Low Flow Data Research Articles

Conversion to pine: Changes in timing and magnitude of high and low flows

Understanding watershed responses to extreme events is important for assessing potential impacts of floods, droughts, episodic pollution, and other external driving variables on watershed resources. In this study, we combine trend and frequency analyses with paired watershed techniques to evaluate the long-term high- and low-flow data from Coweeta Hydrologic Laboratory in North Carolina, USA in an attempt to quantify and interpret responses to extreme flow events in managed and unmanaged watersheds. Two experimental watersheds were converted from mixed deciduous forest cover to pine (Pinus strobus L.) in 1957 and 1956, respectively and two others were kept untreated to serve as control watersheds. Seventy years of annual streamflow (instantaneous maximum, minimum, and mean) time series data from watersheds 1, 2, 17, and 18 were analyzed with a Mann–Whitney–Pettitt test to identify and compare change points. Mean annual streamflow increased in both watersheds for 10–12 years after harvests, but the more effective factor on the flow series was the growth of pine plantations. Maximum flows of 2-, 5-, and 10-year return periods decreased after planting pine, but there was no difference for larger flow events compared to deciduous forest. For minimum flows, pine stands were more effective compared to maximum flows as minimum flows decreased for all return periods.

Frequency and intensity of extreme drought in the Delaware Basin, 1600–2002

The frequency and severity of drought in the Delaware Basin between 1600 and 2002 are examined using the Palmer Drought Severity Index (PDSI) estimated from tree ring data and correlated with reconstructed annual low flows. In the Delaware Basin, the most severe drought in nearly a century occurred during 1995–2002 as the Brandywine River, Delaware's largest surface water supply, ran dry at its mouth and declined to the lowest flow on record since 1912. To evaluate the long‐term context of the 1995–2002 droughts given a variable hydroclimate, tree ring and PDSI data were correlated to reconstruct flows along the river to 1600, the beginning of European exploration to the Delaware Bay. Reconstructed PDSI and low flows were fit using general extreme value (GEV) distributions to estimate drought frequency. Some variability is present as reconstructed low flows tend to overestimate recorded streamflow in severe dry years, a finding reported by others. Some uncertainty appears in the correlations as the coefficient of multiple determination (CRSQ) between recorded and estimated PDSI from tree ring data is 0.50–0.54, a level of variance considered to be “quite good,” and the coefficient of determination (r2) between PDSI and low flow is 0.52. Given the uncertainty, PDSI and reconstructed low flow data both agree that the most extreme drought in 400 years occurred during 1635, and the drought of 1995–2000 was historically extreme with differences only in the degree of severity. On the basis of PDSI, the 2002, 1999, and 1995 droughts were the sixth, twelfth, and seventeenth most severe in 400 years with frequencies of once every 50, 33, and 16 years, respectively. Based on low flow, the 2002, 1999, and 1995 droughts were the second, fourth, and ninth most severe since 1600 with frequencies of once every 200, 100, and 50 years, respectively. The record drought of 2002 has a low probability of reoccurring in any given year (2.0% by PDSI and 0.5% by low flow), but droughts nearly as severe have occurred during the 1630s, 1680s, 1820s, 1840s, 1860s, 1930s, 1940s, and 1960s. Increased intensities of drought low flows in Delaware during the late twentieth century through 2002 were coincident with population growth, watershed urbanization, and atmospheric warming although these associations were not correlated and further study is needed. Over 400 years of tree ring, PDSI, and reconstructed streamflow data indicate that the Delaware Basin record drought of 1995–2002 was a historically severe event with important implications for water supply and drought management. Droughts more severe than the record 2002 event have occurred in the past, and droughts may become even more intense should watershed urbanization and atmospheric warming continue in the future.

