In situ water vapor and ozone measurements in Lhasa and Kunming during the Asian summer monsoon

Authors


Abstract

[1] The Asian summer monsoon (ASM) anticyclone circulation system is recognized to be a significant transport pathway for water vapor and pollutants to enter the stratosphere. The observational evidence, however, is largely based on satellite retrievals. We report the first coincident in situ measurements of water vapor and ozone within the ASM anticyclone. The combined water vapor and ozonesondes were launched from Kunming, China in August 2009 and Lhasa, China in August 2010. In total, 11 and 12 sondes were launched in Kunming and Lhasa, respectively. We present the key characteristics of these measurements, and provide a comparison to similar measurements from an equatorial tropical location, during the Tropical Composition, Cloud and Climate Coupling (TC4) campaign in July and August of 2007. Results show that the ASM anticyclone region has higher water vapor and lower ozone concentrations in the upper troposphere and lower stratosphere than the TC4 observations. The results also show that the cold point tropopause in the ASM region has a higher average height and potential temperature. The in situ observations therefore support the satellite-based conclusion that the ASM is an effective transport pathway for water vapor to enter stratosphere.

1. Introduction

[2] The Asian summer monsoon (ASM) anticyclone is a prominent circulation system in the Northern Hemisphere upper troposphere and lower stratosphere (UTLS) in the summer season. Numerous studies, using satellite observations, show that the region within the anticyclone is associated with significantly enhanced tropospheric tracers, such as carbon monoxide [Filipiak et al., 2005; Li et al., 2005; Fu et al., 2006; Park et al., 2007], biomass burning products, such as HCN [Randel et al., 2010], aerosols [Vernier et al., 2011], and water vapor [Rosenlof et al., 1997; Gettelman et al., 2004]. Based on CALIPSO satellite data, significant fractions of cirrus cloud are found to be above the tropopause level in this region [Pan and Munchak, 2011]. In addition, numerical simulations identify the ASM circulation to be a prominent boreal summer pathway of tropospheric air into the tropical stratosphere [Bannister et al., 2004; Gettelman et al., 2004; Wright et al., 2011].

[3] The studies above have used satellite data and models, which often do not provide sufficiently resolved details for characterizing the transport process. In situ measurements inside the anticyclone are scarce [Zheng et al., 2004]. In this letter, we report the first coincident in situ measurements of water vapor and ozone in the region of the anticyclone and present the measurement highlights. These new high-resolution and high-accuracy water vapor data are significant for quantifying the moisture transport associated with the ASM anticyclone, and the paired measurements of water vapor and ozone provide additional information for identifying the transport pathway. Results from the two locations within the ASM anticyclone are compared to differentiate the conditions in the center from those near the edge of the anticyclone. Similar measurements from an equatorial tropical location are used as a reference to contrast the region of anticyclone with the deep tropics. Additional detailed analyses and modeling efforts using these data are ongoing.

2. Campaign Information

[4] Figure 1 highlights climatological features of the ASM anticyclone using 31 year (1980–2010) July–August National Center for Environmental Prediction (NCEP)/ National Center for Atmospheric Research (NCAR) reanalysis fields. The anticyclone is denoted by the 16770 m geopotential height (GPH) contour at 100 hPa and 14350 m GPH contour at 150 hPa. These GPH contours are empirically selected anticyclone boundaries. The 3° north–south shift between the two contour levels shows the northward tilting behavior of the anticyclone. The ASM anticyclone spans the subtropical Asian continent: from the east coast of China (∼120°E) to the east shore of the Mediterranean Sea (∼30°E), between 20°N–40°N. Also shown in the figure are locations of low outgoing longwave radiation (OLR), highlighting the region of deep convection and associated upper tropospheric heating that initiates and maintains the anticyclone [Hoskins and Rodwell, 1995], and a low temperature region (T < = 195°K) at the 100 hPa level, illustrating the typical temperature response to the heating [Gettelman and Birner, 2007]. The region of the anticyclone also exhibits a higher tropopause (i.e. a lower tropopause pressure). The region of lowest tropopause pressure is located over the southern Tibetan Plateau (represented by the 95 hPa tropopause pressure contour), baring the importance of topography. While the figure represents the climatology on a seasonal time scale, it is known that the core of the ASM anticyclone has a marked east-west oscillation, where it alternates between a Tibetan plateau mode and an Iranian plateau mode, on a quasi-biweekly time scale [Zhang et al., 2002].

Figure 1.

