ORIGINAL PAPER
Recovery, statistical validation and analysis of a historical
meteorological dataset collected at the Hanbury Botanical Gardens
(Liguria, northwestern Italy) from 1900 to 1940
Greta Vagge
1
&
Laura Cutroneo
1
&
Daniela Gandolfi
2
&
Gabriele Ferretti
1
&
Davide Scafidi
1
&
Marco Capello
1
Received: 28 June 2017 / Accepted: 20 May 2018
#
Springer-Verlag GmbH Austria, part of Springer Nature 2018
Abstract
A historical set of daily meteorological data collected at the Hanbury Botanical Gardens (Liguria, northwestern Italy) from 1900 to
1940 was recovered from five manually entered registers. They were digitised and statistically analysed to check their reliability
and study their trends and variations. In particular, air temperature, precipitation and their extreme values were considered,
together with wind direction, sea state, sky conditions and relative humidity. The results show a decreasing trend in mean annual
temperature of approximately 0.2 °C/decade due to a decrease in maximum air temperature. Annual cumulative precipitation
increased by 65.2 mm/decade over the study period. The data analysis showed a summer temperature decrease in 1912 and a
severe drought in 1921. Moreover, the years with most days with extreme temperatures were associated to the negative phases of
the North Atlantic oscillation (NAO). During the study period, SW winds were prevailing. Sky conditions followed seasonal
trends, while slight sea was the most frequent sea state.
1 Introduction
Climate systems are clearly affected by man (IPCC 2013). The
effects of natural climate variability combine with the effects
of human actions. Therefore, in order to really understand how
climate evolves in an environment, as much information as
possible must be gathered about past climate changes. For this
reason, historical meteorological data are key to get to know
the exact extent of observed changes (Franke et al. 2017).
They are also crucial in climatological forecasts, because they
allow researchers to make forecasts on future climate changes
with ever greater accuracy (Alexander and Arblaster 2017).
For these reasons, great efforts are being made to retrieve
historical data. Historical series are increasingly being re-
trieved and analysed (Brunetti et al. 2000;Ntegekaand
Willems 2008; Piccarreta et al. 2013; Slonosky 2002a)despite
the difficulties due to historical (e.g. the World Wars I and II)
and technical reasons: namely, notes written on paper, manu-
ally copied data on paper or recorded by instruments on paper,
difficulty to retrieve archives, changes and transfers of stations
or instruments (Brönnimann et al. 2006;Comani1987;
Maugeri et al. 2002). Collecting information on the years be-
fore 1950 is particularly important, for example by publishing
historical series through the digitisation of paper registers. As
a matter of fact, many historical data can be found on paper
registers generally entered by amateur meteorologists and
stored in archives in libraries or museums (e.g. Añel et al.
2012; Bergström and Moberg 2002; Manley 1974;Murphy
et al. 2018; Slonosky 2002b). In the majority of cases, these
Highlights
• A1900–1940 meteorological dataset was recovered from hand-filled
registers.
• Decreasing trend of 0.2 °C/decade in mean annual air temperature was
found.
• Increasing trend of 65.2 mm/decade in annual cumulative precipitation
was found.
• Evidence of the strong world drought of 1921 was found also in the
study dataset.
• Extreme high temperatures were associated to the negative phases of the
NAO.
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s00704-018-2524-3) contains supplementary
material, which is available to authorized users.
* Laura Cutroneo
laura.cutroneo@dipteris.unige.it
1
DISTAV, University of Genoa, 26 Corso Europa,
I-16132 Genoa, Italy
2
International Institute of Ligurian Studies, 39 Via Romana,
I-18012 Bordighera, IM, Italy
Theoretical and Applied Climatology
https://doi.org/10.1007/s00704-018-2524-3
@Phil_Robichaud
@deepthiw
@JoseServera