Signal-to-Noise Ratios of Geophysical and Environmental Time Series

Research Article| November 01, 2010 Signal-to-Noise Ratios of Geophysical and Environmental Time Series SEAN W FLEMING SEAN W FLEMING 1Generation Resource Management, BC Hydro, 6911 Southpoint Drive, Burnaby, BC V3N 4X8, Canada 1Also at: Geography Department, University of British Columbia, Vancouver, BC V6T 1Z4, Canada; phone: 604-528-2242; email: sean.fleming@bchydro.com. Search for other works by this author on: GSW Google Scholar Author and Article Information SEAN W FLEMING 1Also at: Geography Department, University of British Columbia, Vancouver, BC V6T 1Z4, Canada; phone: 604-528-2242; email: sean.fleming@bchydro.com. 1Generation Resource Management, BC Hydro, 6911 Southpoint Drive, Burnaby, BC V3N 4X8, Canada Publisher: Association of Environmental & Engineering Geologists Received: 01 Jun 2009 Revision Received: 01 Dec 2009 First Online: 02 Mar 2017 Online ISSN: 1558-9161 Print ISSN: 1078-7275 Copyright © 2010 EEGS Environmental & Engineering Geoscience (2010) 16 (4): 389–399. https://doi.org/10.2113/gseegeosci.16.4.389 Article history Received: 01 Jun 2009 Revision Received: 01 Dec 2009 First Online: 02 Mar 2017 Cite View This Citation Add to Citation Manager Share Icon Share Facebook Twitter LinkedIn MailTo Tools Icon Tools Get Permissions Search Site Citation SEAN W FLEMING; Signal-to-Noise Ratios of Geophysical and Environmental Time Series. Environmental & Engineering Geoscience 2010;; 16 (4): 389–399. doi: https://doi.org/10.2113/gseegeosci.16.4.389 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu toolbar search search input Search input auto suggest filter your search All ContentBy SocietyEnvironmental & Engineering Geoscience Search Advanced Search Abstract An observational time series can normally be viewed as consisting of two general components: one that we are interested in (signal), and another that we are not (noise) and that usually obscures the signal. For analysis of data sets capturing the temporal evolution of complex, open, multi-faceted systems—such as those routinely encountered in environmental and engineering geoscience—the operational definitions of signal and noise will frequently be extremely context-dependent, reflecting whatever happen to be the scientific interests of, or the societal questions posed to, a particular researcher. As such, either signal or noise may be mechanistic, random, or both in origin. Further, signal and noise may often be of comparable strength. In this paper, existing concepts are compiled and integrated to offer a flexible and general metric for the strength of an arbitrarily defined signal embedded within an observational time series. One potential application is the estimation of the relative importance or clarity of a specified physical process or effect. Brief, illustrative examples are drawn from geoscientific trend analysis. The results of these preliminary studies suggest that signal-to-noise ratios can be a useful, robust, and readily applied descriptive aid to support experimental characterization of geophysical patterns (as applied here to ecological low-flow data from the Cowichan region of British Columbia), exploration of fundamental relationships (suggesting a non-linear relationship between data set length and trend clarity on the basis of numerical experiments), and science communication (following from a straightforward analogy to consumer electronics). You do not have access to this content, please speak to your institutional administrator if you feel you should have access.

Determining Hydraulic Conductivity Using Pumping Data from Low‐Flow Sampling

Hydraulic conductivity values computed using the steady-state discharge and drawdown attained while low-flow sampling were evaluated to determine if they were equivalent to those determined from slug testing. Based on testing 12 wells, it was found that the results were statistically equivalent. Conductivity values computed using low-flow sampling parameters were also evaluated as to their reproducibility in actual practice by analyzing consultant data for three wells sampled over three quarterly monitoring periods by four field technicians. The results were found to be reproducible within about a factor of 2 or better. Since the method is based on only one pair of parameters, diligence is required in attaining steady state and in accurately measuring the flow rate and drawdown. Conductivity values computed using this approach can enhance the use of low-flow data gathered in water quality sampling, avoid the need for slug testing in a subsequent phase of investigation, and help reduce the cost of characterizing sites when multilevel samplers are used. Given the practical range of discharge in low-flow sampling, the method was found to be applicable at conductivity values somewhat greater than 10(-6) cm/s. Given the typical accuracy of water level meters and pressure transducers and a maximum discharge of 1 L/min, as mandated by regulatory guidance, the method has a calculated upper conductivity limit in the range of 10(-3) to 10(-2) cm/s.

