A century-long homogenized dataset of near-surface wind speed observations since 1925 rescued in Sweden, HomogWS-se

The main reasons for the lack of data rescue and homogenization of early near-surface wind speed (WS) observations before the 1960s are insufficient manpower and lack of funding. Funding from the Swedish Research Council for Sustainable Development (FORMAS) for a joint project (ref. 2019-00509) ‘Assessing centennial wind speed variability from a historical weather data rescue project in Sweden (WINDGUST)’ among the Swedish Meteorological and Hydrological Institute, the University of Gothenburg, and the Spanish National Research Council, presents a great opportunity to rescue and homogenize the early paper-based WS data in Sweden, for creating a century-long homogenized WS dataset.

Here, we rescued paper-based WS records dating back to the 1920s at 13 stations in Sweden and established a four-step homogenization procedure to generate the first 10-member centennial homogenized WS dataset (HomogWS-se) for community use. First, background climate variation in the rescued WS series was removed, using a verified reanalysis series as a reference series to construct a difference series. A penalized maximal F test at a significance level of 0.05 was then applied to detect artificial change-points. About 38% of the detected change-points were confirmed by the known events recorded in metadata, and the average segment length split by the change-points is ~11.3 years. A mean-matching method using up to five years of data from two adjacent segments was used to adjust the earlier segments relative to the latest segment. The homogenized WS series was then obtained by adding the homogenized difference series back onto the subtracted reference series. Compared with the raw WS data, the homogenized WS data is more continuous and lacks significant non-climatic jumps. The homogenized WS series presents an initial WS stilling and subsequent recovery until the 1990s, whereas the raw WS fluctuates with no clear trend before the 1970s. The homogenized WS shows a 25% reduction in the WS stilling during 1990-2005 than the raw WS, and this reduction is significant when considering the homogenization uncertainty from reference series. The homogenized WS exhibits a significantly stronger correlation with the North Atlantic Oscillation (NAO) than that of the raw WS (0.54 vs 0.29). These results highlight the importance of the century-long homogenized WS series in increasing our ability to detect and attribute multidecadal variability and changes in WS. HomogWS-se will be released on an open-access data repository for community uses, including studying WS changes, assessing model simulations, and constraining future projections of WS and wind energy potential. The proposed homogenization procedure enables other countries or regions to rescue their early climate data and jointly build global long-term high-quality datasets.