Stratospheric intrusion (SI) events are a topic of ongoing research, especially because of their ability to change the oxidation capacity of the troposphere and their contribution to tropospheric ozone levels. In this work, a novel tool called STEFLUX (Stratosphere-to-Troposphere Exchange Flux) is presented, discussed, and used to provide a first long-term investigation of SI over two global hot-spot regions for climate change and air pollution: the southern Himalayas and the central Mediterranean Basin. The main purpose of STEFLUX is to obtain a fast-computing and reliable identification of the SI events occurring at a specific location and during a specified time window. It relies on a compiled stratosphere-to-troposphere exchange (STE) climatology, which makes use of the ERA-Interim reanalysis dataset from the ECMWF, as well as a refined version of a well-established Lagrangian methodology. STEFLUX results are compared to the SI observations (SIO) at two high-mountain WMO/GAW global stations in these climate hot spots, i.e., the Nepal Climate Observatory-Pyramid (NCO-P, 5079 m a.s.l.) and Mt. Cimone (2165 m a.s.l.), which are often affected by SI events. Compared to the observational datasets at the two specific measurement sites, STEFLUX is able to detect SI events on a regional scale. Furthermore, it has the advantage of retaining additional information concerning the pathway of stratospheric-affected air masses, such as the location of tropopause crossing and other meteorological parameters along the trajectories. However, STEFLUX neglects mixing and dilution that air masses undergo along their transport within the troposphere. Therefore, the regional-scale STEFLUX events cannot be expected to perfectly reproduce the point measurements at NCO-P and Mt. Cimone, which are also affected by small-scale (orographic) circulations. Still, the seasonal variability in SI events according to SIO and STEFLUX agrees fairly well. By exploiting the fact that the ERA-Interim reanalysis extends back to 1979, the long-term climatology of SI events at NCO-P and Mt. Cimone is also assessed in this work. The analysis of the 35-year record at both stations denies the existence of any significant trend in the SI frequency, except for winter seasons at NCO-P. Furthermore, for the first time, by using the STEFLUX outputs, we investigate the potential impact of specific climate factors (i.e. ENSO, QBO, and solar activity) on SI frequency variability over the Mediterranean Basin and the Himalayas.
Stratosphere-to-troposphere exchange (STE) represents one of the natural
processes that have substantial impacts on meteorology and atmospheric
chemistry, and is an important aspect of climate change
High-mountain stations are appropriate sites for investigating the transport
of stratospheric air masses into the troposphere, because stratospheric
air masses can already be identified at mid-tropospheric levels. Furthermore,
they are less influenced by polluted air masses due to local or regional
anthropogenic emissions
In this work we present a novel tool, which aims at objectively identifying
SI events reaching a “target” geographical region and during a specific time
window. The tool, called STEFLUX (Stratosphere-to-Troposphere Exchange Flux),
is a relatively fast-computing algorithm which makes use of the pre-computed
trajectories composing the STE climatology by
The paper is structured as follows. In Sect.
Datasets of daily SI occurrences are available at two high-altitude WMO/GAW
global stations, i.e., the Nepal Climate Observatory-Pyramid (NCO-P, 5079 m a.s.l., Nepal) and Mt. Cimone (2165 m a.s.l., Italy) starting from 2006 and 1998,
respectively. In this section, a brief description of the two measurement
sites is provided, together with the description of the methodology used to
detect SI events based on the analysis of in situ stratospheric tracers'
variability (coupled with additional model data). Hereinafter, these datasets
will be referred to as “stratospheric intrusion observations” (SIO).
Technical details on the different parameters considered are given in the
Supplement and in the papers by
NCO-P (27.95 significant variations in daily significant daily TCO increases and presence of back-trajectories with values of PV significant variations in daily the presence of RH values lower than 60 % and significant negative correlation of O
The significant variations are obtained as follows: first, three iterations of a 21-day running mean
Mt. Cimone (44.19 significant daily TCO increases and presence of back-trajectories with values of PV significant daily presence of RH values lower than 40 % and presence of back-trajectories with values of PV presence of RH values lower than 40 % and significant TCO daily value increases.
Again, the significant variations are defined in the same way as done for
NCO-P, and at least one criterion must be valid for tagging the selected day
as influenced by SI. The period of study for Mt. Cimone spans from January
1998 to December 2010.
