Engineering topics

Einaudi, Franco

Publications and source records attributed to Einaudi, Franco.

At least 109 records · Page 6

NASA NTRS2001

Weather Prediction Improvement Using Advanced Satellite Technology

We discuss in this paper some of the problems that exist today in the fall utilization of satellite data to improve weather forecasts and we propose specific recommendations to solve them. This discussion can be viewed as an aspect of the general debate on how best to organize the transition from research to operational satellites and how to evaluate the impact of a research instrument on numerical weather predictions. A method for providing this transition is offered by the National Polar-Orbiting Operational Environmental Satellite System (NPOESS) Preparatory Project (NPP). This mission will bridge the time between the present NOAA and Department of Defense (DOD) polar orbiting missions and the initiation of the converged NPOESS series and will evaluate some of the Earth Observing System (EOS) instruments as appropriate for operational missions. Thus, this mission can be viewed as an effort to meet the operational requirements of NOAA and DOD and the research requirements of NASA. More generally, however, it can be said that the process of going from the conception of new, more advanced instruments to their operational implementation and full utilization by the weather forecast communities is not optimal. Instruments developed for research purposes may have insufficient funding to explore their potential operational capabilities. Furthermore, instrument development programs designed for operational satellites typically have insufficient funding for assimilation algorithms needed to transform the satellite observations into data that can be used by sophisticated global weather forecast models. As a result, years often go by before satellite data are efficiently used for operational forecasts. NASA and NOAA each have unique expertise in the design of satellite instruments, their use for basic and applied research and their utilization in weather and climate research. At a time of limited resources, the two agencies must combine their efforts to work toward common goals of full utilization of satellite data. This is a challenge that requires the assimilation of myriad new data into increasingly sophisticated numerical forecast models that run on increasingly sophisticated computer systems. In section II, we briefly outline the impact of satellite data on the quality of the National Centers for Environmental Prediction (NCEP) forecasts. In section III, we describe the present status of the utilization of satellite data in NCEP models and the challenges that lie ahead. In section IV, we propose solutions whose goals are summarized in section V.

Einaudi, Franco↗

NASA NTRS2001

TOMS Tropical Tropospheric Ozone Data Sets at the University of Maryland Website

Since 1997, shortly after the launch of the Earth-Probe TOMS (Total Ozone Mapping Spectrometer) satellite instrument, we have been processing data in near-real time to post maps of tropical tropospheric ozone at a website: metosrv2.umd.edu/-tropo. Daily, 3-day and 9-day averages of tropical tropospheric ozone column depth (TTO) are viewable from 10N to 10S. Data can be downloaded (running 9-day means) from 20N-30S. Pollution events are trackable along with dynamically-induced variations in tropospheric ozone column. TOMS smoke aerosol (toms.gsfc.nasa.gov) can be used to interpret biomass burning ozone, as for example, during the extreme ozone and smoke pollution period during the ENSO-related fires of August November 1997. During that time plumes of ozone and smoke were frequently decoupled and ozone from Indonesian fires and from Africa merged in one large feature by late October 1997. In addition to the Earth-Probe TOMS record, data as half-month averages and as daily 9-day means from the Nimbus 7 TOMS instrument are at the metosrv2.umd.edu/-tropo website. A guide to the website and examples of ozone time-series and maps will be shown.

Kochhar, A. K.↗

NASA NTRS2001

Model of Wave Driven Flow Oscillation for Solar Cycle

At low latitudes in the Earth's atmosphere, the observed zonal flow velocities are dominated by the semi-annual and quasi-biennial oscillations with periods of 6 months and 20 to 32 months respectively. These terrestrial oscillations, the SAO and QBO respectively, are driven by wave-mean flow interactions due to upward propagating planetary-scale waves (periods of days) and small-scale gravity waves (periods of hours). We are proposing (see also Mayr et al., GRL, 2001) that such a mechanism may drive long period oscillations (reversing flows) in stellar and planetary interiors, and we apply it to the Sun. The reversing flows would occur below the convective envelope where waves can propagate. We apply a simplified, one dimensional, analytical flow model that incorporates a gravity wave parameterization due to Hines (1997). Based on this analysis, our estimates show that relatively small wave amplitudes less than 10 m/s can produce zonal flow amplitudes of 20 m/s, which should be sufficient to generate the observed variations in the magnetic field. To produce the 22-year period of oscillation, a low buoyancy frequency must be chosen, and this places the proposed flow in a region that is close to (and below) the base of the convective envelope. Enhanced turbulence associated with this low stability should help to generate the dynamo currents. With larger stability at deeper levels in the solar interior, the model can readily produce also oscillations with much longer periods. To provide an understanding of the fluid dynamics involved, we present numerical results from a 2D model for the terrestrial atmosphere that exemplify the non-linear nature of the wave interaction for which a mechanical analog is the escapement mechanism of the clock.

