Back to the programme printer.gif Print

Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Wind speed predictions: Are we at the limit of our knowledge or can we improve?' taking place on Wednesday, 12 March 2014 at 11:15-12:45. The meet-the-authors will take place in the poster area.

Clive Wilson Meteorological Office, United Kingdom
Clive Wilson (1) F P Graeme Candlish (1)
(1) Meteorological Office, Exeter, United Kingdom (2) Universidad de Concepción, Concepción, Chile

Printer friendly version: printer.gif Print


Wind trends and variability from reanalyses and their validation using conventional and elevated wind observations


Several global reanalyses are now being used in wind resource assessments. In a few years regional higher resolution data sets will become available. We have shown that direct use of the reanalysed winds at sites in the UK are inferior to mesoscale model and downscaled site-specific winds when compared against mast observations. A prototype regional reanalyses for Europe shows better results. We now examine trends and interannual variability of the reanalyses in Europe and validate them against conventional observations, gridded long- term UK observations and high quality research mast observations.


Horizontally and vertically interpolated reanalysis winds are compared to elevated mast observations at over 100 sites in the UK and Europe. Long term trends and variations in the reanalyses at individual sites, for a 5km grid over the UK and country-sized areas are derived and validated against stable and well-maintained long-term conventional meteorological observations. We attempt to differentiate between trends and variability which have observational support and likely spurious ones; we also advance possible reasons for these. Pitfalls in the use of reanalyses are exposed and we attempt to relate the changes in variability to changes in atmospheric circulation.

Main body of abstract

Several reanalyses data sets, including a new European regional reanalysis from the EURO4M programme, have been evaluated as a direct proxy for conventional wind observations in resource assessment. We have previously identified the added value of additional stages in the modelling chain from 1) downscaling using a higher resolution mesoscale model (the MetUM) and 2) further post-processing to adjust to local surface and orography.

Although the EURO4M is an improvement on the global datasets in terms of accuracy, doubts still remain that reanalyses contain spurious trends and variability arising from changes in observation types, inhomogeneous network coverage and different observational practice and biases. We have adopted a number of approaches to answer this question: intercomparing the reanalyses, looking for regional differences; comparison with a homogenised 43 year UK gridded climatology; comparison against wind observations from long term surface stations and research masts.

Problems of changes in exposure, instrument or location of some meteorological stations makes clear identification of trends and variability difficult or impossible at some sites. Excluding these we show that generally reanalyses have overstrong wind speeds over land and appreciable differences in mean annual winds which strongly influence percentage changes in mean annual wind variations with large disagreement between reanalyses. Observations also have large differences in percentage variation because of wide variation in the mean speed at sites. Standard deviation is a better indicator of variability and is more coherent and closely related to the larger-scale changes in atmospheric circulation on which the reanalyses agree more closely.


We have evaluated the quality of reanalyses for use in resource assessment with particular emphasis on how well they capture observed wind variations and trends. The study has revealed pitfalls in some approaches to analysis and the difficulties in validation caused by changes in exposure and instrumentation at observing sites over long periods. We highlight where reanalyses and observations are in better agreement and also where there is a much higher level of uncertainty.

Learning objectives
Delegates will learn which reanalyses are best suited to resource assessment over the UK and Europe,and what aspects of wind trends and variation are most reliable and supported by observations.