The prediction of reservoir sedimentation across England and Wales
The principal aim of this thesis is to produce a comprehensive assessment of the spatial
variation of reservoir sedimentation for the population of surface water storage reservoirs,
exhibiting a degree of catchment-coupling, across England and Wales. This is the first time
that such a study has been undertaken. To achieve this, a valid model had to be developed
to predict reservoir sedimentation for those sites where no direct measurements were
available. Using readily available secondary data describing catchment and reservoir basin
factors identified to influence reservoir sedimentation, two lumped-empirical models were
developed. The first model used the traditional modelling technique of multiple regression;
the second used fuzzy set multi-criteria evaluation (MCE), a technique not previously
applied to reservoir sedimentation modelling. Both techniques were applied to a population
data set (established herein). The models were validated using primary data from
Through multiple regression, models predicting reservoir annual percentage capacity loss
(APL) could not be verified and validated due to violations of the assumptions of the
technique. A major reason for these failures was the inability of the approach to account for
uncertainty and error associated with the quantification of the dependent and independent
variables, and the specification of the respective relationships.
Through fuzzy set MCE, a model predicting the degree of reservoir sedimentation for the
population data set was verified and validated to a reasonably good level; as was an
associated defuzzification model converting the predicted fuzzy number to a predicted APL
rate. The fuzzy set MCE modelling technique was identified as the most suitable for
accomplishing the principal aim because, unlike multiple regression, it was able to account
for uncertainty and error associated with the quantification and specification of the
catchment and reservoir basin factors that influence reservoir sedimentation.
Upon operationalization, the fuzzy set MCE model and associated defuzzification model
identified the upland areas of Cumbria, Wales, the southern Pennines and the southwest as
experiencing the largest levels of reservoir sedimentation. This is primarilly attributed to
these areas having high mean annual rainfall and large areas of highly erosive soil types.
The outputs of the fuzzy set MCE model and associated defuzzification model allow
attention to be focused by reservoir undertakers on areas where sustainability of water
supply may be most sensitive to loss of reservoir capacity from sedimentation.