Mayr, Hans G.↗

NASA NTRS2001

Retrieval of Polar Stratospheric Cloud Microphysical Properties From Lidar Measurements: Dependence on Particle Shape Assumptions

Knowledge of particle sizes and number densities of polar stratospheric clouds (PSCs) is highly important, because they are critical parameters for the modeling of the ozone chemistry of the stratosphere. In situ measurements of PSC particles are rare. the main instrument for the accumulation of PSC data are lidar systems. Therefore the derivation of some microphysical properties of PSCS from the optical parameters measured by lidars would be highly beneficial for ozone research. Inversion of lidar data obtained in the presence of PSCs formed from crystalline particles type 11 and the various nitric acid tri Ydrrate (NAT) types cannot be easily accomplished, because a suitable scattering theory for small faceted crystals has not been readily available tip to now. As a consequence, the T-matrix method is commonly used for the interpretation of these PSC lidar data. Here the assumption is made that the optical properties of an ensemble of spheroids resemble those of crystalline PSCs, and microphysical properties of the PSC are inferred from the optical signatures of the PSC at two or more wavelengths. The problem with the T-matrix approach is that the assumption of spheroidal instead of faceted particles can lead to dramatically wrong results: Usually cloud particle properties are deduced from analysis of lidar profiles of backscatter ratio and depolarization ratio. The particle contribution to the backscatter ratio is given by the product of the particle number density and the backscattering cross section. The latter is proportional to the value of the particle's scattering phase function at 180 degrees scattering angle. At 180 degrees however, the phase functions of rough, faceted crystals and of spheroids with same maximum dimension differ by a factor of 6. From this it follows that for a PSC consisting of faceted crystals, the particle number density is underestimated by roughly the same factor if spheroidal particles are unrealistically assumed. We are currently developing a retrieval technique for determining the microphysical parameters of crystalline PSCs that takes into account the faceted shape of the PSC particles. This approach utilizes finite-difference time-domain (FDTD) calculations of particle optical properties. The accuracy and the free choice of the shape of the scattering particle make the FDTD technique a promising tool for the inversion of PSC lidar data. A first comparison of FDTD and T-matrix calculations will be presented.

Reichardt, Susanne↗

NASA NTRS2001

Tropical Tropospheric Ozone and Smoke Interactions: Satellite Observations During the 1997 Indonesian Fires

Biomass burning generates hydrocarbons, nitrogen oxides and carbon monoxide that lead to tropospheric ozone pollution. Other combustion products form soot and various aerosol particles that make up smoke. Since early 1997 smoke and tropospheric ozone have been monitored in real-time from TOMS (Total Ozone Mapping Spectrometer) at toms.gsfc.nasa.gov (smoke aerosol) and metosrv2.umd.edu/-tropo (tropospheric ozone). The striking increase in smoke and tropospheric ozone observed during the 1997 Indonesian fires was the first extreme episode observed. During the August-November period, plumes of excess ozone and smoke coincided at times but were decoupled at other times, a phenomenon followed with trajectories. Thus, trans-boundary evolution of smoke and ozone differed greatly. The second discovery of the 1997 TOMS record was a dynamical interaction of ozone with the strong El Nino Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) that led to a jump in tropospheric ozone in March 1997 over the entire Indian Ocean, well ahead of the intense burning period. A climatology of smoke and tropospheric ozone from a 1980's TOMS instrument shows offsets in the timing of these pollutants - further evidence that factors other than biomass burning exert a strong influence on tropical tropospheric ozone.

