Abstracts
Back and forth nudging algorithm for data assimilation problems
by
Didier Auroux
University of Toulouse, France
Coauthors: Jacques Blum (University of Nice, France)
The standard nudging algorithm consists in adding to the state
equations of a dynamical system a feedback term, which is proportional
to the difference between the observation and its equivalent quantity
computed by the resolution of the state equations. The model appears
then as a weak constraint, and the nudging term forces the state
variables to fit as well as possible to the observations. The backward
nudging algorithm consists in solving the state equations of the
model backwards in time. A nudging term, with the opposite sign
compared to the standard nudging algorithm, is added to the state
equations, and the final obtained state is in fact the initial state
of the system.
The back and forth nudging algorithm (BFN) consists in solving first
the forward nudging equation and then the direct system backwards
in time with a feedback term. After resolution of this backward
equation, one obtains an estimate of the initial state of the system.
We repeat these forward and backward resolutions until convergence
of the algorithm. The theoretical convergence of this algorithm
can be obtained in the case of a linear system, or a very simple
nonlinear ODE (Lorenz equations), under some observation hypothesis.
From a numerical point of view, we compared this algorithm to the
4D-VAR on nonlinear systems such as Lorenz, Burgers' equations,
and a quasi-geostrophic ocean model. In all cases, the first iterations
of the BFN algorithm were more efficient than the 4DVAR ones. Moreover,
it is still the case in presence of model or observation errors.
Finally, its implementation is very easy because it requires neither
the linearization of the equations nor any minimization process.
--------------------------------------------------------------------------------
Surface measurements of size and composition
of particulate matter at Eureka, Nunavut
by
Asan Bacak
Department of chemistry university of waterloo waterloo ontario
Coauthors: Thomas Kuhn, James Sloan
Significant levels of synthetic organic pollutants (so-called persistent
organic pollutants, or POPs) such as pesticides, PCBs and semi volatile
industrial chemicals have been found in the Arctic snow pack and
wildlife, providing unequivocal evidence for the long range transport
of these materials into the Arctic. Both the origins of these materials
and the mechanisms responsible for their transport have been investigated
for many years.
It is generally accepted that gas phase transport occurs via the
“grasshopper” mechanism, which is an annual cycle in which
the materials are sequentially volatilized and dispersed in the
summer and deposited back to the surface in the winter. The lower
average temperatures at high latitudes cause the retention of the
materials that are deposited there, with the result that they become
more concentrated with time in the high Arctic and bio-accumulate
in the ecosystem to levels at which they become a hazard to the
health of the species at the top of the food chain. Using the methods
of regional chemical transport modelling, we have shown previously
that at mid-latitudes a significant fraction of the transport of
these semi-volatile organics results from their partitioning to
atmospheric particulate matter (PM). In the size range below about
1 µm, intercontinental transport of PM is possible, providing
an additional mechanism for the transport of these organics to the
Arctic. To determine whether this mechanism is responsible for some
of the POPs detected in the Arctic, we have installed an Aerosol
Mass Spectrometer to measure size distribution and composition of
the PM arriving at the Polar Environmental Atmospheric Research
Laboratory (PEARL), at 80N;86W near Eureka, Nunavut. We will report
the results of our initial measurements, which were made during
2007. When combined with semi-Lagrangian trajectory modelling, these
results will identify the most probable sources of the contaminated
PM.
--------------------------------------------------------------------------------
The impact of ground based ozone monitoring on stratospheric
ozone assessments: A case study using sequential and variational
data assimilation
by
Frank Baier
German Aerospace Center, Oberpfaffenhofen, 82234 Wessling
Coauthors: Th. Erbertseder, H. Elbern, J. Schwinger and M. Bittner
Data assimilation using chemistry-transport models is commonly
used to derive consistent analysis of ozone related stratospheric
chemistry. However, based on historic satellite data it is also
possible to simulate the potential information gain by ground based
ozone soundings. Today, ground based observations are indispensable
to validate satellite instruments. Their potential role for routine
stratospheric data assimilation is therefore of special relevance
for planning future satellite missions. Here, we present results
of an Observing System Experiment (OSE) using ERS-2/GOME observations
to simulate different ozone sonde station networks with respect
to existing networks like that of GAW and NDAAC. The sequential
assimilation system ROSE/DLR is applied to study the long-term impact
of routine ground based observations on assimilated ozone fields.
To highlight the respective change in information content, analysis
and first-guess errors are discussed. In parallel the 4DVAR SACADA
system is used to analyse the sensitivity of ozone related chemistry
to ozone sonde observations for certain atmospheric conditions,
e.g., ozone hole versus dynamically distorted episodes. Our results
show that continuous ground based ozone soundings can lead to significantly
improved ozone maps via data assimilation. While the Northern Hemisphere
has an already well established station network, the greatest impact
is expected by increasing the number of observations in the Southern
Hemisphere and especially the Tropics.
--------------------------------------------------------------------------------
How will a changing stratosphere affect
high-latitude climate?
by
Mark P. Baldwin
NorthWest Research Associates, Bellevue, WA, USA
Recent research has shown that changes to the circulation of the
lower stratosphere affect tropospheric weather and climate, especially
at high latitudes. For example, the ozone hole in the Southern Hemisphere
spring has not only cooled the polar lower stratosphere and increased
the strength of the stratospheric circumpolar winds, but also affected
Antarctic surface climate.
During the past 25+ years the composition of the stratosphere has
changed significantly, with higher abundances of anthropogenic greenhouse
gases and ozone-depleting substances (ODSs), together with a concomitant
thinning of the ozone layer. With the recent stabilization of stratospheric
ODSs following the Montreal Protocol we are near the turnaround
point in ozone depletion, so that the past climate impacts of ODSs
and ozone depletion are about to change sign against a background
of continued increases in most greenhouse gases, which will tend
to cool the stratosphere. Chemistry-climate models predict that
ozone recovery will not be a simple reversal of ozone depletion.
