• Aucun résultat trouvé

Study of the hemispheric difference in observed NO2 trends

N/A
N/A
Protected

Academic year: 2021

Partager "Study of the hemispheric difference in observed NO2 trends"

Copied!
1
0
0

Texte intégral

(1)Study of the hemispheric difference in observed NO2 trends K. Kreher1), J.B. Liley1), P.V. Johnston1), A. Thomas1), B.J. Connor2), B. Hassler1,3), J. Zawodny4), M. van Roozendael5), P. Demoulin6) 1) National Institute of Water and Atmospheric Research, Lauder, New Zealand 2) BC Consulting, Alexandra, New Zealand 3) Meteorological Institute, University of Munich, Munich, Germany 4) NASA Langley Research Center, Hampton, USA 5) Belgian Institute for Space Aeronomy, Brussels, Belgium 6) Institute of Astrophysics and Geophysics, University of Liège, Liège, Belgium. Pictures of the four primary NDACC stations included in this study (from left: Jungfraujoch, Lauder, Mauna Loa (top), Arrival Heights (bottom). Background. Method. As summarized in the WMO 2006 report, Chapter 3 (Global Ozone), evidence suggested that the NO2 (nitrogen dioxide) trends in the Southern Hemisphere (SH) are significantly larger than in the Northern Hemisphere (NH). This finding was based on the trend analysis of long-term observation of NO2 made at Lauder (45.0oS), New Zealand using UV-visible spectroscopy and at the Jungfraujoch (46.5oN), Switzerland using Fourier transform spectroscopy.. To investigate this hemispheric difference further, we present a statistical analysis of longterm time series of NO2 from ground-based measurements made at several locations in the NH and SH (including Lauder and Jungfraujoch), using two independent ground-based techniques (UV-visible spectroscopy and FTIR) and one satellite data set (SAGE II). All the monthly time-series were fitted by multi-linear regression with the same model, as used in our past work (Liley et al, JGR, 2000). In addition to the linear (decadal) trend, basis functions include three sine-cosine pairs for annual, semi-annual, and 4-month periods, 10.7 cm solar flux, Quasi-Biennial Oscillation, Southern Oscillation Index, and volcanic aerosol terms if the measurements date back to the aftermath of Pinatubo or El Chichon. Monthly means are used because there is still substantial autocorrelation in the residuals at that time resolution, showing that there are other effects not accounted for by the statistical model. Higher time resolution would compound these effects even more, meaning that any apparent improvement in statistical confidence would be spurious. To allow for the non-independence of residuals, standard errors of the fitted terms are increased by the autocorrelationdependent factor adopted by Weatherhead et al. (JGR, 1998).. A linear increase of 6.2 ± 1.8% per decade (am) and 5.7 ± 1.1% per decade (pm) was inferred for Lauder, while the trend analysis using the same algorithm for the Jungfraujoch data set produced a linear trend of only 1.5 ± 1.0% per decade. While the trend at Jungfraujoch is consistent with the 2.4% per decade increase in tropospheric N2O (nitrous oxide), the WMO 2003 report concluded that both the observed N2O increase and the decrease in ozone explained a trend in NO2 of 5 ± 1% per decade.. 12. 12. 10. 10. 6.3 +/- 3.5 % per decade (pm). Trend 1984.8 – 2005.6. NO2 slant column (1016 cm-2 ). 4. 2. 1996. 1998. Arrival Heights (77.8oS, 166.7oE). Kiruna (67.8 N, 20.4 E) 2000. 2002. 2004. 2006. 2008. 0 1982. 2010. 1984. 1986. 1988. 1990. 1992. 1994. 1996. 1998. 2000. 2002. 2004. 2006. 2008. 2010. Year. Year. 12. 12. Jungfraujoch (46.5oN, 8.Eo) 10. NO2 slant column (1016 cm-2 ). NO2 slant column (1016 cm-2 ). 10. 8. 6. 4. 2. 0 1990. 1996. 1998. 4. o. 2004. 2006. 2008. 1996. 2010. 1998. 2000. 14.8 +/- 4.1 % per decade (am). 2002. 12. 10. 10. NO2 slant column (1016 cm-2 ). 12. 8. 6. 4. -7.2 +/- 1.8 % per decade (pm). 2. Moshiri (44.4oN, 142oE). 2004. 1992. 1994. 1996. 2000. 2010. 6. 4. 2. 4.6 +/- 0.8 % per decade (pm). -7.6 +/- 2.4 % per decade (am). 1998. 2008. 8. Lauder (45.0oS, 170.0oE) 0 1990. 2006. Year. Year. NO2 slant column (1016 cm-2 ). 14.0 +/- 2.4 % per decade (pm). o. Macquarie Island (54.5 S, 159 E) 2002. 2002. 2004. 2006. 2008. 0 1980. 2010. 1985. 5.4 +/- 1.1 % per decade (am). 1990. 1995. 2000. 2005. 2010. Year. Year 7. 12. Jungfraujoch (46.5oN, 8.Eo). 6. Mauna Loa (19.5oN, 156oW). 6. 4. 2. -3.7 +/- 3.9 % per decade (pm). 1996. 1998. 2000. 2002. 2004. 2006. 2008. -0.2 ± 5.0 -18.0 ± 4.3. -0.5 ± 4.0 -7.5 ± 3.9. -0.2 ± 5.8 -1.6 ± 5.9. -0.9 ± 3.9 -0.5 ± 5.2. 7.1 ± 1.3. 7.5 ± 2.7. 6.2 ± 1.8. 15.8 ± 5.3. 17.2 ± 6.2. -2.2 ± 3.2 20.4 ± 4.5 33.9 ± 10.2. 2.1 ± 2.3 19.3 ± 3.1 37.6 ±12.2. 3.1 ± 1.5. 0.9 ± 2.5. 0.1 ± 3.3. 4.4 ± 1.5 3.5 ± 1.9 3.1 ± 1.4 5.9 ± 2.2. -2.9 ± 3.8 -2.4 ± 5.8 -2.7 ± 4.7 -8.9 ± 7.7. -16.9 ± 2.7 -18.5 ± 3.5 -20.6 ± 2.6 -25.7 ± 4.8. SAGE II (0o-20oS) SAGE II (20o-40oS) SAGE II (40o-50oS) SAGE II (50o-60oS). 7.9 ± 3.0 6.3 ± 1.8 6.5 ± 2.0 5.5 ± 1.6. -1.9 ± 8.5 4.0 ± 4.6 1.4 ± 5.4 10.8 ± 3.9. -17.3 ± 5.3 -16.9 ± 1.8 -14.5 ± 1.8 -5.9 ± 6.7. Table 1 lists the fitted trend in percent (%) per decade in monthly mean NO2 relative to the intercept at year 2000.0 for all ground-based stations and instruments, and the SAGE II sunset data. End date for all calculated trends is 2005.6. Trends in yellow are assumed real with a 95% confidence level when the ratio between the trend and its standard deviation is > 2. The SAGE II sunrise data are omitted because they have been corrected for a number of instrumental effects that might introduce a trend with time with the ‘Thermal Shock’ correction being the most important one.. 4. 