Thompson, A. M.↗

NASA NTRS2001

Evidence of Stratosphere-to-Troposphere Transport Within a Mesoscale Model and TOMS Total Ozone

We present evidence for stratospheric mass transport into, and remaining in, the troposphere in an intense midlatitude cyclone. Mesoscale forecast model analysis fields from the Mesoscale Analysis and Prediction System (MAPS) were compared with total ozone observations from the Total Ozone Measurement Spectrometer (TOMS). Coupled with parcel back-trajectory calculations, the analyses suggest two mechanisms contributed to the mass exchange: (1) A region of dynamical ly-induced exchange occurred on the cyclone's southern edge. Parcels originally in the stratosphere crossed the jet core and experienced dilution by turbulent mixing with tropospheric air. (2) Diabatic effects reduced parcel potential vorticity (PV) for trajectories traversing precipitation regions, resulting in a "PV-hole" signature in the cyclone center. Air with lower-stratospheric values of ozone and water vapor was left in the troposphere. The strength of the latter process may be atypical. These results, combined with other research, suggest that precipitation-induced diabatic effects can significantly modify, (either decreasing or increasing) parcel potential vorticity, depending on parcel trajectory configuration with respect to jet core and maximum heating regions. In addition, these results underscore the importance of using not only PV but also chemical constituents for diagnoses of stratosphere-troposphere exchange (STE).

Olsen, Mark A.↗

NASA NTRS2001

Microphysical Interpretation of Cirrus Measurements With Lidar

Cirrus measurements obtained with a ground-based polarization Raman lidar at 67.9 N in January 1997 reveal a strong correlation between the particle optical properties, specifically depolarization ratio delta and extinct ion-to-backscatter ratio S, for ambient cloud temperatures above approximately -45 C (delta less than approximately 40%), and an anti-correlation for colder temperatures (delta greater than approximately 40%). Over the length of the measurements (4-7.5 hours) the particle properties vary systematically: Initially, delta approximately equal to 60% and S approximately equal to 10sr are observed. Then, with decreasing delta, S first increases to approximately 27 sr(delta approximately equal to 40%) before decreasing to values around 10 sr again (delta approximately equal to 20%). The particle optical properties distinctly depend on the ambient temperature. For the microphysical analysis of the lidar observations. ray-tracing computations of particle scattering properties and a size-distribution resolving cirrus model with explicit microphysics have been used. The theoretical studies show that the optical properties and their temporal evolution can be interpreted in terms of size, shape, and growth of the cirrus particles: Near the cloud top in the early stage of the cirrus development, light scattering by small hexagonal columns with aspect ratios close to one is dominant. Over time the cloud base height extends to lower altitudes with warmer temperatures, the ice particles grow and get morphologically diverse (the scattering contributions of hexagonal columns and plates are roughly the same for large S and depolarization values of approximately 40%). In the lower ranges of the cirrus clouds, light scattering is predominantly by plate-like or complex ice particles. Mid-latitude cirrus data measured with the same instrument at 53.4 N between 1994 and 1996 follow closely the correlation between delta and S found in the warmer regions of the Arctic cirrus clouds (delta less than approximately 40%). Cirrus clouds with higher depolarization ratios are rarely observed, even for very cold ambient temperatures. Atmospheric parameters other than temperature, e.g., the availability of water vapor, are also important for the growth and the morphology of cirrus particles.

Reichardt, Jens↗

NASA NTRS2001

Are US Great Plains Droughts Predictable on Seasonal and Longer Time Scales?

The United States has experienced numerous episodes of unusually dry conditions lasting anywhere from months to several years. In this talk, I will examine the predictability of such episodes and the physical mechanisms controlling the variability of the summer climate of the continental United States. The analysis is based on ensembles of multi-year simulations and seasonal hindcasts generated with the NASA Seasonal-to-Interannual Prediction Project (NSIPP-1) General Circulation Model.

Schubert, Siegfried↗

NASA NTRS2001

Methods for Discerning Cloud Reflectivity Changes due to the Indirect Effect of Aerosol: A Pilot-study for Triana

Understanding the effect of aerosol on cloud systems is one of the major challenges in atmospheric and climate research. Local studies suggest a multitude of influences on cloud properties. Yet the overall effect on cloud albedo, a critical parameter in climate simulations, remains uncertain. NASA's Triana mission will provide, from its EPIC multi-spectral imager, simultaneous data on aerosol properties and cloud reflectivity. With Triana's unique position in space these data will be available not only globally but also over the entire daytime, well suited to accommodate the often short lifetimes of aerosol and investigations around diurnal cycles. This pilot study explores the ability to detect relationships between aerosol properties and cloud reflectivity with sophisticated statistical methods. Sample results using data from the EOS Terra platform to simulate Triana are presented.