SIO at these two measurement sites provide a unique opportunity to test the capacity of STEFLUX in reproducing the main features of SI occurrences (frequency, seasonality, long-term variations) at two locations representative of the Northern Hemisphere midlatitudes and subtropics.
The main purpose of STEFLUX is to obtain a fast-computing and reliable
estimation of SI occurring at a specific location. The database used as input
relies on the trajectories from the STE climatology presented in
STEFLUX detects the air parcels originating in the stratosphere and entering a tropospheric 3-D target box during a specific time window. For this reason, several parameters need to be defined for the STEFLUX tool to work: (i) the time period for which the analysis should be carried out, and (ii) the target box by means of its longitude and latitude boundaries and by its vertical extension from the surface up to the top boundary (defined as a pressure level, in hPa). On request, the PBL height can be used as the top boundary of the target box: this option takes into account the ERA-Interim PBL height, which is a parameter also available along each STE trajectory. In addition to this, another optional parameter allows the temporal resolution for the STE trajectories to be increased from its default value (6 h) up to 1 h. STEFLUX produces several output files; these enclose (i) the trajectory positions and timing found within the box, (ii) the first box crossing positions and timing for each trajectory, (iii) the tropopause crossing position and timing for each trajectory, and (iv) the complete list of the trajectories that have crossed the box.
Example of application of STEFLUX, in a target box around NCO-P position over the period 9–25 January 2007.
Panel
To present an application of STEFLUX and its output files, a SI case in
January 2007 is discussed. The period of study coincides with a case study
discussed in
Figure
Seasonal graph of SI frequency at NCO-P
The first positions of the air parcels after entering the target box are
marked in Fig.
Input parameters for STEFLUX for the comparison with in situ measurements (SIO).
In this section, the SI occurrences from STEFLUX are compared to the ones
from SIO at the two WMO/GAW global stations (see Sect.
The aim of this part is twofold: first, we would like to see how STEFLUX
compares to SIO (in Sects.
The seasonal frequency (in %) of SI days, derived from measurements (SIO),
is presented in Fig.
Box-and-whisker plot of the annual variation in SI frequency at NCO-P
Although the seasonality was a feature well captured by STEFLUX, the
interannual variability was less clearly identifiable. Concerning NCO-P
(Fig.
In this section, we extend the comparison to a higher temporal resolution,
i.e., by considering single SI events. More specifically, in this study, a SI
event was defined (for both STEFLUX and SIO) as the aggregation of contiguous
SI days. Furthermore, cases in which two distinct SI events were separated by
a single no-SI day were treated like a single event covering the entire
period. Generally, an event-based comparison between modeled and observed SI
events and experimental detection is a very challenging task, as pointed out
by previous investigations, concerning the transport and mixing of
stratospheric air deep into the troposphere
The 2
For NCO-P, based upon the SIO criteria, a total of 203 SI events (361 days influenced by SI, representing 13 % of the period) were identified, with duration ranging from 1 to 14 days and an average length of 1.9 days. On the other hand, STEFLUX identified 155 SI events (376 days, 13 %), with duration ranging from 1 to 10 days (average length: 2.6 days). At Mt. Cimone, 299 SI events (433 days, 9 %) were identified by the SIO methodology (with duration ranging from 1 to 8 days and an average length of 1.6 days), while STEFLUX yielded 237 SI events (491 days, 10 %) that lasted from 1 to 10 days (with an average length of 2.2 days).
To assess the STEFLUX performance, the approach presented in
For an overview of the results of this comparison, we computed contingency
tables (Table
Table
Several possible reasons can explain the mismatch between the SIO and STEFLUX
time series. For instance, STEFLUX is not fully able to capture subgrid-scale
processes (like convection, turbulent diffusion, and mixing) along the path
from the stratosphere to the target region. This deficiency becomes
particularly pronounced over mountainous measurement sites, mostly because of
the complex topography and the associated small-scale thermally and
dynamically driven circulations that characterize the area. As shown in
In the case of long travel times from the tropopause to the target region, we
expect a stronger impact of mixing and dilution processes on the air-mass
properties. Hence, when a SI actually affects a specific region, the SIO
criteria might not be able to detect it, because mixing and dilution with
tropospheric air masses could lower stratospheric tracers concentrations
below the thresholds used for detection. Then, we computed the travel times
(hereinafter called
A further point of discrepancy between STEFLUX and SIO results is related to
the “overpasses” phenomenon, i.e., air masses that overpass the station at
altitudes high enough that there is no indication in the measurements record
(but might be observed by STEFLUX). Indeed, during a study conducted at the
Zugspitze (Germany, 2962 m a.s.l.),
In summary, although correctly depicting the typical seasonal variability in SI frequency at NCO-P and Mt. Cimone, the STEFLUX and SIO time series differ for several reasons. These differences point out that the complete approach to study and assess SI is to deploy together modeling tools and observations, because they are complementary and address together several scientific questions. In particular, in situ observations have the advantage of capturing short and transient SI events associated to transport processes occurring at subgrid scales, while STEFLUX has the advantage of detecting the arrival of stratospheric-affected air masses, irrespective of the degree of mixing and dilution along the transport within the troposphere.