31 year (July–August, 1980–2010) climatology of the ASM anticyclone and the campaign locations. The anticyclone is shown using GPH at 100 hPa (solid black contours, 16.77 km) and 150 hPa (dashed black, 14.35 km), and the 100 hPa wind vectors (purple). Also shown are selected contours for tropopause pressure (red, hPa), outgoing longwave radiation (OLR) (filled blue, Wm−2), and temperature at 100 hPa (blue, 195°K). Grey shading identifies terrain with altitude over 3 km. NCEP/NCAR reanalyses fields are used.

[5] Two campaigns are conducted, one at Kunming (KM) (25.01°N, 102.65°E, elevation 1,889 m, WMO station number 56778) in summer of 2009 and the other at Lhasa (LH) (29.66°N, 91.14°E, elevation 3,650 m, WMO station number 55591) in summer of 2010. Climatologically, KM is located at the northeast corner of the low OLR area and within the southeast edge of the ASM anticyclone (Figure 1). Because the position of the anticyclone is variable, it is important to note KM was inside the anticyclone limit most times during the campaign but influenced by the air mass from outside. In contrast, LH is located slightly east of the ASM anticyclone center on the Tibetan Plateau, and is consistently within the anticyclone limit.

[6] In both campaigns, balloon borne payloads are launched to measure water vapor, ozone, and meteorological variables. The payload used at both sites consists of a Cryogenic Frostpoint Hygrometer (CFH) [Vömel et al., 2007] and an electrochemical concentration cell (ECC) ozonesonde, while Vaisala RS80 radiosondes [Bian et al., 2011] were used at KM and iMet radiosondes from International Met Systems (www.intermetsystems.com) were used at LH. Thus, every flight provides observations of water vapor, ozone, pressure, temperature, and wind between the surface and the middle stratosphere. The water vapor uncertainty is estimated to be ∼10% [Vömel et al., 2007]. The ozone uncertainty is estimated to be better than 5–10% [Smit et al., 2007]. In addition, three sondes at LH were launched with a Compact Optical Backscatter AerosoL Detector (COBALD) made in Eidgenössische Technische Hochschule (ETH) Zürich. This instrument observes cirrus clouds and aerosols; results will be introduced in future publications.

[7] Detailed summaries of the launches are provided in Tables S1 and S2, included in the auxiliary material. A total of 11 soundings were launched during the KM campaign (August 7–13, 2009, see Table S1). Seven soundings were launched in the daytime (Beijing local time 14:00), and four soundings at night (Beijing local time 19:00). A total of 12 soundings were launched from LH (August 22–28, 2010, see Table S2). Launch times were irregular due to the effort to avoid frequent rain showers during this period.

3. Key Characteristics of Water Vapor, Ozone, and Relative Humidity

[8] The water vapor, ozone, and relative humidity over ice (RH) profiles from both LH and KM measurements are presented in Figure 2. In each case, we show the mean profiles from both campaigns for comparisons. To put these measurements in the context of tropical UTLS observations, we also included the mean profiles from a campaign with similar measurements in an equatorial tropical location. The chosen tropical reference is the Tropical Composition, Cloud and Climate Coupling (TC4) campaign, which launched 16 sondes from Alajuela, Costa Rica (9.98°N, 84.21°W, 899 m) during July and August of 2007 [Selkirk et al., 2010]. Although not shown in this figure, it is important to note that the mean cold point tropopause (CPT) height and potential temperature during TC4 (16.7 km; 375.8 K) are lower than that at KM (17.8 km; 390.1 K) and LH (17.7 km; 389.2 K), which have similar means.

Figure 2.

(left) Water vapor mixing ratio, (middle) ozone mixing ratio, and (right) RH profiles at (top) KM and (bottom) LH. Point measurements are shown in grey, and the mean profiles are shown as red lines. Black lines indicate the mean profile for the other location to provide a comparison between KM and LH. The light blue lines show the mean profile from TC4. See text for discussions of the two outlier profiles, shown in purple and green. The orange histograms in the RH plots show the layer fraction of super-saturation over ice. The mean CPT level is shown by the blue dashed line, with the bar indicating its 10–90% variation.

