Backtracking the Holuhraun exceptional SO2 event in September 2014 C. Maurer, G. Wotawa, and D. Arnold Central Institute for Meteorology and Geodynamics (ZAMG) Corresponding author: Christian Maurer, [email protected] Background On August, 28th, 2014, the effusive „Holuhraun“ fissure eruption started near the vent of the Bardarbunga volcano on Iceland leading to SO2 emissions of between 35000 and 90000 tons per day for several weeks. This amount of daily SO2 exceeded the daily emissions of all European smokestacks together by more than a factor of two. Whereas concentrations of up to 21000 micrograms per cubic meter [μg/m3] measured in Icelandic towns did not come as a surprise, remarkable concentrations could be found in other parts of Europe. This was especially true for the Alpine area, where health-relevant concentrations well above 200 μg/m3 could be found in south-eastern Austria. In the present work measurements of mountainous low-background stations (e.g. Sonnblick, 3105 m a.s.l.) were used together with backward (srs-)fields from the Lagrangian dispersion model FLEXPART using ECMWF meteorological input data to apply a back- trajectory statistic and to calculate the PSR (Possible Source Region) fields in accordance with the standard method applied for CTBT verification. For this so-called correlation method, also a new approach was tested, namely to correlate logarithmic measurement and model values. This approach considers also smaller measurement/model values and is not dominated by major concentrations. It will be demonstrated whether the Holuhraun eruption can be properly identified as source location for the SO2 measured, and whether realistic estimates of the source strength can be provided. Synoptic situation At the turn of Sep. 20th/21st cyclogenesis over Scandinavia intensified advecting air masses from Iceland rapidly and a massive cold front spanning from northern Scandinavia to the British Isles started heading southwards. In the evening of 21st the cold front finally reached Austria. Descending air was forced down on the Lee side due to the pressure perturbation of the wave regime induced by the Alps. Thus, a North Foehn situation prevailed the whole 22nd leading to the observed maxima in SO2. Surface pressure and front analysis from 20140922, 00 UTC. Data & Forward Runs Hourly averaged SO2-measurements [μg/m3] (solid lines) from the four mountain stations and model values (solid lines with dots) from a FLEXPART simulation with hourly 0.125° ECMWF input data (40° W to 40° E / 40° to 85° N) , 0.1° outgrid resolution and assuming a continuous 50kt/d emission up to 3km above surface from Sep. 16th onward. SO2-measurement data from four Alpine mountain observatories were used to compare with a FLEXPART V- 8.23 model run: • Hoher Sonnblick (1010—SON1; 12.96°E, 47.05°N, 3105m): half-hourly data • Jungfraujoch (JFJ01; 7.98°E, 46.55°N, 3471m): hourly data • Hoher Peissenberg (HPB01; 11.01°E, 47.80°N, 977m): minute-by- minute data • Zugspitze/Schneefernerhaus (ZUG01; 10.98° E, 47.42° N, 2650m): minute-by-minute data Total column [μg/m3] (up to 22km) SO2 concentration as modeled by FLEXPART. Left panel shows the combination of two parts of the plume forming the main plume arriving in Austria on the 22nd. Backward Runs & PSR-Fields: Classic CTBT & logarithmic Integrated Maximum PSR-field back to Sep. 15th, 6UTC, from Jungfraujoch and Sonnblick backward fields, assuming a limit for background concentrations and one for the emission scenario and using the standard CTBT-verification-approach (left panel) as well as calculating correlations after applying a logarithmic transformation (right panel). Conclusions • The Holuhraun fissure eruption next to the vent of Badarbunga caused a distinct SO2 signal at all four considered stations. Since the location of the emission is perfectly known it was a very good opportunity to test the standard CTBT- approaches for backtracking. • Due to the complex meteorological situation (no straight transport of the emitted plume towards Europe for several days, different unifying plume branches) the forward simulation proves to be successful only for the westernmost and highest station Junfraujoch despite using the best currently available ECMWF meteorological input data. For Sonnblick and Zugspitze the simulated plume arrives too early, for Hoher Peissenberg it is correct in phase, but not in magnitude. • The standard CTBT-PSR-field calculated back to Sep. 15th gives a broad region of influence. Applying