Premium
Homogenization and trend analysis of monthly mean and maximum wind speed time series in Finland, 1959–2015
Author(s) -
Laapas Mikko,
Venäläinen Ari
Publication year - 2017
Publication title -
international journal of climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.58
H-Index - 166
eISSN - 1097-0088
pISSN - 0899-8418
DOI - 10.1002/joc.5124
Subject(s) - wind speed , homogenization (climate) , environmental science , climatology , anemometer , time series , meteorology , mathematics , atmospheric sciences , statistics , geography , geology , biodiversity , ecology , biology
Observed near‐surface monthly mean and maximum wind speed time series were homogenized in this study. Observations from 144 weather stations were used in the homogenization process, of which 33 covered the entire study period of 1959–2015. Homogenized time series were used to assess the linear trends in wind speed time series for various periods of interest, including the period of 1979–2008 for comparison with other studies. Trends were analysed on annual and seasonal levels. Based on statistical homogenization, by applying the HOMER homogenization tool and use of station history metadata, 95% of the analysed time series were found to be at least to some degree inhomogeneous. Almost half of the detected inhomogeneities were verified by metadata, most of which were due to changes in station location and height of the anemometer. Adjustment of detected inhomogeneities resulted in both temporally and spatially more consistent time series, therefore improving the quality of observational wind speed time series of Finland. Trends in homogenized wind speed time series, in both mean and maximum, were found to be generally negative around the whole study area. The mean linear trend of the annual mean and maximum wind speed of 33 weather stations for the period of 1959–2015 were estimated to be −0.09 and −0.32 m s −1 decade −1 , respectively. Negative trends were consistently found, regardless of data period and season. A majority of these trends were statistically significant ( p < 0.05).