[9] On average, water vapor profiles (Figure 2, left) at KM and LH have a similar structure. A number of outliers are evident in KM profiles, which are consistent with KM's position at the anticyclone edge where it is under the influence of multiple transport pathways. The most prominent outlier profile (shown in purple in Figures 2 and 4) has much lower water vapor mixing ratios (<100 ppmv) and relatively high ozone in the middle troposphere. Detailed analysis of this case found that the sounding sampled a filament of a stratospheric intrusion from midlatitudes (L. L. Pan et al., manuscript in preparation, 2012). A second outlier profile (shown in green) is evident with extremely low water vapor (∼2 ppmv) and low ozone (∼35 ppbv) near the CPT. This outlier will be discussed later in association with Figure 4. On average, both KM and LH have higher water vapor mixing ratios than the TC4 observations throughout the troposphere and stratosphere. Between the two locations in the anticyclone, LH shows higher water vapor mixing ratios. A wetter stratosphere at LH supports the satellite studies that find enhanced water vapor and other tropospheric tracers within the ASM anticyclone. At 100 hPa, water vapor mixing ratios are 6.0 and 6.4 ppmv at KM and LH, respectively. These results qualitatively support the water vapor observations from MLS that show concentrations greater than 5.4 ppmv within the anticyclone [Park et al., 2007], although the satellite retrieval-based means show a dry bias.

[10] Ozone profiles (Figure 2, middle) at KM and LH differ significantly in the troposphere where, on average, LH measures lower ozone mixing ratios. These differences suggest that the tropospheric air over the plateau is less polluted. Figure 3 (top) provides a comparison of ozone measurements in the 7–9 km altitude range, illustrating the stark contrast in tropospheric ozone. The large ozone variation near the cold point tropopause (CPT) over KM suggests that the UTLS air near the anticyclone edge is of varying origins. Around the tropopause, ozone values from both Asian measurement sites are lower than the TC4 results, indicating the reduced concentrations of stratospheric tracer within the anticyclone. This result is also consistent with the satellite observations. Ozone mixing ratios at 100 hPa are 124.2 and 142.4 ppbv at KM and LH, respectively, supporting the results from MLS measurements that show ozone mixing ratios to be less than 160 ppbv within the anticyclone. These values are lower than the stratospheric average at the same level outside the anticyclone [Park et al., 2007].

Figure 3.

Frequency distributions of (top) ozone within the 7–9 km layer, (middle) RH within the 11–13 km layer, and (bottom) the level of lapse rate minimum for KM (orange) and LH (grey). Ozone and RH distributions are given in relative frequency, i.e., the distribution is normalized by the total number of observations in the layer. The LRM plot is given in absolute frequency.

[11] RH profiles (right column) are calculated with respect to ice. Mean profiles indicate that the RH is higher over LH. Super-saturation over ice is found at both sites, but LH measurements show a higher frequency of super-saturation, especially around the level of main convective outflow (∼13 km). To provide more details, the RH distributions in the 11–13 km layer are contrasted for the two locations inFigure 3 (middle). Also shown in Figure 3 (bottom) are distributions of the level of potential temperature lapse rate minimum (LRM). The LRM acts as a good proxy for the main convective outflow level [Gettelman and de F. Forster, 2002]. Together, Figure 3(middle and bottom) illustrates that the level of main convective outflow is higher at LH (12.8 km, 358.3 K) than KM (11.5 km, 354.3 K), and the distribution of RH is significantly different at the outflow level for the two locations. Climatologically, KM is more humid than LH. During the campaign year 2009, however, KM was under anomalous condition of drought. Day-to-day variability is also an important factor contributing to the result. To account for that, the weather conditions during the sounding are noted in the launch summary tables, provided in theauxiliary material.

[12] The ozone and water vapor relationships for the two campaigns are shown in Figure 4. A log-log scale is used to highlight the range of measurements. The two tracers have a familiar “L” shaped relationship in the tracer-tracer space [Pan et al., 2007], but the log-log scale slightly distorts the linear relationship in the stratosphere and troposphere. Two outlier profiles at KM, discussed previously, are marked with color. The most prominent outlier profile in the tracer-tracer space (green points inFigure 4) shows an anomalous positive ozone-water vapor correlation near the corner of “L,” represented by the low ozone and low water vapor data points. At the extreme point, the air mass has ∼35 ppbv of ozone and ∼2 ppmv of water vapor. The extremely low ozone value is evidence that the air parcel has experienced rapid deep convective ascent, while the extremely low water vapor value indicates that the air parcel has experienced a low temperature near the tropopause level. The estimated temperature that produces the corresponding saturation vapor mixing ratio is ∼187°K at 100 hPa. The transport history of the air mass involved in this case is investigated and details of the case study will be reported in a follow up paper. Briefly, a 5-day back-trajectory calculation identifies the air mass to be advected at approximately the 100 hPa level from the western Pacific in the vicinity of a synoptic scale cold region. The temperature (based on the NCEP analyses), however, is a few degrees warmer than that required to explain the observed water vapor mixing ratio. Possible association with wave perturbations and other errors in analyses will need to be investigated.

