Historical reanalyses that span more than a century are needed for a wide range of studies, from understanding large‐scale climate trends to diagnosing the impacts of individual historical extreme weather events. The Twentieth Century Reanalysis (20CR) Project is an effort to fill this need. It is supported by the National Oceanic and Atmospheric Administration (NOAA), the Cooperative Institute for Research in Environmental Sciences (CIRES), and the U.S. Department of Energy (DOE), and is facilitated by collaboration with the international Atmospheric Circulation Reconstructions over the Earth initiative. 20CR is the first ensemble of sub‐daily global atmospheric conditions spanning over 100 years. This provides a best estimate of the weather at any given place and time as well as an estimate of its confidence and uncertainty. While extremely useful, version 2c of this dataset (20CRv2c) has several significant issues, including inaccurate estimates of confidence and a global sea level pressure bias in the mid‐19th century. These and other issues can reduce its effectiveness for studies at many spatial and temporal scales. Therefore, the 20CR system underwent a series of developments to generate a significant new version of the reanalysis. The version 3 system (NOAA‐CIRES‐DOE 20CRv3) uses upgraded data assimilation methods including an adaptive inflation algorithm; has a newer, higher‐resolution forecast model that specifies dry air mass; and assimilates a larger set of pressure observations. These changes have improved the ensemble‐based estimates of confidence, removed spin‐up effects in the precipitation fields, and diminished the sea‐level pressure bias. Other improvements include more accurate representations of storm intensity, smaller errors, and large‐scale reductions in model bias. The 20CRv3 system is comprehensively reviewed, focusing on the aspects that have ameliorated issues in 20CRv2c. Despite the many improvements, some challenges remain, including a systematic bias in tropical precipitation and time‐varying biases in southern high‐latitude pressure fields.

Instrumental meteorological measurements from periods prior to the start of national weather services are designated “early instrumental data.” They have played an important role in climate research as they allow daily to decadal variability and changes of temperature, pressure, and precipitation, including extremes, to be addressed. Early instrumental data can also help place twenty-first century climatic changes into a historical context such as defining preindustrial climate and its variability. Until recently, the focus was on long, high-quality series, while the large number of shorter series (which together also cover long periods) received little to no attention. The shift in climate and climate impact research from mean climate characteristics toward weather variability and extremes, as well as the success of historical reanalyses that make use of short series, generates a need for locating and exploring further early instrumental measurements. However, information on early instrumental series has never been electronically compiled on a global scale. Here we attempt a worldwide compilation of metadata on early instrumental meteorological records prior to 1850 (1890 for Africa and the Arctic). Our global inventory comprises information on several thousand records, about half of which have not yet been digitized (not even as monthly means), and only approximately 20% of which have made it to global repositories. The inventory will help to prioritize data rescue efforts and can be used to analyze the potential feasibility of historical weather data products. The inventory will be maintained as a living document and is a first, critical, step toward the systematic rescue and reevaluation of these highly valuable early records. Additions to the inventory are welcome.

The Svalbard Airport composite series spanning the period from 1898 to the present represents one of very few long-term instrumental temperature series from the High Arctic. A homogenized monthly temperature series is available since 2014. Here we increase the resolution from a monthly to daily basis, and further digitization of historical data has reduced the uncertainty of the series. The most pronounced changes in the 120-year record occur during the last three decades. For the 1991–2018 period the number of days warmer than 0 and 5 °C has increased by 25 (21%) and 22 (59%), respectively, per year compared to the 1961–1990 standard normal. Likewise, comparing the same periods, the number of days colder than −10 and −20 °C has decreased by 42 (32%) and 27 (62%), respectively. During the entire time span of the series, the western Spitsbergen climate has gone through stepwise changes, alternating between cold and warm regimes: 1899–1929 was cold, 1930–1961 warm, 1962–1998 cold and 1999–2018 warm. The latest cold regime was 1.0 °C warmer than the first cold one, and the latest warm regime was 1.7 °C warmer than the previous warm one. For the whole series the linear trend for annual means amounts to 0.32°C/decade, which is about 3.5 times the increase of the global mean temperature for the same period. Since 1991, the rate of warming at Svalbard Airport is 1.7 °C/decade, which is more than twice the Arctic average (0.8 °C/decade, north of 66 °N) and about seven times the global average for the same period.

Recent air temperature changes in the high Arctic (HA) have been investigated based on mean seasonal and annual data calculated for the period 1951–2015 and for two sub-periods 1976–2015 and 1996–2015. Two kinds of air temperature data (observational and reanalysis) have been used in the research. The observational data were compared with data taken from six reanalysis products (20CRv2c, CERA-20C, ERA-Int, MERRA-2, NCEP-CFSRR, JRA-55). The scale of the HA warming for the period 1996–2015 relative to the reference period 1951–1990 reached 1.6 °C for annual mean and was greatest in autumn (1.9 °C) and in winter (1.7 °C), while it was smallest in summer (0.9 °C). Evidently, the greatest warming was observed in the Atlantic and Siberian climatic regions, while in the rest of the HA, the rate of warming was usually weaker than trends calculated for the period 1976–2015. Air temperature tendencies in all study periods 1951–2015, 1976–2015 and 1996–2015 showed a predominance of positive trends that were statistically significant at the level of 0.05. In the two latter periods, the rate of warming was on average 2–3 times faster than for the entire study period. In the HA, there has not been a slowdown in the rate of warming (“hiatus”) in the last two decades (in contrast to that which was noted for the Northern Hemisphere). Our results reveal that, in most cases, the closest fit to observations was obtained for two reanalysis products (the ERA-Interim and JRA-55, since 1979) and the six reanalysis average. Two new polar amplification (PA) metrics based on scaled and standardised values of surface air temperature (SAT) reveal the non-existence of this phenomenon in the period 1951–2015. One of the metrics shows very small PA in the periods 1976–2015 and 1996–2015.

Air temperature variations and gradients along the coast and fjords of western Spitsbergen

In this paper, the results of an investigation into the thermal conditions in the Arctic in the period from 1801 to 1920 are presented. For this ‘early instrumental’ period limited meteorological data exist. Generally, the first meteorological stations in the Arctic were established in the second half of the 19th century and almost all of them were located in the coastal parts of Greenland. In order to get at least a rough idea of thermal conditions in the Arctic in the study period, data from different land and marine expeditions were collected. A total of 118 temperature series of monthly means have been gathered. Although the area and time periods covered by the data are variable, it is still possible to describe the general character of the temperature conditions.The results show that the areally averaged Arctic temperature in the early instrumental period was 0.8 °C lower than the next 60‐year period (1861–1920). In comparison to present‐day conditions, winter and autumn were significantly colder (winter by 1.6 °C and autumn by 0.9 °C) than were summer (colder by 0.4 °C) and spring (colder by only 0.2 °C). The air temperature in the real Arctic during the first International Polar Year (IPY) was, on average, colder than today by 1.0–1.5 °C. Winter was exceptionally cold with the average temperature being lower by more than 3 °C in all months except February. On the other hand, spring (March–May) was slightly warmer than today, and April was exceptionally warm (1.1 °C above present norm).The temperature differences calculated between historical and modern mean monthly temperatures show that majority of them lie within one standard deviation (SD) from present long‐term mean. Thus, it means that the climate in the early instrumental period was not as cold as some proxy data suggest. Copyright © 2009 Royal Meteorological Society