3. 2. 0.7 +/- 1.4 % per decade (pm). 1. 2010. sunset. -2.6 ± 3.5 -7.8 ± 2.7 -6.4 ± 2.2. 5. -4.1 +/- 4.4 % per decade (am) 0. sunrise. -3.7 ± 4.3 -9.8 ± 2.9 -7.1 ± 3.4. SAGE II (50o-60oN) SAGE II (40o-50oN) SAGE II (20o-40oN) SAGE II (0o-20oN). NO2 column (1015 cm-2). NO2 vertical column (1015 cm-2 ). NO2 slant column (1016 cm-2 ). 10. 8. sunset. 6. 2. 2000. 8.3 ± 1.6. Jungfraujoch FTIR. 0 1994. Trend 1996 – 2005.6. sunrise. 8. -11.2 +/- 2.4 % per decade (pm) -13.7 +/- 2.5 % per decade (am). 1992. Lauder (45.0oS) Macquarie Island (54.5oS) Arrival Heights (77.8oS). 0 1984. 1986. Year. 1988. 1990. 1992. 1994. 1996. 1998. 2000. 2002. 2004. 2006. 2008. 6. 5. (50o- 60oN). 4 4.4 +/- 1.5 % per decade (pm) 3 1984. 1986. 1988. 1990. 1992. 1994. 2010. Year. Figure 1: NO2 long-term time series of UV-visible spectroscopy measurements from seven ground-based stations are shown together with the FTIR NO2 data from Jungfraujoch. Displayed is the trend analysis for the longest time series available for each of the locations. The four Northern Hemisphere stations are shown on the left, the three Southern Hemisphere stations on the right; the bottom right panel shows the Jungfraujoch FTIR data (green). The monthly mean values are displayed as red (sunrise) and blue (sunset) stars, the thin solid line is the fitted model, the filled circles are the trend plus residual, and the thick solid line the linear trend. The reference date for the percentage change is the start of year 2000.. Discussion. 1996. Summary  The Lauder UV/visible (7.1%) and Jungfraujoch FTIR (3.1%) trends agree very well with the SAGE II 40-50oS and 40-50oN trends respectively (see table) for the longest overlapping time period (> 20 years). These trends are the most reliable ones and show for the ground-based stations as well as for SAGE II only positive trends with considerably higher positive trends in the SH.  The SAGE II data seem to show an increase in NO2 from the mid-80s to mid90s, a turn-around in the mid- to late 90s and possibly a thereafter.. 2000. 2002. 2004. 2006. 6. 5. (50o- 60oS). 4 5.5 +/- 1.6 % per decade (pm) 3 1984. 1986. 1988. 1990. 1992. 1994. Year. 5. (40o- 50oN). 4 3.5 +/- 1.9 % per decade (pm) 3 1984. 1986. 1988. 1990. 1992. 1994. 1996. 1998. 2000. 2002. 2004. 2006. (20o- 40oN). 4 3.1 +/- 1.4 % per decade (pm) 1986. 1988. 1990. 1992. 1994. 5. 1996. 1998. 2000. 2002. 2004. 2006. (0o- 20oN). 3 5.9 +/- 2.2 % per decade (pm) 1986. 1988. 1990. 1992. 1994. 2004. 2006. 2004. 2006. 2004. 2006. 2004. 2006. (40o- 50oS). 6.5 +/- 2.0 % per decade (pm) 3 1984. 1986. 1988. 1990. 1992. 1994. 1996. 1998. 2000. 2002. 6. 5. (20o- 40oS). 4 6.3 +/- 1.8 % per decade (pm) 3 1984. 1986. 1988. 1990. 1992. 1994. 1996. 1998. 2000. 2002. Year. 5. 2 1984. 2002. 4. Year. 4. 2000. Year. 6. 3 1984. 1998. 6. Year. 5. 1996 Year. 6. 1996 Year. Our analysis highlights that the calculated NO2 trends can depend strongly on the chosen time period and in some cases on the season (not shown here). For this study, we took care to only include ground-based data sets from stations for which we know that the data analysis is done consistently over the whole time period. All data series used in this study seem to be long enough for reasonable orthogonality between basis functions, so that the trends as quoted are sufficiently robust against variations in the model, and statistically significant within the model used.. 1998. NO2 column (1015 cm-2). 1994. NO2 column (1015 cm-2). 1992. o. NO2 column (1015 cm-2). o. -7.6 +/- 2.4 % per decade (am) 0 1990. Kiruna (67.8oN) Jungfraujoch (46.5oN) Moshiri (44.4oN) Mauna Loa (19.5oN). 6. 2 -7.2 +/- 1.8 % per decade (pm). Trend 1991 – 2005.6. sunset. 1998. 2000. 2002. 2004. 2006. NO2 column (1015 cm-2). 4. sunrise. 8. NO2 column (1015 cm-2). 6. NO2 column (1015 cm-2). 8. Site. NO2 column (1015 cm-2). NO2 slant column (1016 cm-2 ). 5.6 +/- 3.0 % per decade (am). 5. 4. (0o- 20oS). 3 7.9 +/- 3.0 % per decade (pm) 2 1984. 1986. 1988. 1990. 1992. 1994. 1996. 1998. 2000. 2002. Year. Figure 2: SAGE II sunset NO2 monthly mean trends for NH and SH zonal means over 20 years (1985 – 2005). The filled circles show the linear trend plus residual and the solid line shows the linear trend..  The hemispheric difference seen at the ground-based stations for NO2 trends calculated for 1991 – 2005.6 is not confirmed as clearly by SAGE II for 2 reasons: firstly, most of the calculated SAGE II trends are not statistically significant within the 95% level and secondly, since the start date is so close to the Pinatubo eruption and the SAGE NO2 measurements have not been carefully evaluated for the impact of lingering volcanic aerosols, the trends have to be viewed with some caution.  There is a distinct difference between the strong negative trends seen in the UV-visible data at the Jungfraujoch station and the statistically not significantly different from zero trends seen in the FTIR data set for the same time period. This will be investigated further.  This study shows a more complicated picture than anticipated and we will continue to investigate the difference in the trends between hemispheres and different data sets in more detail. We would like to acknowledge the use of the NDACC data base for the ground-based column NO2..