Rather, the stratospheric cooling from increasing greenhouse gases
will, overall, accelerate the recovery of the ozone layer, so that
pre-1980 ozone abundances are expected to be reached sometime around
the middle of this century.
Nearly all climate models with well-represented stratospheres predict
an enhancement of the Brewer-Dobson circulation under climate change,
suggesting that the northern polar stratosphere will be warmer during
winter and spring, even while the rest of the stratosphere becomes
colder. Because of the very real possibility of dynamical responses
to the stratospheric changes, predictions of the evolution of polar
climate could be substantially different, especially in the winter
and spring in the Northern Hemisphere and spring and summer in the
Southern Hemisphere.
--------------------------------------------------------------------------------
Measurements of atmospheric trace
gases in the Arctic: First light measurements from the new FTIR
spectrometer at PEARL
by
Rebecca Batchelor
Department of Physics, University of Toronto
Coauthors: Rodica Lindenmaier, Kimberly Strong
In order to fully understand the mechanisms and processes which
result in ozone depletion and climate change, quality measurements
of atmospheric trace gases from high latitude observatories are
essential. The atmospheric observatory at Eureka (80N, 86W) has
recently been rejuvenated by the Canadian Network for the Detection
of Atmospheric Change (CANDAC).
A new Bruker IFS 125HR Fourier transform infrared spectrometer
was installed at the Polar Environment Atmospheric Research Laboratory
(PEARL), 610 m above sea level, in July 2006. With a resolution
of 0.0035 cm-1 and the capability of making automated measurements
of approximately 15 different trace gases in the mid-infrared region,
this instrument promises to be an essential component of the Arctic
observing network.
This presentation will introduce the instrument and present preliminary
results from the 2006 first-light and 2007 polar sunrise campaigns.
--------------------------------------------------------------------------------
Detection of Antarctic Ice Polar Stratospheric
Clouds from AIRS Assimilation
by
Craig Benson
UMBC
Coauthors: Ivanka Stajner, Hui-Chun Liu, Steven Pawson
The distribution and nature of Antarctic polar stratospheric clouds
(PSCs) is determined by synoptic, planetary, and gravity waves,
which vary on interannual, seasonal, monthly, daily, and hourly
time scales. This paper uses AIRS radiance observations with the
GEOS-5 data assimilation system to construct maps of ice PSCs in
the Antarctic. The basis of the work is that the AIRS channels near
6.79 um, primarily used to detect the tropospheric moisture distribution,
are sensitive to the presence of cirrus and PSCs in the atmosphere
above their level of peak sensitivity under clear-sky conditions.
This study uses differences between AIRS radiance observations and
clear-sky radiance computations based on the GEOS-5 background fields
to infer the presence of ice PSCs in the atmosphere, based on observed-minus-forecast
brightness-temperature differences of -2 K. This sensitivity threshold
was determined using meteorological fields from GEOS-5 to construct
Lagrangian trajectories along which the Integrated Microphysics
and Aerosol Chemistry on Trajectories (IMPACT) model is applied
to simulate PSC microphysics, heterogeneous and gas-phase chemistry;
PSC particle densities were used in the Moderate Resolution Atmospheric
Transmission (MODTRAN) radiation code to establish sensitivities
to the presence of PSCs. POAM observations of ice PSCs show a high
degree of correlation with AIRS O-Fs below -2 K. Further comparisons
with PSC data from CALIPSO and POAM provide a basis for validation
and interpretation of the AIRS PSC maps.
--------------------------------------------------------------------------------
Seasonal variations in the MLT tides in
the 120E meridian
by
Zeyu Chen
Institute of Atmospheric Physics, Chinese Academy of Sciences
Coauthors: Lu, Daren (Institute of Atmospheric Physics, Chinese
Academy of Sciences)
The authors report in this presentation the seasonal variations
in major atmospheric tides in the mesosphere and lower thermosphere
region in the meridian at 120 ?E. The SABER/TIMED temperature measurements
covering Nov. 2004 to Jan. 2006 were used to extract tidal components
with Fourier-leastsquares fit and FFT analysis, and to reconstruct
the diurnal, semidiurnal and terdiurnal tides in the meridian. The
migrating diurnal tides increase with altitude and attain maxima
at 97 km height, then decrease sharply with altitude. The tides
of other frequency increase with altitude and attain significant
amplitude at 97 km height. Considering the behavior in the vertical
of the amplitudes, the authors placed their focus on introducing
estimation results obtained at the 97-km altitude. For the tides
of each frequency, the migrating and nonmigrating components were
examined separately, and their respective contribution to the whole
of the tides were evaluated. The primary results showed that migrating
component plays dominant role in characterizing the general temporal
and spatial distribution for both diurnal and semidiurnal tides.
Regarding the diurnal tides, contribution of migrating component
is the most dominant one during spring equinox, which is characterized
by the amplitude maxima at the equator and that at the tropics for
both hemispheres. Moreover, the temporal variation of the diurnal
tides at the tropical latitude in Northern hemisphere is consistent
to the analysis result obtained by using meteor radar wind measurements
taken in Wuhan (30 ?N, 114 ?E). Contributions of nonmigrating tides
are more significant in other seasons. During summer solstitial
time in 2005, tidal modes (1, 0), (1, 2), (1, -3) and (1, -2) contribute
together to form a diurnal tides active area from 10 ?N to 30 ?S
with the maximal amplitude 20K at the Equator. Due to the domination
of migrating component, semidiurnal tides occur at the tropical
latitudes in both hemispheres. In northern hemisphere, the active
tides area centers at autumn equinox with maximum 13K. And in southern
hemisphere, the active area centers at the time in between spring
and summer equinox. The influences of nonmigrating semidiurnal components
are also clear during other seasons as several other centers with
maximal amplitude are seen. Confined in the latitude range 40 ?S-40
?N, terdiurnal components exhibit much weaker activity with much
smaller amplitudes than that of the diurnal and semidiurnal components.