Kinne, S.↗

NASA NTRS2001

The Effect of Deep, Tropical Convection on the Tropical Tropopause Layer

There is evidence that most convective detrainment occurs below 355-K potential temperature (14 km). On the other hand, it is also generally accepted that convection does occasionally reach the altitude of the tropopause, typically approx. 16 km, or higher. The question is whether the relatively infrequent penetrations above 14 km are sufficient to influence the composition of the atmosphere above 14 km. Using measurements of O3 and CO as tracers of convective activity, I quantify the height distribution of mass detraining from convection near the tropopause by means of a simple mixing model similar to those used to calculate mixing into the stratospheric "tropical pipe". I calculate that significant detrainment from convection is occurring as high as approx. 380 K (16 km), an altitude typical of the tropopause. Further, I calculate that as much as 40% of the mass crossing the 380-K surface at this time detrained above 370 K. I conclude that the "mixing barrier" at 14 km is more porous than previously thought.

Dessler, A. E.↗

NASA NTRS2001

Are Droughts in the United States Great Plains Predictable on Seasonal and Longer Time Scales?

The United States Great Plains has experienced numerous episodes of unusually dry conditions lasting anywhere from months to several years, In this presentation, we will examine the predictability of such episodes and the physical mechanisms controlling the variability of the summer climate of the continental United States. The analysis is based on ensembles of multi-year simulations and seasonal hindcasts generated with the NASA Seasonal to-Interannual Prediction Project (NSIPP-1) General Circulation Model.

Schubert, Siegfried D.↗

NASA NTRS2001

Results from Global Land-Surface Data Assimilation Methods

Realistic representation of the land surface is crucial in global climate modeling (GCM). Recently, the Mosaic land-surface Model (LSM) has been driven off-line using GEOS DAS (Goddard Earth Observing System Data Assimilation System) atmospheric forcing, forming the Off-line Land-surface Global Assimilation (OLGA) system. This system provides a computationally efficient test bed for land surface data assimilation. Here, we validate the OLGA simulation of surface processes and the assimilation of ISCCP surface temperatures. Another component of this study as the incorporation of the Physical-space Statistical Analysis System (PSAS) into OLGA, in order to assimilate surface temperature observations from the International Satellite Cloud Climatology Project (ISCCP). To counteract the subsequent forcing of the analyzed skin temperature back to the initial state following the analysis. incremental bias correction (IBC) was included in the assimilation. The IBC scheme effectively removed the time mean bias, but did not remove him in the mean diurnal cycle. Therefore, a diurnal him correction (DBC) scheme was developed, where the time-dependent bias was modeled with a sine wave parameterization. In addition, quality control of the ISCCP data and anisotropic temperature correction were implemented in PSAS. Preliminary results showed a substantial impact from the inclusion of PSAS and DBC that was visible in the surface meteorology fields and energy budget. Also, the monthly mean diurnal cycle from the experiment closely matched the diurnal cycle from the observations.

Radakovich, Jon D.↗

NASA NTRS2001

Wind and Rotation Enhanced Escape from the Early Terrestrial Atmospheres

The earliest atmospheres of the terrestrial planets are thought to have been hotter, have stronger winds and rotate faster than atmospheres of today. Since these primitive atmospheres were weakly bound, they evolved rapidly because atmospheric escape was very strong, often referred to as "blowoff." Such escape has been treated as hydrodynamic, transonic flow; similar to solar wind flow dynamics. However, in many cases, although the outward flow is hydrodynamic at low altitudes, it becomes collisionless at higher altitudes, before sonic speeds are ever attained. Recent models dealing with the transition from fluid to kinetic flow have applied the Jeans escape flux at the exobase. This approach leads to escape rates that are too low, because thermospheric winds and planetary rotation are known to increase the escape flux above the corresponding Jeans flux. In particular, for a given density and temperature at the exobase, the escape flux increases as the wind speed and/or the rotation rate increase. Also, for a given wind speed and rotation rate, the escape flux enhancement over the Jeans flux increases as the mass of an escaping constituent increases, an important factor in isotope fractionation, especially the enrichment of deuterium on Mars. Accounting for a range of possible temperatures, thermospheric wind speeds and planetary rotation rates in the primitive atmospheres of the terrestrial planets, estimates are made of light constituent escape flux increases over the corresponding Jeans fluxes.