Broadband trailing edge noise predictions in the time domain

A recently developed analytic result in acoustics, “Formulation 1B,” is used to compute broadband trailing edge noise from an unsteady surface pressure distribution on a thin airfoil in the time domain. This formulation is a new solution of the Ffowcs Williams–Hawkings equation with the loading source term, and has been shown in previous research to provide time domain predictions of broadband noise that are in excellent agreement with experimental results. Furthermore, this formulation lends itself readily to rotating reference frames and statistical analysis of broadband trailing edge noise. In the present work, Formulation 1B is used to calculate the farfield noise radiated from the trailing edge of a NACA 0012 airfoil in a low Mach number flow, using both analytical and experimental data on the airfoil surface. The acoustic predictions are compared with analytical results and experimental measurements that are available in the literature. Good agreement between predictions and measurements is obtained.

GEV-PWM Model for Distribution of Minimum Flows

Equations for the probability weighted moments (PWMs) of the generalized extreme value (GEV) distribution for minima are derived. The GEV-PWM model and five other models are fitted to the low flow data of 16 streams. The suitability of the models is checked by the probability plot correlation coefficient test. Weibull distribution fits the data better than lognormal. GEV-PWM has the best performance among the three-parameter models, but it frequently leads to infeasible lower bounds. The 10-year minimum flow estimates based on various models are determined. In the majority of cases (11 stations), differences between the largest and the smallest estimates are <20%. However, in a few cases, rather large differences (71%) are found. The estimates from the GEV-PWM and W3-MSO models are usually either the largest or smallest at a station.

Modeling restricted bivariate censored lowflow data

Modeling restricted bivariate censored lowflow data

Modeling restricted bivariate censored lowflow data

Environmental studies often result in censored data. In this article, the lowflow quantiles Q*7,2 and Q*7,10 below a limit are treated as censored data. These streamflow quantiles are important for water resources planning and management. Our partial all-subsets censored regression procedure identifies a few important explanatory variables, such as drainage area, basin slope, soil-infiltration index, rainfall index, and some combinations of them. The proposed maximum likelihood estimation method incorporates the restriction Q*7,2≥Q*7,10 and the bivariate probability distribution of the quantiles to improve model quality. Analyses of the lowflow quantiles obtained from streams in West-Central Florida show that our procedure is more appropriate than the commonly used univariate main-effects models in predicting quantiles. Copyright © 1999 John Wiley & Sons, Ltd.

Derivation of low flow distribution functions using recession curves

Recession analysis and low flow frequency analysis have been treated by tradition as separate problems, although both are related to low flow conditions. The shape and analytical expression of the streamflow decay during rainless periods might give an insight into the behaviour of the tails of low flow distribution functions. Furthermore, models for streamflow recession have, as a rule, a more or less straightforward physical background and this might be of help in the interpretation and regionalization of parameters of low flow distribution functions. The present paper is an attempt to combine results from these two separate fields in order to use recession equations to derive theoretical expressions for low flow distribution functions. The derived distribution approach together with the concept of compound distributions are used to deduce expressions for a family of low flow distributions related to linear and non-linear recession models. Approaches for parameter estimation are elaborated and exemplified on low flow data from the Nordic countries.

Application stochastique du phénomène d'étiage

The volume and duration of low-flow events in river channels are important characteristics for water management and the design of hydraulic structures. These extreme low-flow characteristics were studied separately, as identically distributed random variables, on a time interval of 1 year. A statistical analysis was performed on the low-flow data series of Lepreau River, station 01AQ001 (New-Brunswick, Canada). The observed characteristics were compared with the theoretical results using the Kolmogorov-Smirnov and (or) X 2 tests. It was generally concluded that the proposed model adequately fits the low-water phenomenon.