Current estimation results suggest that nonmigrating terdiurnal
components have amplitudes that are as large as and during most
times larger than that of migrating component, thus predominate
the global distribution of terdiurnal tides in 2005.
--------------------------------------------------------------------------------
Temperature correlation between the lower
atmosphere and the MLT region
by
Young-Min Cho
York University
Coauthors: Young-Min Cho, Marianna Shepherd, Gordon Shepherd
The airglow temperature in the Mesosphere and Lower Thermosphere
(MLT) region has been measured using a Spectral Airglow Temperature
Imager (SATI) at Resolute Bay (74.68 N, 94.90 W) since November
2001. The MLT temperature anomalies are compared to the lower stratospheric
radiosonde temperature anomalies at Resolute Bay for five years.
A positive relationship of the MLT temperature at the altitude 87
km and the lower stratospheric temperature at 22.5 km is found during
the period. The MLT temperatures are also compared to the upper
stratospheric temperatures of SABER. A negative relationship between
the MLT temperature and the upper stratospheric temperature at the
altitude 55 km is found during the period. The lower atmospheric
temperature and the MLT temperature are also compared with the solar
flux variation as a function of QBO phase and season.
--------------------------------------------------------------------------------
Evolution of the Arctic polar vortex
during 2004/05 and 2005/06 winter seasons based on the analysis
of MetO assimilated fields
by
Tatyana Chshyolkova
ISAS, University of Saskatchewan
Coauthors: A. Manson, C. Meek
The atmospheric polar vortex is a dominant feature of winter middle
atmosphere. Knowledge of Arctic polar vortex structure and behavior
can provide information on the background atmosphere for chemical
and dynamical studies at the PEARL station. To characterize polar
vortices the MetO (UK Met Office) assimilated fields have been subjected
to the Q-diagnostic. As a part of the diagnostic, potential vorticity,
(PV), stream function, relative vorticity and the rate of strain
and rotation in wind field (Q) have been calculated at several isentropic
surfaces ( 20-50 km). Evolution of the Arctic polar vortex is demonstrated
for two winter seasons: 2004/05, with only a few relatively weak
stratospheric disturbances, and 2005/06, with a major sudden stratospheric
warming at the end of January. In addition data from 12 meteor and
MF radars have been used to compare dynamical processes at mesospheric
heights with the polar vortex structure. It is shown that the arrangement
of radar wind vectors is consistent with cyclonic motion around
the pole and they matched the MetO winds well at corresponding locations
during “quiet” days; while on occasions during the stratospheric
disturbances radar and MetO winds demonstrated opposite directions.
--------------------------------------------------------------------------------
Pan-Arctic Study of the Coupled Tropospheric
Stratospheric and Mesospheric Circulation
by
Richard L. Collins
Geophysical Institute, University of Alaska Fairbanks, 903 Koyukuk
Drive, Fairbanks, AK 99775
Coauthors: Brentha Thurairajah, David A. Atkinson, V. Lynn Harvey,
Kohei Mizutani, Ruth S. Lieberman
We present an IPY observational study that combines lower-resolution
global data from satellite observations and meteorological analyses
(e.g., National Centers for Environmental Prediction (NCEP), National
Center for Atmospheric Research (NCAR), UK Meteorological Office
(MetO)) with higher-resolution data from an Arctic observing network
of Rayleigh lidar systems to address questions in the aeronomy of
the stratosphere and mesosphere and the linkages to the troposphere.
We will employ an observing network of Rayleigh lidars at the Arctic
Lidar Observatory for Middle Atmosphere Research, Andoya, Norway
(69 N, 16 E), Eureka Stratospheric Observatory, Eureka, Nunavut,
Canada (80°N, 86°W), Poker Flat Research Range, Chatanika,
Alaska, USA (65 N, 147 W), Sondrestrom Upper Atmospheric Research
Facility, Kangerlussuaq, Greenland (67 N, 51 W) to provide the high-resolution
temperature measurements (~ 500 m, 10’s of minutes) of the
stratosphere and mesosphere (~40-80 km). We illustrate this IPY
study with observations from January 2003 that shows the evolution
of the stratospheric vortex and anti-cyclones during a period of
extensive interaction.
--------------------------------------------------------------------------------
General Characteristics of Stratospheric
Singular Vectors
by
Ronald M. Errico
GEST/UMBC and GMAO/NASA
Coauthors: Ronald Gelaro, Elena Novakovskaia, and Ricardo Todling
Leading singular vectors have been computed for a numerical weather
prediction model that can resolve dynamical structures within the
stratosphere and lower mesosphere. The norm applied at the final
time is the commonly used energy norm but confined to measuring
the stratosphere. These stratospheric singular vectors are described
by presenting three examples. They are produced using either of
two initial norms that weight perturbations within the troposphere
versus stratosphere very differently. For either initial norm, singular
values are typically smaller than their tropospheric counterparts
and they are less geographically local. They also retain their relevance
to corresponding nonlinear evolutions for longer periods and larger
amplitudes. For these reasons, stratospheric SVs may be useful for
explaining observed stratospheric dynamical behaviors.
--------------------------------------------------------------------------------
Ground-based Zenith-sky DOAS Measurements
of Trace Gases at Eureka, Nunavut
by
Annemarie Fraser
University of Toronto
Coauthors: Cristen Adams, Peter F. Bernath, Chris Boone, Florence
Goutail, Tobias E. Kerzenmacher, C. Thomas McElroy, Clive Midwinter,
Kimberly Strong, Jennifer Walker, Kaley A. Walker, Hongjiang Wu
The University of Toronto Ground-Based Spectrometer (UT-GBS) is
a portable zenith-sky-viewing UV-Visible spectrometer, assembled
in 1998. Since then it has participated in eight polar sunrise field
campiagns at the Polar Environmental Research Laboratory (PEARL)
in Eureka, Nunavut (80N, 86W, Feb. - Apr. 1999-2001, 2003-2007).