Hartle, Richard E.↗

NASA NTRS2001

Variations and Trends in Global and Regional Precipitation Based on the 22-Year GPCP (Global Precipitation Climatology Project) and Three-Year TRMM (Tropical Rainfall Measuring Mission) Data Sets

This paper gives an overview of the analysis of global precipitation over the last few decades and the impact of the new TRMM precipitation observations. The 20+ year, monthly, globally complete precipitation analysis of the World Climate Research Program's (WCRP/GEWEX) Global Precipitation Climatology Project (GPCP) is used to study global and regional variations and trends and is compared to the much shorter TRMM(Tropical Rainfall Measuring Mission) tropical data set. The GPCP data set shows no significant trend in global precipitation over the twenty years, unlike the positive trend in global surface temperatures over the past century. The global trend analysis must be interpreted carefully, however, because the inhomogeneity of the data set makes detecting a small signal very difficult, especially over this relatively short period. The relation of global (and tropical) total precipitation and ENSO events is quantified with no significant signal when land and ocean are combined. Identifying regional trends in precipitation may be more practical. From 1979 to 2000 the tropics have pattern of regional rainfall trends that has an ENSO-like pattern with features of both the El Nino and La Nina. This feature is related to a possible trend in the frequency of ENSO events (either El Nino or La Nina) over the past 20 years. Monthly anomalies of precipitation are related to ENSO variations with clear signals extending into middle and high latitudes of both hemispheres. The El Nino and La Nina mean anomalies are near mirror images of each other and when combined produce an ENSO signal with significant spatial continuity over large distances. A number of the features are shown to extend into high latitudes. Positive anomalies extend in the Southern Hemisphere (S.H.) from the Pacific southeastward across Chile and Argentina into the south Atlantic Ocean. In the Northern Hemisphere (N.H.) the counterpart feature extends across the southern U.S. and Atlantic Ocean into Europe. Further to the west a negative anomaly extends southeastward again from the Maritime Continent across the South Pacific and through the Drake Passage. In the Southern Hemisphere an anomaly feature is shown to spiral into the Antarctica land mass. The extremes of ENSO-related anomalies are also examined and indicate that globally, during both El Nino and La Nina, more extremes of precipitation (both wet and dry) occur than during the "neutral" regime, with the El Nino regime showing larger magnitudes. The distribution is different for the globe as a whole and when the area is restricted to just land. The recent (1998-present) TRMM observations are compared with the similar period of GPCP analyses with very good agreement in terms of pattern and generally good agreement with regard to magnitude. However, there still are differences among the individual TRMM products using passive and active microwave techniques and these need to be resolved before longer-term products such as the GPCP analyses can be validated.

Adler, Robert F.↗

NASA NTRS2001

A Modeling Study of Oceanic Response to Daily and Monthly Surface Forcing

The goal of this study is to investigate the effect of high-frequency surface forcing (wind stresses and heat fluxes) on upper-ocean response. We use the reduced-gravity quasi-isopycnal ocean model by Schopf and Loughe (1995) for this study. Two experiments are performed: one with daily and the other with monthly surface forcing. The two experiments are referred to as DD and MM, respectively. The daily surface wind stress is produced from the SSM/I wind data (Atlas et al. 1991) using the drag coefficient of Large and Pond (1982). The surface latent and sensible heat fluxes are estimated using the atmospheric mixed layer model by Seager et al. (1995) with the time-varying air temperature and specific humidity from the NCEP-NCAR reanalysis (Kalnay et al. 1996). The radiation is based on climatological shortwave radiation from the Earth Radiation Budget Experiment (ERBE) [Harrison et al. 1993] and the daily GEWEX SRB data. The ocean model domain is restricted to the Pacific Ocean with realistic land boundaries. At the southern boundary the model temperature and salinity are relaxed to the Levitus (1994) climatology. The time-mean SST distribution from MM is close to the observed SST climatology while the mean SST field from DD is about 1.5 C cooler. To identify the responsible processes, we examined the mean heat budgets and the heat balance during the first year (when the difference developed) in the two experiments. The analysis reveals that this is contributed by two factors. One is the difference in latent heat flux. The other is the difference in mixing processes. To further evaluate the responsible processes, we repeated the DD experiment by reducing the based vertical diffusion from 1e-4 to 0.5e-5. The resultant SST field becomes quite closer to the observed SST field. SST variability from the two experiments is generally similar, but the equatorial SST differences between the two experiments show interannual variations. We are investigating the possible mechanisms responsible for the different responses.

