K. K. Kelly

Reactive nitrogen and its correlation with ozone in the lower stratosphere and upper troposphere

Reactive nitrogen (NOy) and O3 were measured simultaneously from a NASA ER-2 aircraft during 1987 through 1989. These high resolution measurements cover a broad range of latitudes in the upper troposphere and lower stratosphere. The data show a striking positive correlation between NOy and O3 in the lower stratosphere at all scales sampled. The ratio NOy/O3 is nearly independent of altitude from the tropopause to above 20 km, with ratios in the stratosphere of 0.0015–0.002 in the tropics and 0.0025–0.004 elsewhere. The ratio is much more constant than either individual species, thus providing an excellent reference point for comparing limited data sets with models. Two-dimensional models reproduce general features of the vertical profile of NOy/O3 but not the gradient in the lower stratosphere between tropics and mid-latitudes. NOy and O3 are better correlated in the lower stratosphere than in the upper troposphere. The magnitude and variability of both NOy mixing ratios and NOy/O3 ratios indicate a source of NOy in the upper troposphere. The most plausible source in the tropics is lightning production of NOx. Condensation of NOy onto aerosol particles is often possible in the tropical upper troposphere and may play a role in determining the vertical distribution of NOy. In the mid-latitude upper troposphere the data suggest long-range transport of NOy. NOy mixing ratios in the tropical upper troposphere were usually between 150 and 600 pptv, enough so that upward transport can affect the NOy abundance in the tropical lower stratosphere.

Stratospheric water vapor increases over the past half‐century

Ten data sets covering the period 1954–2000 are analyzed to show a 1%/yr increase in stratospheric water vapor. The trend has persisted for at least 45 years, hence is unlikely the result of a single event, but rather indicative of long-term climate change. A long-term change in the transport of water vapor into the stratosphere is the most probable cause.

Hemispheric asymmetries in water vapor and inferences about transport in the lower stratosphere

Both satellite water vapor measurements and in situ aircraft measurements indicate that the southern hemisphere lower stratosphere is drier than that of the northern hemisphere in an annual average sense. This is the result of a combination of factors. At latitudes poleward of ∼50°S, dehydration in the Antarctic polar vortex lowers water vapor mixing ratios relative to those in the north during late winter and spring. Equatorward of ∼50°S, water vapor in the lower stratosphere is largely controlled by the tropical seasonal cycle in water vapor coupled with the seasonal cycle in extratropical descent. During the tropical moist period (June, July, and August), air ascending in the Indian monsoon region influences the northern hemisphere more than the southern hemisphere, resulting in a moister northern hemisphere lower stratosphere. This tropical influence is confined to levels beneath 60 mbar at low latitudes, and beneath 90 mbar at high latitudes. During the tropical dry period (December, January, and February), dry air spreads initially into both hemispheres. However, the stronger northern hemisphere wintertime descent that exists relative to that of southern hemisphere summer transports the dry air out of the northern hemisphere lower stratosphere more quickly than in the south. This same hemispheric asymmetry in winter descent (greater descent rates during northern hemisphere winter than during southern hemisphere winter) brings down a greater quantity of “older” higher water vapor content air in the north, which also acts to moisten the northern hemisphere lower stratosphere relative to the southern hemisphere. These factors all act together to produce a drier southern hemisphere lower stratosphere as compared to that in the north. The overall picture that comes from this study in regards to transport characteristics is that the stratosphere can be divided into three regions. These are (1) the “overworld” where mass transport is controlled by nonlocal dynamical processes, (2) the “tropically controlled transition region” made up of relatively young air that has passed through (and been dehydrated by) the cold tropical tropopause, and (3) the stratospheric part of the “middleworld” or “lowermost stratosphere”, where troposphere-stratosphere exchange can occur adiabatically. Satellite water vapor measurements indicate that the base of the “overworld” is near 60 mbar in the tropics, or near the 450 K isentropic surface.

Mixing of polar vortex air into middle latitudes as revealed by tracer‐tracer scatterplots