In August 2006, a second instrument (the PEARL UT-GBS) was permanently
installed at PEARL as part of the refurbishment of the lab by the
Canadian Network for the Detection of Atmospheric Change (CANDAC).
Vertical column density amounts of ozone and NO2 are regularly retrieved,
while slant column densities of BrO and OClO are retrieved when
possible.
We will discuss measurements from the 2004 – 2007 Canadian
Arctic ACE Validation Campaigns, which were held as part of the
validation effort for the ACE (Atmospheric Chemistry Experiment)
satellite. Also participating in these campaigns were three other
UV-Visible zenith-sky viewing spectrometers.
The addition of the second UT-GBS instrument with a suntracker
allows for direct Sun measurements to be taken at the same time
as zenith-sky measurements. Combining the slant columns from both
viewing geometries allows for the separation of BrO into tropospheric
and stratospheric partial columns. Direct Sun measurements were
made for the first time during the 2007 campaign. Plans for future
measurements with both UT-GBS instruments will be discussed.
--------------------------------------------------------------------------------
On the extraction of wind information
from the 4D-var assimilation of chemical constituents
by
Jean de Grandpré
Meteorological Service of Canada
Coauthors: Pierre Gauthier, Cécilien Charette et Monique
Tanguay
In partnership with the European Space Agency (ESA) and the Belgium
Institute for Space Aeronomy (BIRA-IASB), Environment Canada has
developed an assimilation system for addressing chemical weather
issues. It is based on the stratospheric extension of the operational
weather prediction global model with a lid at 0.1 hPa. The model
is coupled with a comprehensive on-line photochemical module to
incorporate dynamical, radiative and photochemical interactions.
In the stratosphere, TOVS-AMSU-a radiances and radiosondes observations
are assimilated whereas MIPAS occultation measurements are used
for the assimilation of temperature and chemical constituents such
as ozone, methane and nitric acid. In the case of 4D-Var cycles,
the system uses long lived constituents from MIPAS as passive tracers
in the tangent-linear and adjoint models for inferring wind increments
in the lower stratosphere where wind observations are sparse. Analysis
of the results will focus on the limitation of using a simplified
chemistry in the incremental 4D-Var.
--------------------------------------------------------------------------------
The mesospheric polar vortices in GEOS,
WACCM, SABER, and EOS-MLS
by
V. Lynn Harvey
University of Colorado/LASP
Coauthors: C. Randall, S. Pawson, R. Garcia, R. Lieberman, G. Manney
Satellite data analysis is combined with global modeling to characterize
the 3-D structure and day-to-day variability of the polar vortex
in the mesosphere. We use satellite temperature and geopotential
height data from the Sounding of the Atmosphere using Broadband
Emission Radiometry (SABER) instrument and temperature, carbon monoxide,
and methane data from the Microwave Limb Sounder (MLS) instrument
to characterize the structure of the upper stratospheric and mesospheric
polar vortex in each hemisphere on a daily basis. The mesospheric
vortex, as seen by these satellite instruments, is then compared
to the representation of the mesospheric vortex in the GEOS-4 and
GEOS-5 data assimilation systems as well as in the Whole Atmosphere
Community Climate Model (WACCM). We will show the 3-D structure
of the mesospheric vortex at times when the stratospheric vortex
is strong as well as how this structure is modified during stratospheric
warming events. This work will conduct model/observation intercomparisons
of the mesospheric vortex as a means to determine the need for data
assimilation in the mesosphere and above.
--------------------------------------------------------------------------------
The benefits of in-line advection -
Assessing the transport characteristics of the CMAM-DAS
by
Michaela I. Hegglin
University of Toronto
Coauthors: Stephen Beagley, Andreas Jonsson, Diane Pendlebury, Saroja
Polavarapu, and Theodore G. Shepherd
DAS-driven chemical transport models (CTMs) are known to be excessively
dispersive and to produce erroneous distributions of long-lived
tracers. This is reflected in too-young age of air, and results
from noise in the assimilated winds. While CTMs are driven off-line
by analyzed wind fields sampled at a certain frequency, CMAM-DAS
calculates advection within the model code (in-line), hence providing
higher temporal resolution. This procedure should mitigate the effects
of noise and make 3D-Var analyses useable for advection of chemical
species. In this study we test this hypothesis by investigating
the transport characteristics of the CMAM-DAS. Comparisons with
ACE-FTS satellite data and ER-2 aircraft measurements show that
mixing barriers are well represented and latitudinal gradients in
N2O, NOy, and O3 are retained in the CMAM-DAS, confirming our expectations.
We conclude that in-line calculation of advection represents a way
ahead in order to improve tracer transport in DAS.
--------------------------------------------------------------------------------
Stratospheric and Mesospheric Assimilation
using the NOGAPS-ALPHA/NAVDAS forecast model
by
Karl Hoppel
Naval Research Laboratory
Coauthors: David Siskind, Larry Coy, Stephen Eckermann, Andrew Kochenash,
John McCormack, Nancy Baker
A high altitude extension of the Navy Operational Global Atmospheric
Prediction System (NOGAPS) Advanced Level Physics, High Altitude
(ALPHA) is under development at the US Naval Research Lab. This
research model has extended the altitude range of NOGAPS from the
lower stratosphere into the mesosphere, and has been used to study
stratospheric warmings [Siskind et al, 2005 JGR] and mesospheric
dynamics [Siskind et al, 2007 GRL] Recently, the NRL Atmospheric
Variational Analysis System (NAVDAS), has also been extended to
begin assimilating measurements over the full pressure range of
NOGAPS-ALPHA. Initial tests are being performed using temperature
measurements from the NASA's Microwave Limb Sounder (MLS) and the
Sounding of the Atmosphere using Broadband Emission Radiometry (SABER)
instruments, along with univariate MLS ozone and water vapor analyses.