Sui, Chung-Hsiung↗

NASA NTRS2001

Influence of 3D Radiative Effects on Satellite Retrievals of Cloud Properties

When cloud properties are retrieved from satellite observations, the calculations apply 1D theory to the 3D world: they only consider vertical structures and ignore horizontal cloud variability. This presentation discusses how big the resulting errors can be in the operational retrievals of cloud optical thickness. A new technique was developed to estimate the magnitude of potential errors by analyzing the spatial patterns of visible and infrared images. The proposed technique was used to set error bars for optical depths retrieved from new MODIS measurements. Initial results indicate that the 1 km resolution retrievals are subject to abundant uncertainties. Averaging over 50 by 50 km areas reduces the errors, but does not remove them completely; even in the relatively simple case of high sun (30 degree zenith angle), about a fifth of the examined areas had biases larger than ten percent. As expected, errors increase substantially for more oblique illumination.

Varnai, Tamas↗

NASA NTRS2001

A New "Cloud Mode" for the AERONET

A key to predicting climate change is to observe an understand the global distribution of clouds and their physical properties such as optical thickness and droplet size. Since clouds change rapidly over short time and space intervals, they are difficult to simulate in computer models. But it is essential that global climate models predict realistic spatial and temporal distribution of cloud optical depth. The best way to verify these distributions is to infer optical depth from global coverage satellite data. However, satellite methods have many sources of uncertainty; thus, independent and reliable ground-based estimates are essential for validation. For aerosol, there is the AERONET - a ground based monitoring network that consists of identical multi-channel radiometers for assessing aerosol optical properties and validating their satellite retrievals. In addition to AEROSOL, we want the network monitoring CLOUD optical properties. It will use AERONET "time" (inappropriate for aerosol studies) to make basic new measurements related to cloud physics. In the presentation we will report on a new technique that retrieves cloud optical thickness for even broken clouds above green vegetation from surface measurements of zenith radiance in the visible (VIS) and near-IR (NIR) spectral regions. The idea of the method is simple: since green vegetation reflects 40-50% of incoming radiation in the NIR and only 5-10% in the VIS region, ground measurements under thin clouds have little spectral contrast between VIS and NIR, while thick clouds reflect much more of the surface-reflected radiation in the NIR than in VIS. Based on this idea, we use a combination of measurements (spectral indices) in VIS and NIR to estimate cloud optical thickness. By analogy with NDVI, the simplest index that can be defined is the Normalized Difference Cloud Index (NDCI) which is a ratio between the difference and the sum of two radiances measured for two narrow spectral bands in VIS and NIR.

Marshak, Alexander↗

NASA NTRS2001

On the Origin of Polar Vortex Air

The existence of the multi-year HALOE CH4 data set, together with some comparisons of forward with back trajectory calculations which we have carried out, has motivated us to reexamine the question of polar vortex descent. Three-dimensional diabatic trajectory calculations have been carried out for the seven month fall to spring period in both the northern hemisphere (NH) and southern hemisphere (SH) polar stratosphere for the years 1992-1999. These computations are compared to fixed descent computations where the parcels were fixed at their latitude-longitude locations and allowed to descend without circulating. The forward trajectory computed descent is always less than the fixed descent due to horizontal parcel motions and variations in heating rates with latitude and longitude. Although the forward calculations estimate the maximum amount of descent that can occur, they do not necessarily indicate the actual origin of springtime vortex air. This is because more equator-ward air can be entrained within the vortex during its formation. To examine the origin of the springtime vortex air, the trajectory model was run backward for seven months from spring to fall. The back trajectories show a complex distribution of parcels in which one population originates in the upper stratosphere and mesosphere and experiences considerable descent in the polar regions, while the remaining parcels originate at lower altitudes of the middle and lower stratosphere and are mixed into the polar regions during vortex formation without experiencing as much vertical transport. The amount of descent experienced by the first population shows little variability from year to year, while the computed descent and mixing of the remaining parcels show considerable interannual variability due to the varying polar meteorology. Because of this complex parcel distribution it is not meaningful to speak of a net amount of descent experienced over the entire winter period. Since the back trajectories indicate that much of the air can come from lower altitudes than would be implied by the forward calculations, using a comparison between pre-winter and post-winter tracer profiles to estimate the amount of descent over this period will give erroneous descent amounts. In order to evaluate the computed descent, spring methane amounts were computed by mapping HALOE fall observations onto the final latitude-altitude locations of the back trajectories. These locations indicate the origin of the spring vortex air. The agreement between the computed means and the spring HALOE means is generally within 0.1-0.2 ppmv in the NH and 0.1-0.4 ppmv in the SH.

Rosenfield, J. E.↗