In this section we present a climatology of SI events, as defined in Sect.
Density of tropopause (TP) crossings for the period 1979–2013, for NCO-P
An important aspect of SI is where and when the SI trajectories actually
crossed the tropopause. First, the location of the crossing is useful to
determine the O
Time series of monthly averaged SI frequency, as retrieved by STEFLUX, at NCO-P
To detect potential trends in the SI frequencies, we adopted the same STEFLUX
climatological datasets presented in Sect.
Same as Fig.
The long-term variability in SI frequency at NCO-P and Mt. Cimone was further
analyzed with respect to potential oscillations and periodicities. To this
aim, we applied the complete ensemble empirical mode decomposition with
adaptive noise method
For Mt. Cimone the situation was different and more difficult to understand.
The seasonal cycle is still present (IMF 4, Fig.
In this work we presented a novel methodology (STEFLUX) to evaluate SI events in a
user-defined region by using as input a Lagrangian STE climatology derived
from the ERA-Interim reanalysis. Besides having shown an illustrative case
study (Sect.
Our results showed that STEFLUX correctly represented the typical seasonal
cycles of SI frequencies over these two areas, with the highest occurrence of
SI in winter for NCO-P and in winter–spring for Mt. Cimone. For both
sites, the lowest SI occurrence was recorded during summer (i.e., the monsoon
for NCO-P). STEFLUX had real skill (higher for Mt. Cimone than NCO-P) in
detecting single events at both regions, especially for robust (i.e., longer
than 1 day) events. The identification of short events was more problematic;
this is in agreement with a similar study by
The observed mismatch between STEFLUX and SIO occurs for several reasons, such as the absence of representation of subgrid-scale processes in STEFLUX (e.g., convection, turbulent diffusion and mixing), along the path from the stratosphere to the target region. Furthermore, one should consider that these subgrid-scale processes can also lead to “local” or transient SI events captured by a single measurement point, which cannot be considered representative/significant for a whole region. In addition to this, another reason for the mismatch might be mixing and dilution processes occurring within air masses from the tropopause crossing to the target region, expected to be maximized for greater travel times. Lastly, as also demonstrated in previous studies, the “overpasses” phenomenon might have a not negligible impact. All of these discrepancies point out that a combination of modeling outputs (e.g., STEFLUX) and in situ observations is still needed to completely study, characterize, and evaluate the occurrence of SI.
Another important feature of STEFLUX is its capability of climatologically
assessing the SI occurrence at the chosen site, since the ERA-Interim
reanalysis extends back to 1979. In this study, it allowed us to obtain a
35-year time series of SI events at NCO-P and Mt. Cimone, which affected 13
and 9 % of the period, respectively. The tropopause crossings during the
whole 35-year period, provided by STEFLUX, were further analyzed. NCO-P showed
two main cluster regions, i.e., central Asia (maximized for events with short
travel times between the tropopause and the target box,
The STEFLUX outputs are available on request by writing an e-mail to the authors, specifying the box characteristics and the period of study chosen.
This work was partially supported by the Project of National Interest NextData. The authors thank the ECMWF and MeteoSwiss for providing access to the meteorological data, the staff at NCO-P for technical and logistic help, F. Calzolari and F. Roccato (CNR–ISAC) for the technical support at Mt. Cimone, and the “Magera team” (C. Magera, P. Giambi, and N. Gherardini) and the Italian Air Force (CAMM Monte Cimone) for the valuable cooperation. Furthermore, the authors thank the NOAA–CPC for the access to the MEI, the Free University of Berlin for the QBO index, the Royal Observatory of Belgium for providing the sunspot numbers, and S. Pfahl for the useful discussions. Edited by: J. Kuttippurath Reviewed by: two anonymous referees