The temperature assimilation is being used to tune the forecast
model and study the impact of improved stratospheric initial conditions
on stratospheric forecasts. This presentation describes the assimilation
system and examines the impact of the assimilation on understanding
and forecasting the stratospheric major warming of Jan/Feb 2006.
--------------------------------------------------------------------------------
Impact of Different Representations of
Ozone on Tropospheric Weather Forecasts
by
Mike Keil
Met Office, UK
Coauthors: David Jackson Camilla Mathison
Ozone can potentially have a large impact on Numerical Weather
Prediction (NWP) products. For example, improved representations
of ozone can lead to better temperature analyses and forecasts (via
more accurate radiative heating rates) and better assimilation of
satellite radiances. Improved ozone analyses can also lead to improved
surface UV forecasts. Until recently, the availability and quality
of ozone observations, and the understanding of the performance
of the ozone assimilation scheme, have not been sufficient for these
advances to have been realised. However, with the introduction of
new ozone observations from satellites and the growing maturity
of ozone assimilation techniques it is now appropriate to revisit
these issues.
The first part of this presentation describes how the addition
of EOS Microwave Limb Sounder (EOS MLS) ozone observations to the
Met Office ozone assimilation scheme improves the quality of the
ozone analysis. The chief positive benefits are seen in the lower
stratosphere, with reductions of mean errors in both the extratropics
and tropics. In particular, the representation of winter stratospheric
ozone depletion, low ozone near the summer pole, and ozone near
the tropical tropopause is considerably improved with the addition
of the EOS MLS data.
In the second part of the presentation a series of experiments
is run to examine how changing the ozone fields used in the forecast
model radiation scheme can impact on the accuracy of tropospheric
forecasts. Currently, the operational Met Office NWP system uses
the ozone climatology of Li and Shine (1995) for this purpose. A
series of experiments was run with several alternative representations
of ozone, including: assimilated ozone from two analyses (one with
EOS MLS data included in the assimilation and one with them excluded),
ECMWF ozone analyses imported into the system and an updated ozone
climatology from SPARC.
The NWP index is a measure of the skill of the tropospheric forecasts.
Our results show a positive impact on the NWP Index when the Li
and Shine ozone climatology is replaced by the SPARC climatology
and by both the assimilated ozone fields. However, the use of the
ozone ECMWF fields caused a deterioration in the NWP index. Longer-period
forecasts also hint at potential benefits of improved representations
of ozone to extended range tropospheric forecasts.
--------------------------------------------------------------------------------
Equatorial waves as a balance relationship
in global data assimilation
by
Heiner Körnich
Department of Meteorology, Stockholm University
Coauthors: Erland Källén
Tropical analysis displays the largest uncertainties in global
data assimilation. Besides the lack of observations, the assimilation
process would also benefit from a tropical balance relationship
incorporated in the background error. It has been suggested to use
equatorial waves as a mass/wind balance relationship. This approach
is tested in the idealized setup of a global shallow water model.
It is demonstrated that the application of an incomplete relationship
based on geostrophy alone leads to a misinterpretation of observational
data and thus to enhanced errors in the analysis. Only the combination
of geostrophy and equatorial waves improves the tropical analysis.
Furthermore, its applicability to a full 3D-GCM extending into the
middle atmosphere is explored. As a first step the background error
covariances are calculated in terms of the equatorial waves. The
horizontal structure of the background error is then interpreted
in terms of the convectively excited equatorial waves. The vertical
correlations provide further information about the vertical propagation
of the waves. Finally, the performance of the equatorial wave approach
is compared with alternative balance relationships, such as semi-geostrophy.
--------------------------------------------------------------------------------
Stratospheric Circulation—AO—Mei-yu
Anomaly
by
Chongyin Li
Institute of Atmospheric Physics, CAS, P.O.Box 9804, Beijing 100029
Coauthors: Wei Gu and Jing Pan
The relationship between stratospheric circulation variation and
climate is an important component of the SPARC/WCRP. The Mei-yu
is an outstanding climate event in East Asia and has important impacts
on the economic and societal development in China and Japan. So
it has been paid more attentions how about the relation of Mei-yu
anomaly to stratospheric circulation variation. Based on the data
analyses, this study will show that the Mei-yu in summer is evident
correlation with the circulation anomaly at the stratosphere in
winter. And the influence process of the tratospheric circulation
on the Mei-yu is also shown: the stratospheric circulation anomaly
will cause AO (Arctic Oscillation) variation at first; then different
pattern of the AO will lead to the tropospheric circulation anomaly
in East Asia in summer, which is favorable to enhance or reduce
the Mei-yu rainfall.
--------------------------------------------------------------------------------
Intercomparison and fusion of EOS/MLS
and TIMED/SABER temperatures
by
Ruth S. Lieberman
Northwest Research Associates
Coauthors: Dennis Riggin
Data from the Earth Observing System (EOS) and the Thermosphere-Ionosphere-Mesosphere
Energetics and Dynamics (TIMED) satellites are combined in order
to maximize their space-time information content for studies of
waves in the mesosphere. A statistical comparison has been carried
out for temperatures from the EOS Microwave Limb Sounder (MLS) and
the TIMED Sounding of the Atmosphere using Broadband Emission Radiometry
(SABER) during spatiotemporal coincidences of the two instruments.
Relative to SABER EOS/MLS has a slight warm bias in the stratosphere,
and a more substantial cold bias at mesospheric altitudes. Both
offsets are removed by applying an empirical correction to MLS temperatures.
