Abstract
The relationship between extreme events and their return periods relates to the design of structures and systems which often rely on an estimated design event X{sub T}, assessed to occur once per T years on average. In order to make inferences about extreme events corresponding to design return periods beyond the time range of observed data, this computer programme EVA (Extreme Value Analysis) was developed. It is based on a frequency analysis approach. The connection between the sample data and design events is obtained by use of a theoretical probability distribution as a model for the extreme events. EVA facilitates and improves the analysis of extreme events originating from time series such as precipitation, runoff, wind, wave, currents. For the estimation of T-year events and associated confidence limits EVA provides verification of inherent assumptions, provision of alternative theoretical distributions as candidates for representing the population of events, estimation of statistical parameters of a selected distribution, appropriateness of fit tests for evaluating the distribution hypothesis, estimation of T-year events, X{sub T} for accepted candidate distributions, assessment of the uncertainty of X{sub T} caused by a limited sample, and confidence limits of data sample. Relevant statistical issues, as an aid to interpretation
More>>
Citation Formats
None.
EVA. A programme for extreme value analyses.
Denmark: N. p.,
1991.
Web.
None.
EVA. A programme for extreme value analyses.
Denmark.
None.
1991.
"EVA. A programme for extreme value analyses."
Denmark.
@misc{etde_10148428,
title = {EVA. A programme for extreme value analyses}
author = {None}
abstractNote = {The relationship between extreme events and their return periods relates to the design of structures and systems which often rely on an estimated design event X{sub T}, assessed to occur once per T years on average. In order to make inferences about extreme events corresponding to design return periods beyond the time range of observed data, this computer programme EVA (Extreme Value Analysis) was developed. It is based on a frequency analysis approach. The connection between the sample data and design events is obtained by use of a theoretical probability distribution as a model for the extreme events. EVA facilitates and improves the analysis of extreme events originating from time series such as precipitation, runoff, wind, wave, currents. For the estimation of T-year events and associated confidence limits EVA provides verification of inherent assumptions, provision of alternative theoretical distributions as candidates for representing the population of events, estimation of statistical parameters of a selected distribution, appropriateness of fit tests for evaluating the distribution hypothesis, estimation of T-year events, X{sub T} for accepted candidate distributions, assessment of the uncertainty of X{sub T} caused by a limited sample, and confidence limits of data sample. Relevant statistical issues, as an aid to interpretation and use of results, and guidelines on operation are given. (AB).}
place = {Denmark}
year = {1991}
month = {Oct}
}
title = {EVA. A programme for extreme value analyses}
author = {None}
abstractNote = {The relationship between extreme events and their return periods relates to the design of structures and systems which often rely on an estimated design event X{sub T}, assessed to occur once per T years on average. In order to make inferences about extreme events corresponding to design return periods beyond the time range of observed data, this computer programme EVA (Extreme Value Analysis) was developed. It is based on a frequency analysis approach. The connection between the sample data and design events is obtained by use of a theoretical probability distribution as a model for the extreme events. EVA facilitates and improves the analysis of extreme events originating from time series such as precipitation, runoff, wind, wave, currents. For the estimation of T-year events and associated confidence limits EVA provides verification of inherent assumptions, provision of alternative theoretical distributions as candidates for representing the population of events, estimation of statistical parameters of a selected distribution, appropriateness of fit tests for evaluating the distribution hypothesis, estimation of T-year events, X{sub T} for accepted candidate distributions, assessment of the uncertainty of X{sub T} caused by a limited sample, and confidence limits of data sample. Relevant statistical issues, as an aid to interpretation and use of results, and guidelines on operation are given. (AB).}
place = {Denmark}
year = {1991}
month = {Oct}
}