The occurrence of mixing of polar vortex air with midlatitude air is investigated by examining the scatterplots of insitu measurements of long-lived tracers from the NASA ER-2 aircraft during the Stratospheric Photochemistry, Aerosols and Dynamics Expedition (SPADE, April, May 1993; northern hemisphere) and the Airborne Southern Hemisphere Ozone Experiment / Measurements for Assessing the Effects of Stratospheric Aircraft (ASHOE/MAESA, March-October 1994; southern hemisphere) campaigns. The tracer-tracer scatterplots from SPADE form correlation curves which differ from those measured during previous aircraft campaigns (Airborne Antarctic Ozone Experiment (AAOE), Airborne Arctic Stratospheric Experiments I (AASE I) and II (AASE II)). It is argued that these anomalous linear correlation curves are mixing lines resulting from the recent mixing of polar vortex air into the middle latitude environment. Further support for this mixing scenario is provided by contour advection calculations and calculations with a simple one-dimensional strain-diffusion model. The scatterplots from the midwinter deployments of ASHOE/MAESA are consistent with those from previous midwinter measurements (i.e., no mixing lines), but the spring CO 2 :N 2 O scatterplots form altitude-dependent mixing lines which indicate that air from the vortex edge region (but not from the inner vortex) is mixing with midlatitude air during this period. These results suggest that at altitudes above about 16 km the mixing of polar vortex air into middle latitudes varies with season: in northern and southern midwinter this mixing rarely occurs, in southern spring mixing of vortex-edge air occurs, and after the vortex breakup mixing of inner vortex air occurs.

Gravity waves generated by a tropical cyclone during the STEP tropical field program: A case study

Overflights of a tropical cyclone during the Australian winter monsoon field experiment of the Stratosphere-Troposphere Exchange Project (STEP) show the presence of two mesoscale phenomena: a vertically propagating gravity wave with a horizontal wavelength of about 110 km and a feature with a horizontal scale comparable to that of the cyclones entire cloud shield (wavelength of 250 km or greater). The larger feature is fairly steady, though its physical interpretation is ambiguous. The 110-km gravity wave is transient, having maximum amplitude early in the flight and decreasing in amplitude thereafter. Its scale is comparable to that of 100-to 150-km-diameter cells of low satellite brightness temperatures within the overall cyclone cloud shield; these cells have lifetimes of 4.5 to 6 hours. Aircraft flights through the anvil show that these cells correspond to regions of enhanced convection, higher cloud altitude, and upwardly displaced potential temperature surfaces. A three-dimensional transient linear gravity wave simulation shows that the temporal and spatial distribution of meteorological variables associated with the 110-km gravity wave can be simulated by a slowly moving transient forcing at the anvil top having an amplitude of 400–600 m, a lifetime of 4.5–6 hours and a size comparable to the cells of low brightness temperature. The forcing amplitudes indicate that the zonal drag due to breaking mesoscale transient convective gravity waves is definitely important to the westerly phase of the stratopause semiannual oscillation and possibly important to the easterly phase of the quasi-biennial oscillation. There is strong evidence that some of the mesoscale gravity waves break below 20 km as well. The effect of this wave breaking on the diabatic circulation below 20 km may be comparable to that of above-cloud diabatic cooling.

Emission Measurements of the Concorde Supersonic Aircraft in the Lower Stratosphere

Emission indices of reactive gases and particles were determined from measurements in the exhaust plume of a Concorde aircraft cruising at supersonic speeds in the stratosphere. Values for NOx (sum of NO and NO2) agree well with ground-based estimates. Measurements of NOx and HOx indicate a limited role for nitric acid in the plume. The large number of submicrometer particles measured implies efficient conversion of fuel sulfur to sulfuric acid in the engine or at emission. A new fleet of supersonic aircraft with similar particle emissions would significantly increase stratospheric aerosol surface areas and may increase ozone loss above that expected for NOx emissions alone.

A Novel Method for Estimating Light-Scattering Properties of Soot Aerosols Using a Modified Single-Particle Soot Photometer

A Single-Particle Soot Photometer (SP2) detects black refractory or elemental carbon (EC) in particles by passing them through an intense laser beam. The laser light heats EC in particles causing them to vaporize in the beam. Detection of wavelength-resolved thermal radiation emissions provides quantitative information on the EC mass of individual particles in the size range of 0.2–1 μm diameter. Non-absorbing particles are sized based on the amount of light they scatter from the laser beam. The time series of the scattering signal of a non-absorbing particle is a Gaussian, because the SP2 laser is in the TEM00 mode. Information on the scattering properties of externally and internally mixed EC particles as detected by the SP2 is lost in general, because each particle changes size, shape, and composition as it passes through the laser beam. Thus, scattered light from a sampled EC particle does not yield a full Gaussian waveform. A method for determining the scattering properties of EC particles using a two-element avalanche photodiode (APD) is described here. In this method, the Gaussian scattering function is constructed from the leading edge of the scattering signal (before the particle is perturbed by the laser), the Gaussian width, and the location of the leading edge in the beam derived from the two-element APD signal. The method allows an SP2 to determine the scattering properties of individual EC particles as well as the EC mass. Detection of polystyrene latex spheres, well-characterized EC particles with and without organic coatings, and Mie scattering calculations are used to validate the method.