RMS differences range between 2K and 4K in the stratosphere, and
rise steeply in the mesosphere. The datasets have been merged in
order to analyze diurnal tides over sub-yaw length intervals.
--------------------------------------------------------------------------------
Assimilation of Multiple Ozone Products
into the NCEP Operational Forecast Model.
by
Craig Long
NOAA/National Weather Service/NCEP/Climate Prediction Center
Coauthors: Shuntai Zhou, Russ Treadon
Recently, there has been a vast increase in the number of total
and profile ozone data products available from either operational
or research satellites. With the current assimilation techniques
and schemes there is little distinction between data coming from
the operational instruments and instruments of opportunity. At NCEP
the operational ozone data assimilated has always and only been
from the SBUV/2. In preparation for its replacement (the Ozone Mapper
and Profiler Suite, OMPS) on the National Polar-orbiting Operational
Environmental Satellite System (NPOESS), NCEP has revised the way
ozone data is assimilated into the Gridpoint Statistical Interpolation
(GSI) scheme expanding its capabilities to assimilate additional
total and profile ozone data. We will discuss the benefits of assimilating
the new SBUV/2 version 8 data and the OMI total ozone data into
the operational GFS model not only in terms of ozone forecasting
but also how other fields, such as temperature and winds, are affected.
Additionally, we will present test results from assimilating the
MLS, HIRDLS, and OMI ozone profile data, which are not currently
available in ‘near real time’ or of sufficient quality
for operational assimilation. We will discuss the process and results
of validating the total and profile ozone analyses and forecasts.
Lastly, we will discuss our plans for assimilation of MetOp GOME-2
total ozone data into the GFS model. The GOME-2 total and profile
data are the next ozone data sources planned to be assimilated operationally
into the GFS.
--------------------------------------------------------------------------------
Global aerosol forecasting and data assimilation
in GFS/GSI: Overview and Progresses
by
Sarah Lu
NCEP EMC, 5200 Auth Rd, Camp Springs, MD
Coauthors: Ho-Chun Huang, NOAA/NCEP EMC Yu-Tai Hou, NOAA/NCEP EMC
Jeff Mcqueen, NOAA/NCEP EMC Mian Chin, NASA GSFC Arlindo da Silva,
NASA GSFC
The NOAA National Centers for Environmental Prediction (NCEP) Global
Forecast System/Global Data Analysis System (GFS/GDAS) is the decision
support tool used by NOAA for medium-range numerical weather prediction.
The forecast model, GFS, is a global spectral model with meteorology
data assimilation using the Gridpoint Statistical Interpolation
(GSI). The chemical and aerosol components that are important to
the radiation transfer computation and the description of long-range
transport impact on the US air quality are currently treated in
a primitive way in the GFS/GDAS. In specific, aerosols in GFS are
currently prescribed based on a climatology and GDAS assumes background
aerosol conditions. The NOAA/NWS/NCEP/EMC is developing an aerosol
forecasting and assimilation capability in GFS/GDAS to provide the
lateral boundary conditions to the CMAQ-WRF regional air quality
model as well as to improve the radiation feedback in the GFS simulations.
The aerosol modules are based on the NASA Goddard Chemistry Aerosol
Radiation and Transport (GOCART) model. Both offline and in-line
approaches are in progress. We will overview the progress of the
NCEP global aerosol modeling system and the approach on the chemical
data assimilation.
--------------------------------------------------------------------------------
Polar Stratopause and Tropopause Evolution
and Transport and Implications for Assimilated Analyses
by
Gloria Manney
Jet Propulsion Laboratory (also at New Mexico Tech)
Coauthors: Michael Schwartz, Kirstin Krueger, Saroja Polavarapu,
Shuzhan Ren, Karl Hoppel, Lawrence Coy, Steven Pawson, Kevin Strawbridge,
William Daffer, and the MLS, ACE and SABER Science Teams
Until very recently, global, multi-annual datasets covering the
upper stratosphere/lower mesosphere (USLM) and upper troposphere/lower
stratosphere (UTLS) were largely unavailable; detailed knowledge
of both regions is critical to understanding climate change and
ozone recovery. Some operational assimilated analyses, including
those from the European Center for Medium-Range Weather Forecasts
(ECMWF) and NASA's Global Modeling and Assimilation Office (GMAO),
are now provided at levels into the mesosphere; however, with no
direct data constraints and few data with which to compare them,
their quality is highly dependent on the underlying general circulation
models and assimilation methods and is largely unknown. The Aura
Microwave Limb Sounder (MLS), Atmospheric Chemistry Experiment-Fourier
Transform Spectrometer (ACE), and Sounding of the Atmosphere with
Broadband Emission Radiometry (SABER) instruments now provide unprecedented
global, multi-annual, multi-species datasets covering the upper
troposphere through the mesosphere, including high-quality temperature
data through the mesosphere. We use MLS, SABER, ACE-FTS, and ground-based
data, to detail the evolution of the stratopause during recent polar
winters, and to assess the ability of analyses to capture observed
behavior. In addition to ECMWF and GMAO analyses, we show preliminary
results from Canadian Middle Atmosphere Model (CMAM) assimilations
with a model top above 0.001 hPa and a comprehensive online chemistry-transport
module, and from preliminary experiments assimilating MLS and SABER
data in NRL's NOGAPS-ALPHA system. Relationships to vortex structure
in assimilated analyses, and consistency of that structure with
long-lived trace gas transport is also explored. While analyses
are better constrained by temperature data near the tropopause than
in the upper stratosphere and above, many uncertainties remain,
and comprehensive trace gas datasets have heretofore been sorely
lacking. We compare tropopause structure in MLS data and analyses,
and present examples aimed at evaluating the consistency of MLS
trace gas data with transport (both on- and off-line) by assimilated
winds. The examples shown here will focus on the polar winter, to
show analyses that will be done for IPY, and include stratopause/tropopause
evolution during stratospheric sudden warmings and examination of
interannual and interhemispheric variability.
