Duration Series model are estimated. It is shown that the regional approach significantly reduces the uncertainty of the T-year event estimator compared to estimation based solely on at-site data. In addition, the regional approach provides quantile estimates at non-monitored sites.