a logarithmic transformation to both measurements and model values yields better results for this test case. Due to the highly skewed distributions of both measurements and model values the transformation seems to be adequate in order to avoid artificial correlation values. • A good source estimation is hampered by the meteorological conditions. Data from and model fields for Jungfraujoch yield an estimate more than 3 times too high compared with already published data (e.g. Gouhier, 2015, EGU- abstract: tropospheric emissions ~86kt/d over three months). A source term estimation was tried using the measurements and backward fields for station Jungfraujoch, since this station performs best regarding a measurement – simulation intercomparison (see time series above). First the hourly mass values of the backward field for a given 3-hourly release were divided by the release in order to yield hourly dilution (or sensitivity) values of each simulated value with regard to the source. Further, the following relationship was applied: ∑ mi *si = c ; mi : mass [kg], si : dilution [1/m3], c: concentration [kg/m3] Assuming a constant eruption rate the sum is formed over all hourly values from Sep. 16th to 20th. This period is probably the most relevant for the measured values on Sep. 22nd. As a result the average daily mass eruption rate is obtained yielding a value of 285kt/d. Back-Trajectory-Statistics Source term estimate Statistics of measurements weighted by the grid cell values of the backward fields for station Jungfraujoch (left panel) and Sonnblick (right panel). It demonstrates the dominance of north-westerly flow. Since the maximum life time of SO2 adds up to around 4 to 6 days, the statistics was confined to 6 days back from the time of the measurement. Conclusions Backward fields (valid for 0-5km above surface) for station Junfraujoch, for a release from 15 to 18 UTC on Sep. 22nd, the time of maximum measured concentrations. The plume hits Iceland first on morning of Sep. 20th (left panel) and again on Sep. 19th (right panel) staying around Iceland until early 17th. Unification of two parts of the plume T1.5-P6
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Backtracking the Holuhraun exceptional SO2 event in September 2014 C. Maurer, G. Wotawa, and D. Arnold
Central Institute for Meteorology and Geodynamics (ZAMG) Corresponding author: Christian Maurer, [email protected]
Background
On August, 28th, 2014, the effusive „Holuhraun“ fissure eruption started near the vent of the Bardarbunga volcano on Iceland leading to SO2 emissions of between 35000 and 90000 tons per day for several weeks. This amount of daily SO2 exceeded the daily emissions of all European smokestacks together by more than a factor of two. Whereas concentrations of up to 21000 micrograms per cubic meter [μg/m3] measured in Icelandic towns did not come as a surprise, remarkable concentrations could be found in other parts of Europe. This was especially true for the Alpine area, where health-relevant concentrations well above 200 μg/m3 could be found in south-eastern Austria. In the present work measurements of mountainous low-background stations (e.g. Sonnblick, 3105 m a.s.l.) were used together with backward (srs-)fields from the Lagrangian dispersion model FLEXPART using ECMWF meteorological input data to apply a back-trajectory statistic and to calculate the PSR (Possible Source Region) fields in accordance with the standard method applied for CTBT verification. For this so-called correlation method, also a new approach was tested, namely to correlate logarithmic measurement and model values. This approach considers also smaller measurement/model values and is not dominated by major concentrations. It will be demonstrated whether the Holuhraun eruption can be properly identified as source location for the SO2 measured, and whether realistic estimates of the source strength can be provided.
Synoptic situation
At the turn of Sep. 20th/21st cyclogenesis over Scandinavia intensified advecting air masses from Iceland rapidly and a massive cold front spanning from northern Scandinavia to the British Isles started heading southwards. In the evening of 21st the cold front finally reached Austria. Descending air was forced down on the Lee side due to the pressure perturbation of the wave regime induced by the Alps. Thus, a North Foehn situation prevailed the whole 22nd leading to the observed maxima in SO2.
Surface pressure and front analysis from 20140922, 00 UTC.