--------------------------------------------------------------------------------
Near-Real-Time Processing Plans for
Aura MLS Data for Use in Data Assimilation
by
Gloria L. Manney (presenter, not first author)
Jet Propulsion Laboratory, California Institute of Technology
Coauthors: Nathaniel J. Livesey, Alyn Lambert, William G. Read,
Lucien Froidevaux, Michael Schwartz, Gloria L. Manney (presenting
author), David C. Cuddy, Vincent S. Perun, Paul A. Wagner
The Microwave Limb Sounder (MLS) aboard the Aura satellite, launched
July 15, 2004, has already obtained 3 years of daily global atmospheric
data. The dataset includes temperature, geopotential height, and
a number of constituents such as O3, H2O, CO, N2O, HNO3, HCl, ClO,
BrO, OH, HO2, with retrievals spanning the upper troposphere to
the mesosphere (or higher) in some cases. Close to 3500 profiles
are retrieved (per product) each day; ice water content (IWC) and
ice water path (IWP) information is also obtained. A significant
amount of these 3 years of data has been reprocessed using the version
2.2 algorithms (the 2nd public release of MLS data). Several groups
have already performed successful assimilation experiments using
MLS data, especially O3 and temperature; some of this work is reported
on elsewhere at this Workshop.
The MLS team is working on a fast but accurate production stream
for a select number of products (mainly T, O3, and H2O), in order
to enable near-real-time processing and data assimilation using
some of the MLS capabilities. We discuss the plans for creating
such a data stream from MLS and early test results using a preliminary
subset as an example.
--------------------------------------------------------------------------------
An overview of the dynamics of the
mesosphere and lower thermosphere
by
Charles McLandress
Dept of Physics, University of Toronto
The mesosphere and lower thermosphere (MLT) contain a rich spectrum
of atmospheric waves ranging from global-scale tides and planetary
waves to meso-scale gravity waves. A detailed picture of the MLT
is emerging as satellite observations of this region increase. In
tandem with this growth in global observations, have been advancements
in general circulation models which now extend upward to include
the MLT. The assimilation of observations in the MLT is therefore
an obvious step which would not only consistently blend models and
measurements but also enhance the usefulness of daytime-only measurements,
for example. However, data assimilation in the MLT is currently
not possible due to difficulties arising from the short time scales
in the MLT. This talk overviews the current understanding of the
dynamics of the MLT through a series of observational and modelling
results. Since a good model is a requirement for a good assimilation
system, focus is placed on the ability of models to simulate the
observed large-scale flow in the MLT. Benefits of assimilating MLT
measurements, such as daytime-only data, are also discussed.
--------------------------------------------------------------------------------
Coupled chemistry-dynamics data assimilation
by
Richard Menard
Atmospheric Science and Technology Directorate, Environment Canada
Coauthors: Simon Chabrillat3, Cécilien Charette2, Pierre
Gauthier2, Jean de Grandpré1, Alain Robichaud1, Yves Rochon1,
Yan Yang1, Alexander Kallaur1, Thomas von Clarmann4 and Jacek Kaminski5
1. Air Quality Research Division, Atmospheric Science and Technology
Directorate, Environment Canada 2. Meteorological Research Division,
Atmospheric Science and Technology Directorate, Environment Canada
3. Belgium Institute for Space Aeronomy, 3 Avenue Circulaire, 1180
Brussels, Belgium 4. Institut für Meteorologie und Klimaforschung,
Forschungzentrum Karlsruhe GmbH, Postfach3640, 76021 Karlsruhe,
Germany 5. Department of Earth and Space Science and Engineering,
York University, Toronto
In partnership with the Belgium Institute for Space Aeronomy (BIRA-IASB),
the Institute for Meteorology and Climatology (IMK) in Karlsruhe
Germany, York University in Toronto, Environment Canada has developed
a coupled chemistry-dynamics model and assimilation system for addressing
chemical weather issues. The model is based on the stratospheric
extension of the Canadian operational weather prediction Global
Multiscale Environmental (GEM) model with a lid at 0.1 hPa. The
model is coupled with a comprehensive on-line photochemical module
developed at BIRA for the Belgian Assimilation System for Chemical
Observations from Envisat (BASCOE), and incorporates chemical-radiative
interactions. The Canadian variational assimilation system has also
been extended to include chemical variables, and in particular cross
error covariances between ozone and the meteorological variables
and between long-lived species with balance operators. The incremental
4D Var extension to chemical variables uses the adjoint of tracer
transport. The study focuses on the stratosphere and the use of
Envisat observations. In addition to the standard meteorological
observations used in NWP operations, AMSU-a radiances of channels
11-14, MIPAS ESA and MIPAS IMK retrievals as well as GOMOS dark
limb measurements of temperature and chemical constituents (such
as ozone, water vapor, methane, nitrous oxide, nitrogen dioxide,
nitric acid) were assimilated. Several data assimilation cycles
throughout the period August-November 2003 have been performed for
the validation of the different components of the system. Both 3D-Var
and 4D-Var cycles have been performed to evaluate the impact of
the assimilation methods on the results. An overview of the study
will be presented. Some of the highlights are: An improved AMSU-a
bias correction, the impact of radiative feedback from ozone on
the model predictability, the impact of MIPAS ESA temperature on
transport of long-lived species, the assimilation of several chemical
species using 3D-Var, and the use of 4D-Var assimilation of long-lived
species to infer winds.
