Homogenization of instrumental time series of air temperature in Central Italy (1930-2015)

Abstract

Long-term and high quality instrumental air temperature data are important for reducing the uncertainty of past temperature trends at local and global scales, and for the projection of future expected changes. Currently this type of data is limited in the Mediterranean, which is particularly important since this region is considered a hot spot for climate change. To cover the data gap for Central Italy, a set of territorially dense long-term time series of temperature data covering different climate areas in the Abruzzo region is presented in this work. Due to the possible presence of inhomogeneities (non-climatology irregularities in the data set), a homogenization process was applied to the data. Monthly maximum and minimum temperatures measured at 22 stations were homogenized for the period 1930−2015 using the software HOMER v.2.6. All stations had at least 1 break in the time series, for a total of 89 and 80 inhomogeneities identified in the maximum and minimum temperature series, respectively. The annual amplitude of breaks in the annual series ranged between 0−4.44°C for the maximum temperatures and 0.01−3.75°C for the minimum temperatures. The trend of annual mean temperatures showed increasing temperatures at a regional level starting in the early 20 th century, with a greater rate especially after 1980 (0.060°C yr −1).

Building a local climate reference dataset: Application to the Abruzzo region (Central Italy), 1930–2019

“…In the last four decades, the warming rate has been around 5–6°C/century for TX and 3–5°C/century for TN. These magnitudes are consistent with trends previously reported for Italy and the Abruzzo region (Aruffo and Di Carlo, 2019; Brunetti et al ., 2006; Fioravanti et al ., 2019; Scorzini and Leopardi, 2019).…”

“…Similarly to most places on the globe, temperature in Italy has been found to have increased in the last century, with a trend of the order of 1°C/century: in Central and Southern Italy, minimum temperature increased at a rate of about 1.3°C/century, about twice as much as the 0.7°C/century of maximum temperature (Brunetti et al ., 2006). A similar increasing trend was reported also for the Abruzzo region, where Aruffo and Di Carlo (2019) estimated a 20‐fold mean temperature increasing rate from ~0.3°C/century in the period 1924–1979 to ~6.0°C/century in 1980–2015, in line with the nation‐wide study by Fioravanti et al . (2019).…”

“…Similarly to most places on the globe, temperature in Italy has been found to have increased in the last century, with a trend of the order of 1 C/century: in Central and Southern Italy, minimum temperature increased at a rate of about 1.3 C/century, about twice as much as the 0.7 C/century of maximum temperature (Brunetti et al, 2006). A similar increasing trend was reported also for the Abruzzo region, where Aruffo and Di Carlo (2019) estimated a 20-fold mean temperature increasing rate from 0.3 C/century in the period 1924-1979 to~6.0 C/century in 1980-2015, in line with the nation-wide study by Fioravanti et al (2019). The extreme temperature indices are also displaying significant trends in the recent decades: for example, on average, in the Abruzzo region the number of 'summer days' (maximum temperature TX higher than 25 C) is increasing by one every year, and the number of 'frost days' (minimum temperature TN less than 0 C) is decreasing by one every 2 years (Scorzini et al, 2018).…”

“…As a check on the automatic procedure, we first visually inspected the results of Climatol break detection and fine‐tuned the statistical thresholds, then we compared them with the information available on station maintenance documented in UIM yearbooks, as well as with the breaks documented in the two previous studies from independent research groups that homogenized the same dataset (Aruffo and Di Carlo, 2019; Scorzini et al ., 2018), both using the homogenization software HOMER (Mestre et al ., 2013). The HOMER package was developed during the European COST Action ES0601 «HOME», Advances in Homogenization Methods of Climate Series: An Integrated Approach (Venema et al ., 2012): we verified that the code is still currently available, but not updated or integrated in the CRAN (Comprehensive R Archive Network, https://cran.r-project.org/mirrors.html) since 2013.…”

“…The selection of stations for the analysis appears to be always affected by some degree of subjectivity and to be dependent on the time period selected for the analysis: for example, regarding the analysis of the temperature dataset, the previous studies focused on different periods and used slightly different selection procedures, which resulted in a different mix of selected stations for the final analysis (Aruffo and Di Carlo, 2019; Scorzini et al ., 2018; Scorzini and Leopardi, 2019). Here we build on the same criteria used in the construction and continuous update of the European ECA&D database (ECA&D, 2013), now called E‐OBS and integrated in the European Copernicus service (Copernicus Climate Change Service, 2020), which requires a minimum length of 20 years of the time series and 70% of valid data over the period of interest.…”