University of Twente Student Theses

Login

Statistical trend analysis of River discharge using a twentieth century weather re-analysis

Booij, D.G.R. (2017) Statistical trend analysis of River discharge using a twentieth century weather re-analysis.

[img]
Preview
PDF
2MB
Abstract:It is expected that river systems are nowadays changing due to an ascending surface temperature and the subsequent ascending number of extreme situation such as heavy precipitation and droughts. The River Rhine is probably also affected by these climate induced changes, which can cause higher river discharges according to experts. This in combination with the fact that the Netherlands is very vulnerable to floods, because of its low-lying land and the large number of densely populated areas, will cause a rise of the flood risk. To remain safe, it is important that flood protections, such as levees, are implemented close to all submersible areas. This process is time consuming and expensive and thence good policy must be developed. Therefore, it is important for Dutch water managers and policy makers to gain more certainty about the effects of climate change on river systems, such as the River Rhine. However, the outcomes of current river discharge trend analyses are not very certain. This is due to the insufficient quality and amount of data and especially in the River Rhine coincidence of the weather cannot be excluded as the cause for a visible or invisible trend. It is therefore important to try to take into account the uncertainties in the observed discharge time series in trend analyses. It has been tried to lower the uncertainties by using ten other possible discharge time series of the past in the trend analyses. These series are constructed by using a numerical weather product called ERA-CLIM in combination with a hydrological model to transform precipitation and temperature into a discharge at Lobith. ERA-CLIM concerns the re-analysed global weather in the period 1901 until 2010 only based on the average daily sea-surface temperature, sea-ice fraction and radiation scheme. These initial conditions of ERA-CLIM’s re-analysis are ten times slightly perturbed to obtain ten equally likely but different weather ensemble members. Variations in the ensemble members approach the uncertainties in the available observational sources on which the ERA-CLIM product is based and therefore they can represent the uncertainties in the observed discharge. Therefore, the objective of this study is as follows: The objective of this study is to find confirmation for average, seasonal and extreme trends in the River Rhine discharge at Lobith making use of discharge time series created with a hydrological model in combination with a numerical weather product. First, five types of annual discharge time series were created from the daily observed and ERA-CLIM discharge time series corresponding to five discharge characteristics; a mean discharge characteristic, two seasonal discharge characteristics and two extreme discharge characteristics. The annual discharge time series were also created for a daily time series created with a hydrological model in combination with observed input data, HYRAS discharge time series, to determine the influence of the hydrological model. After comparing both series was concluded that the influence of the hydrological model is negligible and that therefore the ERA-CLIM trends can directly be compared with the trends from the observed series. Thereafter, the reasonability of the ERA-CLIM discharge time series with respect to the observed series was investigated through a comparison. It turned out that the ERA-CLIM can represent the same kind of weather as observed when looking at the daily series. However, taking a closer look to the annual discharge time series it turned out that the ERA-CLIM series cannot represent the observed weather as same as good for all discharge characteristics. The ERA-CLIM ensembles can for example represent mean and high-flow season discharges quite good in contrast to for example the low-flow season discharges. However, it was concluded that ERA-CLIM can represent the observed series, when keeping in mind the differences between the series. Subsequently the independency of all ERA-CLIM series is tested. It turned out that the ERA-CLIM series are not dependent on each other and also not dependent on the observed series. Therefore, it is confirmed that the ERA-CLIM series can be treated as ten possible, independent discharge series. Subsequently, the trends of all discharge time series were calculated for the periods 1901 until 1951, 1951 until 2006 and 1901 until 2010. It turned out that the significances and direction of the trends found in the observed and HYRAS series are almost equal. Moreover, it was concluded that the trends and directions between the several ERA-CLIM series differ. The period for which the trends were calculated had also a major influence on the significance and direction of a trend. Finally, the calculated ERA-CLIM trends were compared with the trends found in the observed series. It turned out that most of the significances and directions of the trends found in the observed and ERA-CLIM series are very different. It could however be concluded that some trends in the observed series could be confirmed by the ERA-CLIM series. Although no trends in the extreme and mean series were confirmed making use of the ERA-CLIM series, the seasonal trends found in the observed series were confirmed making use of the ERA-CLIM series and therefore this study will provide circumstantial evidence for these seasonal trends in the observed series.
Item Type:Essay (Bachelor)
Faculty:ET: Engineering Technology
Subject:56 civil engineering
Programme:Civil Engineering BSc (56952)
Link to this item:http://purl.utwente.nl/essays/74658
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page