Table of Contents * Previous Chapter * Next Chapter
Between February and November 1994 NASA's ER-2 aircraft made a total of 45 flights during the ASHOE (Airborne Southern Hemisphere Ozone Experiment) / MAESA (Measurements for Assessing the Effects of Stratospheric Aircraft) campaign. The flights covered the latitude range from 60N to 70S. As well as four deployments flying out of Christchurch New Zealand the ER-2 flew out of NASA Ames (California), Hawaii and Fiji. The ER-2 had two payload configurations: the so-called dynamical payload and the full chemistry payload. The main aim of the mission was to investigate the causes of ozone loss in the wintertime middle latitudes in the southern hemisphere.
As well as the longer lived tracers the ER-2 simultaneously measured an impressive range of short-lived species including OH, HO2, ClO, BrO, NO and NO2. These measurements will go a long way to constraining the fast chemistry of these species. Some notable measurements were made in early October in the polar vortex. At 400K the ER-2 sampled air with very low ozone (< 0.5ppmv) where all of the inorganic chlorine was in the form of HCl. There was also high NO (1ppbv) compared to total NOy (2ppbv). The implication here is that the ER-2 saw air in which ClO had recovered rapidly into HCl as a result of the low ozone - a result which had been predicted by models but not yet observed.
Martyn Chipperfield
During aircraft campaigns, there is an obvious need for detailed meteorological forecasts to aid in flight planning and immediate post flight interpretation. The Airborne Southern Hemisphere Ozone Experiment (ASHOE) aircraft campaign involved the use of the NASA ER-2 high altitude research aircraft. For this type of aircraft, there were two requirements of such a forecast; firstly, there was a need to know the local weather conditions for take-off and landing, and secondly, there was a need to know the position of large-scale dynamical structures in the stratosphere. This second requirement involved accurate positioning of the Austral vortex edge, and the possible location of filaments which might contain vortical air. In the past, the principle diagnostic has been the production of potential vorticity (PV) synoptic maps for various isentropic surfaces, generally associated with the cruise height (475 K) and dive height (400 K). These maps are used to indicate the position of the vortex edge, which is thought to be collocated with the steepest gradient found in the PV field. A second tool has been the use of Contour Advection with Surgery (CA/S) (Waugh & Plumb, 1993; Norton, 1993), with the aim to locate filamentary structure with PV values associated with vortical air.
However, both aim to give the location of air which is likely to have been processed on polar stratospheric clouds (PSCs). They assume that air with high absolute values of PV will have been activated if air of these values has seen very cold temperatures. It would be an improvement on the standard meteorological forecasts if we could predict the chemical structure in the stratosphere. A three-dimensional (3D) chemical forecast model is able to indicate the likely composition of processed air taking into account recovery into reservoir species.
To produce a chemical forecast we employed the following methodology. The UGCM was initialised using ECMWF analyses, extrapolating above the 10 mbar lid where necessary, and run for 10 days at T106. The winds and temperatures were then truncated to T42 resolution to force TOMCAT. The TOMCAT model was initialised chemically using a TOMCAT model run that was kept up to date using ECMWF analysed winds and temperatures. This model run had itself been initialised from a 2D model dataset. This run was on the levels prescribed by the ECMWF analyses, so we had to interpolate the chemical data onto the 23 model levels of the UGCM and extrapolate above the 10 mbar lid before we could run TOMCAT in forecast mode. The UGCM forecast was the slowest step in this procedure. Run on the DRAL YMP-8 Cray supercomputer, this step took 40000 CPU seconds (e3 queue). The 10-day TOMCAT chemical forecast took between 6000 CPU seconds at T21 (e4 queue) and 22000 CPU seconds at T42 (e7 queue). We were generally able to make the chemical forecasts available within 3 days of the forecast start date at mission headquarters in Christchurch, New Zealand.
During phase II of the ASHOE campaign, we produced one chemical forecast which was started at 18h00 May 26th and run at T21. This forecast successfully predicted the penetration of the chemically perturbed region (CPR) at the cruise altitude, even though the PV contours appeared quite slack in this region. Furthermore, the forecast predicted the tongue of air extending out of the vortex on the 400 K isentropic surface, which was sampled during the dive. This flight took off before dawn, and so ClO was only measured on penetration of the CPR.
The first figure shows a synoptic map of the ClO
amount on the 400 K isentropic surface, output at 00h00 GMT, which roughly corresponded to the middle
of the flight. Data from this output time was then linearly interpolated to form the flight track comparison
with the measured ER-2 data seen in the
second figure. The model data does not take account of the diurnal variation in the ClO, this accounts for the differences seen in the flight track comparison.
Adrian Lee, Glenn Carver, Martyn Chipperfield, and John Pyle.
The portfolio of off-line chemistry and transport models, developed by Martyn Chipperfield and used by the UGAMP community, employ the same accurate Prather numerical advection scheme (JGR, 6671-6681, 91, 1986). The scheme preserves tracer structures by conserving the second-order moments of the spatial distribution of tracer during advection. The tracer mixing ratio is represented within a rectilinear grid box by a second-order polynomial in three dimensions. The advection follows the basic upstream method, but applied to all moments to second-order. Prather noted it is often desirable of the advection method that the tracer mixing ratios remain positive even in the vicinity of large discontinuities in tracer abundance. He described a simple positive definite fix, which maintained positive tracer concentrations. To do this limits are placed on the high-order moments in terms of the total amount of tracer (zeroth moment), essentially rearranging the moments to remove any negative found in the box. This has the effect of introducing some internal diffusion.
During the ASHOE campaign, TOMCAT and SLIMCAT were used to forecast stratospheric chemical tracers. In diagnosing observational data, pairs of long-lived tracers are frequently found to display compact relationships in the lower stratosphere. Plumb & Ko (JGR, 10145-10156, 97, 1992) derived a theoretical basis in the context of a 2D model. However, when we diagnosed TOMCAT in a similar manner we found that the inter-relationships between certain pairs of tracers were diverging from those in the initial dataset. The largest deviation was found in relationships between long-lived tracers (like N2O) and total inorganic chlorine (Cly).
The figure shows Cly versus N2O for the 21st May, 8 days into the run, with
and without the positive definite fix. The breakdown is seen at the smallest N2O values, which occur in
and around the austral polar vortex. It was found that the Prather fix was turning on around the austral
polar vortex edge, where there were sharp gradients in the chlorine-containing tracers. By rearranging the
distribution of the tracer within a grid box, the fix results in a loss of consistency with this tracer and other
tracers during the advection step. It was concluded that in order to preserve the inter-relationships between
the long-lived tracers the advection must be independent of species distribution.
Adrian Lee
Table of Contents * Previous Chapter * Next Chapter