Data rescue of historical wind observations in Sweden since the 1920s

The ongoing climate change raises the question if, how and why the wind climate is changing. IPCC stated in AR6WG1 that the confidence in wind changes is “low” to “medium”; there is limited knowledge of historical wind changes and multidecadal variability. A fundamental challenge for the climate research community is to rescue old climate observations from weather archives at the National Weather Services and derive homogeneous and complete data series. At SMHI most of the meteorological data is available digitally since the 1960s, but before that only a minor part of the data is previously digitized.

Historical wind speed and direction observations from 1920th to 1940th from 13 stations in Sweden have been rescued and digitized making 165 additional station years of wind data available through Swedish Meteorological and Hydrological Institute (SMHI) open data portal. Stations with instrumental measurements of wind were selected and in the early 1900-century the accordingly equipped stations were mainly found at lighthouses along the coast and at airports. The dominant type of anemometer was of cup-type and different versions are described in the article. The work followed the protocol “Guidelines on Best Practices for Climate Data Rescue” of the World Meteorological Organization. Along with digitize the wind observations meta data of the measurements done at the stations was collected and compiled as a support to the following quality control and homogenization of the wind data. The meta data showed that the most common identified possible homogeneity break was change of observer, but also change of instrument type and position were found in the records.

The presentation is a part of the first work package of the WINDGUST project, which is a collaboration between the SMHI, the University of Gothenburg - Regional Climate Group (GU-RCG) and the Spanish National Research Council (CSIC). The project aim is to fill the key gap of short availability (since 1939) and temporal inhomogeneity of wind datasets in Sweden. Especially, the results could contribute to futures studies on the causes driving the current “stilling” and “reversal” debate in a global warming climate.

Since previous presentations of the current work package, a data screening has been performed for the wind observations from the 13 digitized stations to visualize the data cover and monthly variability and wind speed range. Two distinct categories of stations with separate wind patterns can be established: coastal and inland stations where inland stations typically has a weak annual variation while coastal stationstypically experience the highest wind speed in November and December.