Data & Forward Runs
Hourly averaged SO2-measurements [μg/m3] (solid lines) from the four mountain stations and model values (solid lines with dots) from a FLEXPART simulation with hourly 0.125° ECMWF input data (40° W to 40° E / 40° to 85° N) , 0.1° outgrid resolution and assuming a continuous 50kt/d emission up to 3km above surface from Sep. 16th onward.
SO2-measurement data from four Alpine mountain observatories were used to compare with a FLEXPART V- 8.23 model run: • Hoher Sonnblick (1010—SON1; 12.96°E, 47.05°N, 3105m): half-hourly
data • Jungfraujoch (JFJ01; 7.98°E, 46.55°N, 3471m): hourly data • Hoher Peissenberg (HPB01; 11.01°E, 47.80°N, 977m): minute-by-
minute data • Zugspitze/Schneefernerhaus (ZUG01; 10.98° E, 47.42° N, 2650m):
minute-by-minute data
Total column [μg/m3] (up to 22km) SO2 concentration as modeled by FLEXPART. Left panel shows the combination of two parts of the plume forming the main plume arriving in Austria on the 22nd.
Backward Runs & PSR-Fields: Classic CTBT & logarithmic
Integrated Maximum PSR-field back to Sep. 15th, 6UTC, from Jungfraujoch and Sonnblick backward fields, assuming a limit for background concentrations and one for the emission scenario and using the standard CTBT-verification-approach (left panel) as well as calculating correlations after applying a logarithmic transformation (right panel).
Conclusions
• The Holuhraun fissure eruption next to the vent of Badarbunga caused a distinct SO2 signal at all four considered stations. Since the location of the emission is perfectly known it was a very good opportunity to test the standard CTBT-approaches for backtracking.
• Due to the complex meteorological situation (no straight transport of the emitted plume towards Europe for several days, different unifying plume branches) the forward simulation proves to be successful only for the westernmost and highest station Junfraujoch despite using the best currently available ECMWF meteorological input data. For Sonnblick and Zugspitze the simulated plume arrives too early, for Hoher Peissenberg it is correct in phase, but not in magnitude.
• The standard CTBT-PSR-field calculated back to Sep. 15th gives a broad region of influence. Applying a logarithmic transformation to both measurements and model values yields better results for this test case. Due to the highly skewed distributions of both measurements and model values the transformation seems to be adequate in order to avoid artificial correlation values.
• A good source estimation is hampered by the meteorological conditions. Data from and model fields for Jungfraujoch yield an estimate more than 3 times too high compared with already published data (e.g. Gouhier, 2015, EGU-abstract: tropospheric emissions ~86kt/d over three months).
A source term estimation was tried using the measurements and backward fields for station Jungfraujoch, since this station performs best regarding a measurement – simulation intercomparison (see time series above). First the hourly mass values of the backward field for a given 3-hourly release were divided by the release in order to yield hourly dilution (or sensitivity) values of each simulated value with regard to the source. Further, the following relationship was applied:
∑ mi *si = c ; mi : mass [kg], si : dilution [1/m3], c: concentration [kg/m3] Assuming a constant eruption rate the sum is formed over all hourly values from Sep. 16th to 20th. This period is probably the most relevant for the measured values on Sep. 22nd. As a result the average daily mass eruption rate is obtained yielding a value of 285kt/d.
Back-Trajectory-Statistics Source term estimate
Statistics of measurements weighted by the grid cell values of the backward fields for station Jungfraujoch (left panel) and Sonnblick (right panel). It demonstrates the dominance of north-westerly flow. Since the maximum life time of SO2 adds up to around 4 to 6 days, the statistics was confined to 6 days back from the time of the measurement.
Conclusions
Backward fields (valid for 0-5km above surface) for station Junfraujoch, for a release from 15 to 18 UTC on Sep. 22nd, the time of maximum measured concentrations. The plume hits Iceland first on morning of Sep. 20th (left panel) and again on Sep. 19th (right panel) staying around Iceland until early 17th.
Unification of two parts of the plume
T1.5-P6
Related Documents