Frequency analysis of low flows using the Akaike information criterion

It is the current practice in frequency analysis of low flows to consider only three-parameter distributions in which one of the parameters represents a nonzero lower bound. When applied to the small samples typically available, this practice results in incorrect low flow estimates. These errors are related to errors in the estimated lower bound. To preclude this possibility, it is proposed that the current practice be changed to include the selection of a two-parameter distribution in certain situations. To assess this proposal, the Akaike information criterion (AIC) is used to compare the suitability of the most commonly used three-parameter distribution (three-parameter Weibull) and three two-parameter distributions (two-parameter Weibull, Gumbel, and lognormal) to low flow data for 51 long-term hydrometric stations across Canada. For 75% of the stations, a two-parameter distribution is selected over the three-parameter distribution if the selection criterion is minimum AIC. In about one third of the remaining 25% of the stations where the three-parameter Weibull distribution gave the minimum AIC, the estimated lower bound is sufficiently close to the minimum observed low flow to indicate overfitting and hence unreliable quantile estimates. When the AIC is supplemented with visual examination of goodness of fit on probability plots, it is found that the lognormal distribution could very well fit those cases where the AIC selected the three-parameter Weibull distribution. Key words: low flow frequency, goodness of fit, information criterion, probability plot.

NON‐ZERO LOWER LIMIT IN LOW FLOW FREQUENCY ANALYSIS?1

ABSTRACT: The implications of fitting distributions with non‐zero lower limits to low flow data are examined. The 3‐parameter Weibull distribution is fitted to annual minimum flow series from 60 long term stations in Canada. The relations between the estimated lower limit and three sampling variables (skewness, smallest, and largest observations) were investigated. The lower limit strongly depends on the sample skewness; it varies directly with the sample skewness, which in turn is highly influenced by the largest observation. For a given skewness, the value of the estimated lower limit is determined by the value of the smallest observation. Therefore, the lower limit cannot be accurately determined, and the resulting low flow estimates will be either too small or too high.

Nonparametric kernel estimation of low flow quantiles

A nonparametric kernel estimation model is proposed and developed for estimating low flow quantiles. Based on annual minimum low flow data and Monte Carlo simulation tests, the descriptive and predictive ability of the proposed model is compared with that of Weibull models. The results indicate that it has small bias and root mean square error in low flow quantile estimates. Application of the model to data from the Blue Nile has shown that the nonparametric approach is a viable alternative to the Weibull models.

Ambient Fluid Velocity Influences Proximal Isovelocity Surface Area Calculations

Background: Recent reports have described a simplified mass balance equation designed to predict volumetric flow rate through regurgitant and shunt orifices. The proximal isovelocity surface area (PISA) technique relies on the assumption that constant velocity surfaces (isovels) that are hemispheric in shape can be constructed proximal to the orifice, and flow passing through these surfaces must also pass through the orifice by conservation of mass. The method has been validated in simplified in vitro models, but these models have been symmetric in design and do not consist of ambient crossflows that are found in the beating heart. This study addressed the hypothesis that ambient crossflow will affect the validity of the PISA method. Methods and Results: An in vitro model was designed to test the effect of crossflowing streams on the PISA calculation while maintaining control of volumetric flow through the orifice. PISA predicted flows showed a good correlation but overestimated actual flows by 10%–110% (Y = 2.16X‐118, r = 0.96). Addition of low flow data slightly reduced the overestimation (Y = 1.89X‐55.2, r = 0.97). Only under high output state conditions (high orifice and crossing flows) was good agreement found (Y = 0.92X + 9.78, r = 0.95). Conclusions: The presence of crossflowing streams in the vicinity of PISA images will distort the converging flow and its image in such a way that the simplified PISA theory may correlate with true flow but only rarely produces true agreement. Only when competing factors fortuitously cancel does the PISA method appear to be accurate. Since PISA images are readily obtained in patients and since the method has received extensive attention in the literature, systematic derivation of correction factors accounting for crossflow is indicated. In the meantime, caution should be exercised when interpreting PISA images in the presence of crossflowing streams.

Discharge Data Collection and Analysis Strategies in Low Flow Studies

The objective of this paper is to determine the most appropriate data collection strategy and analysis techniques which should be used to assess the low flow regime of a catchment. The data used were: a) synchronous discharge measurements during low flow periods, and b) continuous daily flow records. The analyses based on both types of data were able to distinguish different low flow regimes within a 114 km2 Danish catchment. Despite the limited spatial variation in climate and geomorphology there was a high spatial variability in low flows caused by differences in the lithology of sediments. This demonstrates the difficulties in using simple indices of catchment geology in regional low flow estimation. The results highlight the benefits of using synchronous discharge measurements, both for estimating low flows at ungauged sites, and for understanding groundwater flow paths. Analyses of daily flow records from six gauging stations in the catchment showed that a baseflow index was more useful than the flow duration curve for classifying low flow regimes when only short records were available. The paper illustrates the importance of estimating the uncertainty of discharge measurements when interpreting low flow data.