(2)

Figure

Figure 1:  NO 2 long-term time series of UV-visible spectroscopy measurements from seven  ground-based stations are shown together with the FTIR NO 2 data from Jungfraujoch

Références

Documents relatifs

In the present study we observed a higher expression of NR2A mRNA in the left medial striatum and of vGluT1 mRNA in the left Cg in healthy rats.. This suggests a hemispheric

The left-hand side of Figure 6 depicts for three sub-cubes the faces (in dark) delimiting the feasible regions of the individual reciprocal preference relations satisfying the

We interpret this result – the lower efficiency of public investment on output when the ‘Code for Fiscal Stability’ is removed from the sample – as a first, though rough, evidence

ةياور « كوكناب ينطايش » قازرلا دبعل ةّيرهاوط ميقلاو ينماضلما في ةسارد فوّصلاوب ظيفلحا دبع ةعماج – ةليم – ...:عجرلما ... يبرعلا بدلأاو ةغّللا

Furthermore, inspired from the proof in dependent theories as well as using techniques developed for almost centralizers in this thesis, we are able to find definable envelopes up

Key in the interpretation and understanding of WISDOMs ground penetrating RADAR (GPR) measurements is the capability to correctly (and efficiently) simulate the instrument

Coupling clay swelling properties to hydro-geochemical conditions: a reactive transport modeling approach.. Christophe Tournassat, Stéphane Gaboreau,

These various instruments present different bandwidth and detector densities: AERA measures showers at frequencies between 30 MHz and 80 MHz, while CODALEMA/EXTASIS bandwidth