Measurements of high number densities of ice crystals in the tops of tropical cumulonimbus

Imaging and light scattering instruments were used during the January/February 1987 STEP Tropical Experiment at Darwin, Australia, to measure ice crystal size distributions in the tops of tropical cumulonimbus anvils associated with tropical cyclones and related cloud systems. Two light scattering instruments covered particles from 0.1-μm to 78-μm diameter. Particles larger than 50-μm diameter were imaged with a two-dimensional Grey optical array imaging probe. The measurements were made at altitudes ranging from 13 to 18 km at temperatures ranging from −60° to −90°C. Additional measurements made in continental cumulonimbus anvils in the western United States offer a comparative data set. The tropical anvil penetrations revealed surprisingly high concentrations of ice crystals. Number densities were typically greater than 10 cm−3 with up to 100 cm−3 if one includes all particles larger than 0.1 μm and can approach condensation nuclei in total concentration. In order to explain the high number densities, ice crystal nucleation at altitude is proposed with the freezing of fairly concentrated solution droplets in equilibrium at low relative humidities. Any dilute liquid phase is hypothesized to be transitory with a vanishingly short lifetime and limited to cloud levels nearer −40°C. Homogeneous nucleation of ice involving H2SO4 nuclei is attractive in explaining the high number densities of small ice crystals observed near cloud top at temperatures below −60°C. The tropical size distributions were converted to mass using a spherical equivalent size, while the continental anvil data were treated as crystalline plates. Comparisons of the ice water contents integrated from the mass distributions with total water contents measured with NOAA Lyman-alpha instruments require bulk densities equivalent to solid ice for best agreement. Correlation between the two data sets for a number of flight passes was quite good and was further improved by subtraction of water vapor density values ranging between ice and water saturation. Ice water contents up to 0.07 g m−3 were observed in the tropical anvils with over 0.1 g m−3 in continental anvils. The size distributions in tropical anvils generally reveal mass modes at sizes of 20–40 μm. With rare exceptions, particles larger than 100 μm were not observed near the cloud tops. In continental cumulonimbus anvils, much larger plate crystals approaching 1 mm in size account for the majority of the ice water. Most of the ice crystal mass lofted to anvil altitudes falls to lower levels prior to evaporating. The anvils can experience strong radiational heating as well as cooling depending upon lower cloud cover, particle size distribution, and time of day.

Chemical loss of ozone in the arctic polar vortex in the winter of 1991-1992.

In situ measurements of chlorine monoxide, bromine monoxide, and ozone are extrapolated globally, with the use of meteorological tracers, to infer the loss rates for ozone in the Arctic lower stratosphere during the Airborne Arctic Stratospheric Expedition II (AASE II) in the winter of 1991-1992. The analysis indicates removal of 15 to 20 percent of ambient ozone because of elevated concentrations of chlorine monoxide and bromine monoxide. Observations during AASE II define rates of removal of chlorine monoxide attributable to reaction with nitrogen dioxide (produced by photolysis of nitric acid) and to production of hydrochloric acid. Ozone loss ceased in March as concentrations of chlorine monoxide declined. Ozone losses could approach 50 percent if regeneration of nitrogen dioxide were inhibited by irreversible removal of nitrogen oxides (denitrification), as presently observed in the Antarctic, or without denitrification if inorganic chlorine concentrations were to double.

Polar stratospheric cloud processed air and potential voracity in the northern hemisphere lower stratosphere at mid‐latitudes during winter

Small-scale (<1000 km) features in ER-2 measurements of ClO, O3, H2O, N2O, and NOy, outside the lower stratospheric Arctic vortex of 1988–1989 are compared with features on potential vorticity maps from the European Centre for Medium-range Weather Forecasts (ECMWF). The potential vorticity maps are obtained from Tl06 analyses and forecasts. Some of the plots have been truncated to lower resolution (T63 or T42) which smooths out the finer-scale structure. Comparison of these lower resolution plots shows how much detail is lost by excessive smoothing. It is also evident that the forecast plots lose fine-scale structure due to dissipation in the model resulting mainly from horizontal diffusion. We conclude that blobs of air on the maps at latitudes between the vortex edge and 25°N having potential vorticities characteristic of the vortex, did indeed originate from the vortex, but that the real atmosphere is more sharply differentiated (inhomogeneous) than the meteorological analyses, implying that the potential vorticity maps underestimate the amount of peeled-off material. Areal budgets of the ex-vortex air are considered for ER-2 flight days, and are performed for 24-hour forecasts at T63, and analyses at T42, T63, and T106 resolution at θ = 475 K. Finally, it is concluded that the lower stratospheric Arctic vortex of 1988–1989 spread considerable amounts of air to mid-latitudes which had been processed by polar stratospheric clouds, and that this mechanism is a realistic explanation for the wintertime loss of ozone observed over northern mid-latitudes during the last decade.