Markov Chain Model for Seasonal-Water Quality Management

A nonstationary Markov chain approach to designing seasonal risk-equivalent water-quality management strategies is presented in this paper. Unlike many existing models for determining seasonal waste load allocations, this approach considers a wider range of possible low-flow conditions than those in the observed historical record. It explicitly incorporates the probabilistic relationships between low-flow data at adjacent gauging stations. The approach is demonstrated for the management of biochemical oxygen demanding (BOD) waste on the Willamette River in Oregon. The results of this application are compared with those of an existing model for seasonal risk-equivalent waste management. The treatment strategies derived for the Markov chain approach are similar, but slightly more conservative than those for the alternative program. Moreover, the ability of these treatment strategies to achieve an acceptable probability of water-quality violation is potentially less sensitive to the length of the historical record.

Frequency Estimation for Hydrological Samples with Zero Values

This paper develops an estimation method for discontinuous distributions used for fitting hydrologic series with several zero events. Such series are observed in arid and semiarid regions, and may include monthly precipitation in dry seasons, annual low flow, etc. The study consists of a derivation of the basic equation of the frequency distribution, derivation of distribution moments and parameters, plotting positions for positive values that form a part of the series instead of the full sample as conditional distribution, and transformation of the conditional distribution parameters into those of the marginal distribution. Using monthly precipitation and annual low-flow data from China, and annual maximum peak discharge data from the United States, the suitability of the distribution and the estimation method is assessed.

A hydrological regionalization of streams in Victoria, Australia, with implications for stream Ecology

Annual, monthly, low and peak flow data were used to classify and ordinate 138 stream gauges in Victoria. Sixteen hydrological variables were used and low-flow and entire-flow regionalizations were derived. The low-flow regionalization was spatially indistinct and therefore unusable, but the entire-flow regionalization produced five distinctive and spatially significant regions. Least-squares relationships were calculated between mean annual runoff, catchment area and coefficient of variation of annual flows, and the 16 variables. Rivers in the dry western districts of Victoria exhibit high variability of annual, monthly and peak flows, and low specific low flows. The converse is true for rivers in the western highlands of Victoria. Stream regionalizations are a useful tool for stream ecologists, and may be used for generating hypotheses, for detecting representative rivers and for producing baseline stream surveys.

On derivation of the extreme value (EV) type III distribution for low flows using entropy

ABSTRACT The extreme value type III distribution was derived by using the principle of maximum entropy. The derivation required only two constraints to be determined from data, and yielded a procedure for estimation of distribution parameters. This method of parameter estimation was comparable to the methods of moments (MOM) and maximum likelihood estimation (MLE) for the low flow data used.

Atraumatic rCBF measurement: an aid in the timing of surgery and the management of spasm following SAH.

We tested the ability of a non-invasive regional cerebral blood flow (rCBF) measurement to show vasospasm. The correlations between the flow values, the clinical condition, the CT scan and angiography were also studied. During none of the measurements were there any negative side effects on the patients. A bad clinical condition was usually accompanied by low flow data during the rCBF investigation. In accordance with local clinical signs (e.g., hemiparesis) we observed corresponding local hypoperfusion. These results were also confirmed by several follow-up studies in which good correlations between the rCBF findings, the angiography and the CT scan were found. As the rCBF investigation can be repeated within a very short time, we used this method for the testing of vasoactive drugs. Nimodipine (Bayer) showed promising results which suggest further clinical investigations. The main problem connected with the rCBF measurement is the limited ability of the patient to cooperate. For the detection of vasospasm the low resolution both in localisation and depth of this method is of only minor importance. We partly overcame this disadvantage by developing a special respirator enabling us to investigate artificially ventilated patients. We think that the measurement of rCBF is an important aid for the timing of the operation.