A23C-3246:
Assessing the Impact of Different Measurement Time Intervals on Observed Long-Term Wind Speed Trends

Tuesday, 16 December 2014
Cesar Azorin-Molina1, Sergio M. Vicente-Serrano1, Tim McVicar2, Sonia Jerez3, Jesús Revuelto1 and J. Ignacio López Moreno1, (1)Instituto Pirenaico de Ecología, Zaragoza, Spain, (2)CSIRO, Black Mountain, Australia, (3)University of Murcia, Physics of the Earth, Murcia, Spain
Abstract:
During the last two decades climate studies have reported a tendency toward a decline in measured near-surface wind speed in some regions of Europe, North America, Asia and Australia. This weakening in observed wind speed has been recently termed “global stilling”, showing a worldwide average trend of -0.140 m s-1 dec-1 during last 50-years. The precise cause of the “global stilling” remains largely uncertain and has been hypothetically attributed to several factors, mainly related to: (i) an increasing surface roughness (i.e. forest growth, land use changes, and urbanization); (ii) a slowdown in large-scale atmospheric circulation; (iii) instrumental drifts and technological improvements, maintenance, and shifts in measurements sites and calibration issues; (iv) sunlight dimming due to air pollution; and (v) astronomical changes. This study proposed a novel investigation aimed at analyzing how different measurement time intervals used to calculate a wind speed series can affect the sign and magnitude of long-term wind speed trends. For instance, National Weather Services across the globe estimate daily average wind speed using different time intervals and formulae that may affect the trend results. Firstly, we carried out a comprehensive review of wind studies reporting the sign and magnitude of wind speed trend and the sampling intervals used. Secondly, we analyzed near-surface wind speed trends recorded at 59 land-based stations across Spain comparing monthly mean wind speed series obtained from: (a) daily mean wind speed data averaged from standard 10-min mean observations at 0000, 0700, 1300 and 1800 UTC; and (b) average wind speed of 24 hourly measurements (i.e., wind run measurements) from 0000 to 2400 UTC. Thirdly and finally, we quantified the impact of anemometer drift (i.e. bearing malfunction) by presenting preliminary results (1-year of paired measurements) from a comparison of one new anemometer sensor against one malfunctioned anenometer sensor due to old bearings.