This work sits at the fringe of atmospheric variability research: its nearest neighbors focus on pulsars, stratospheric weather, and seismic noise, not temperature profiles over a specific region. Individual global oscillations (MJO, AO, QBO, ENSO, NAO) have been studied elsewhere, but their combined influence on the vertical temperature structure in mid-latitudes, especially over the Po Valley, remained poorly described. This paper fills that gap by connecting local radiosonde data to well-known climate modes.
Radiosonde data from twice-daily launches at San Pietro Capofiume station (44°39'N, 11°36'E, 11 m above sea level) from August 1987 to March 2010 were used. Temperature profiles were interpolated onto a vertical grid from 100 m to 20 km with 100 m spacing, and gaps were filled by linear interpolation with 12-hour steps. To suppress mesoscale processes (periods shorter than 10 days), 10-day running means were applied. Then the data were separated into intraseasonal (periods shorter than a year) and interannual (longer than a year) components using a 30-day moving average and removal of annual and semiannual cycles. Singular value decomposition (SVD), equivalent to principal component analysis, was applied to each anomaly matrix, with normalization that returns empirical orthogonal functions (EOFs) in kelvins. Spectra of the principal components (PCs) were computed using the Lomb–Scargle periodogram.
For intraseasonal variability, the first four EOFs explain 93–95% of the variance. The leading mode EOF1 (about 70%) has a maximum amplitude of about 1 K in the troposphere at 1–8 km altitude, a sign change near the tropopause, and gradual decay in the lower stratosphere. Its PC1 contains periods of 30–120 days with peaks at 51, 54, 73, and 98 days, linked to the MJO and AO and their interaction. EOF2–EOF4 (together about 30%) are concentrated in the upper troposphere and lower stratosphere (UTLS) and are modulated by annual and semiannual cycles. For interannual variability, four EOFs also explain about 95% of the variance. EOF1 (about 68%) resembles the intraseasonal EOF1: a maximum of about 1 K in the troposphere (1–9 km), a phase change at the tropopause, and a secondary maximum at 13 km. PC1 has periods of 1–7 years with peaks at 1.3, 1.7, 2.3, 2.9, and 6.8 years; the power ratio of the 2.9 and 6.8 year peaks is close to that of the ENSO index spectrum, indicating a dominant El Niño influence. The secondary modes EOF2–EOF4 (together about 27%) are localized in the UTLS and are associated with the 21-year solar cycle, a 10.3-year cycle, and a 4–8 year band (NAO).
Unlike neighboring studies that examine individual oscillations or other parameters (seismic noise, stratospheric weather), this research is the first to systematically decompose vertical temperature profiles over a specific mid-latitude site into modes linked to several global climate patterns simultaneously. This allows a quantitative estimate of the contribution of MJO, AO, QBO, ENSO, and NAO to local temperature variability and reveals the altitude structure of their influence, including the phase change at the tropopause.
The topic could be extended by applying similar analysis to other regions and longer data records, which would test the universality of the found connections. Additional parameters (humidity, wind) could also be included, and comparisons with climate model outputs could assess their ability to reproduce the observed modes.
The results can be used to validate and improve regional climate models and to better understand the mechanisms by which global oscillation signals propagate to mid-latitudes.
The next step could be extending the time series with more recent data and performing similar analyses for other European stations to identify spatial inhomogeneities in the response.
This work addresses a key problem in atmospheric science: how global climate modes (MJO, AO, QBO, ENSO, NAO) manifest in local temperature variability at different altitudes. The physical mechanisms behind the phase change at the tropopause and the propagation into the stratosphere remain open questions.
The San Pietro Capofiume station sits just 11 meters above sea level, yet its radiosondes ascend to 20 km, which is 1800 times higher than the launch point. Over 22 years, more than 16,000 temperature profiles have been collected, which would fill a book about 3 meters thick if each profile took up a page.