Figure 4.

The tracer-tracer relationship between water vapor and ozone at (top) KM and (bottom) LH in a log-log scale. The two outlier-profiles discussed in the text appear in color.

[13] Figure 5 provides quantitative analysis of the lower stratospheric water vapor from the two campaigns, contrasted with the deep tropical results from TC4. Distributions of water vapor from the three campaigns are compared at four isentropic layers. The lowest layer, in 375–400 K potential temperature range, contains the mean CPT at all three locations. The water vapor distributions in this layer largely overlap. Although the three campaigns have comparable mean CPT temperature (∼194 K), they have significant tropopause height differences. Because the CPT is at lower potential temperature levels in TC4 (average is ∼376 K), the TC4 data in this layer are largely stratospheric and the distribution shows a greater number of dry samples. Distributions of water vapor at KM and TC4 shift toward lower concentrations in the layers above the CPT (400–430 K and 430–470 K), consistent with the signature of the “tape recorder” [Mote et al., 1996] (i.e. a signature of water vapor from the previous winter). It is interesting to note that the “tape recorder” signature over KM (∼25°N) is fairly consistent with that over Alajuela (∼10°N), although in Mote et al., tape recorder signal is computed from 12°S to 12°N. The distribution of water vapor at LH (∼30°N) within the 400–430 K layer, however, remains similar to its distribution at the CPT layer (375–400 K). Moving to higher potential temperatures, the water vapor distribution at LH also shifts toward lower mixing ratios, but LH has a higher mean water vapor concentration at all levels.

Figure 5.

Frequency distributions of water vapor mixing ratio in four isentropic layers over KM (orange), LH (grey), and Costa Rica (red) are shown.

[14] The interesting differences in the lower stratospheric water vapor between LH and other two campaigns will be a topic of future studies. A number of factors may contribute to these differences. One of the complications is that, based on the tropopause height, LH (∼30°N) is located within the tropical belt during boreal summer but outside during boreal winter; thus, the stratospheric water vapor concentrations at this level are likely influenced by several different transport pathways. However, agreement between KM and TC4 is interesting given that measurements at TC4 are representative of a tropical site that is within the latitudinal range of satellite data used to derive the tape recorder signal (12°S–12°N from Mote et al. [1996]). Among many other factors, the Quasi-Biennial Oscillation (QBO) may contribute to the differences in water vapor during the winter phase, since the measurements are taken in different years. An estimate of QBO contribution at a similar level based on 50°S–50°N data was ∼0.5 ppmv [Randel, 2010], while observed differences between KM and LH are ∼1 ppmv. Continued multi-year measurements following these two campaigns may provide more information on processes contributing to the differences. Overall, the results support enhanced tropospheric tracers within the anticyclone at the UTLS level seen by satellite data. Given that this region has higher tropopause heights, the upper tropospheric air mass in the anticyclone is at a higher isentrope and can subsequently be transported into the stratosphere isentropically in both tropical and extratropical directions [Dethof et al., 1999].

4. Summary and Outlook

[15] The in situ measurements from KM and LH provide the first set of high vertical resolution, high-accuracy water vapor and ozone profiles inside the ASM anticyclone. Compared to similar measurements from the TC4 campaign in an equatorial location, the ASM results show higher water vapor and lower ozone in the UTLS. The results also show that the level of CPT over the Tibetan plateau during ASM is higher in altitude and potential temperature than in the deep tropics as represented by the TC4 results. These findings confirm the satellite retrieval based conclusion that the anticyclone is a significant transport pathway for water vapor to enter the stratosphere. The convectively lifted air mass within the ASM anticyclone may enter the tropical stratosphere isentropically, by-passing the tropical tropopause. Multi-year water vapor and ozone sounding campaigns are planned for the two ASM locations. These new in situ measurements will complement the satellite observations and facilitate convective transport studies using modeling tools.

Acknowledgments

[16] This work is supported by the National Basic Research Program of China (2010CB428602) and the National Natural Science Foundation of China (40830102 and 41175040). This work is also supported by the U.S. National Science Foundation through its support for the National Center for Atmospheric Research. The authors thank Yuejian Xuan, Jinqiang Zhang, Zhixuan Bai for their great efforts in the field campaign.

[17] The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.

Ancillary