--------------------------------------------------------------------------------
Gravity Waves in Four-Dimensional Data
Assimilation
by
Lisa Neef
KNMI / University of Toronto
Coauthors: Theodore Shepherd, Saroja Polavarapu
In the mesosphere and upper stratosphere, modeled flows are dominated
by gravity waves. For data assimilation, it is difficult to formulate
covariance models for these regions, since traditional balance constraints
do not represent the full system. A related problem occurs in the
tropics, where there is no clear timescale separation between vortical
modes, gravity waves, and equatorial waves. Both regions have fewer
observations than the midlatitude troposphere, making accurate data
assimilation a greater necessity.
It is not clear to what extent 4D assimilation schemes are able
to develop accurate covariance fields in the context of an unbalanced
state and/or unclear timescale separation. We investigate this problem
using a simplified model of a chaotic vortical mode coupled to a
linear gravity wave of a given amplitude and frequency, and compare
the three most basic types of 4D data assimilation: the Extended
Kalman Filter (EKF), Ensemble Kalman Filter (EnKF), and 4D variational
assimilation (4D-Var). It is shown that each method's ability to
recover the vortical mode, when an gravity wave is present in the
true state but not observed, depends on the accuracy of modeled
covariances between fast and slow normal-mode variables. dLikewise,
the ability to extract both modes from observations which contain
both timescales of motion depends both on the estimated fast-slow
covariances, as well as the estimated error variance ascribed to
the gravity wave.
The EKF is shown to be very limited in the estimation of covariances
between fast and slow variables, and thus tends to return faulty
analyses, but it nevertheless remains useful as long as the fast
variables are observed and observations are very frequent. The EnKF
and 4D-Var offer two ways of alleviating the problems found in the
EKF, but the accuracy of each method depends several other complicating
factors.
--------------------------------------------------------------------------------
A study of the CMAM_DAS using simulated
observations
by
Yulia Nezlin
University of Toronto
Coauthors: Yves Rochon, Meteorological Service of Canada
The forecast errors in CMAM-DAS were investigated using simulations
with the known truth (taken from a free model run). Observations
were simulated at locations of actual measurements. The results
may be helpful in a few aspects. 1. This is an independent approach
to the estimation of error covariances when, contrary to other methods,
we know the truth. 2. This provides the ability to investigate the
sensitivity of the assimilation to its different components in isolation
[model, observations (variances, bias, type and distribution), assimilation
system components (3D-Var approach, the minimization method)]. 3.
This is a simple way to investigate the extent of predictability
in the mesosphere and the possibility of assimilating mesospheric
observations.
--------------------------------------------------------------------------------
Assimilation of Lagrangian Data in Oceanography
by
Maelle Nodet
University of Grenoble
Within the framework of Global Ocean Data Assimilation Experiment
(GODAE), an increasing amount of data are available. A crucial issue
for oceanographers is to exploit at best these observations, in
order to improve models, climatology, forecasts, etc.
Thanks to the international program Argo and to more localized
experiments (such as SAMBA, ARCANE-Eurofloat, ACCE), a new type
of data is now available: positions of floats drifting at depth
in the ocean. Unlike other data, mainly Eulerian, these ones are
Lagrangian: the measuring instrument move in the flow.
I will present methods and results about 4D-Var assimilation of
Lagrangian data in the OPAVAR ocean model: implementation, sensitivity
studies, assimilation of noisy observations, comparison with a classical
method, complementarity with temperature data.
--------------------------------------------------------------------------------
Estimating ozone radiative forcing
based on satellite observations
by
Mark Parrington
University of Toronto
Coauthors: Adam Diamant, Dylan Jones, Kevin Bowman, Helen Worden,
Larry Horowitz
Recently available satellite measurements, such as those from the
Tropospheric Emission Spectrometer (TES) on the NASA EOS Aura platform,
are providing for the first time, global datasets of the distribution
of trace gases in the lower atmosphere. One application of this
data is in constraining physical and chemical processes in numerical
models.
Presented here are initial results from a study of the impact of
constraining the global tropospheric ozone distribution, with ozone
profiles retrieved from TES, on the radiative forcing in a general
circulation model. Ozone data from TES are first assimilated into
a version of the Geophysical Fluid Dynamics Laboratory's chemistry-climate
model AM2-Chem in which the model dynamics are constrained by nudging
to NCEP reanalyses. The assimilated ozone field is then used to
constrain the radiation calculation in AM2-Chem, with no dynamical
constraints imposed, to assess the impact on the radiative forcing
through changes to the outgoing longwave radiation and model dynamics
(T, u, v, w etc).
--------------------------------------------------------------------------------
The constraint of data assimilation in
the stratosphere and troposphere on mesospheric motions
by
Shuzhan Ren
Department of Physics, University of Toronto
Coauthors: Saroja Polavarapu, Ted Shepherd
Due to the poor data coverage in the mesosphere, motions in the
mesosphere are largely unconstrained by observations in most data
assimilation systems. It is well known that the mesosphere is largely
controlled by vertically propagating waves from below. Since these
waves (gravity waves, planetary waves etc.), originating in the
troposphere and propagating through the stratosphere, can be better
represented below the mesosphere by a data assimilation system,
it is expected that the information of data assimilated below the
mesosphere can be carried into the mesosphere by the “corrected”
waves, and consequently can drive the mesosphere close to reality.
In the Canadian middle atmosphere data assimilation system (CMAM-DAS)
the forecast model (CMAM) has the model lid at 100km and the observations
assimilated in the 3dvar system are below the stratopause (1mb).
Therefore it is an ideal tool to examine the constraint on motions
in the mesosphere imposed by the data assimilation below.
We first launch two assimilation experiments starting from different
mesospheres. The differences in the mesosphere drop very quickly
after a few assimilation cycles and become very small after one
month indicating a strong constraint on the mesosphere from the
data assimilation below. To see if the constraint is realized mainly
via gravity waves, another assimilation experiment with different
initial mesospheres and non-orographic gravity wave drag turned
off is launched. Unlike the first two experiments mesospheric differences
in this experiment increase steadily with time (temperature difference
in the two poles and wind difference almost everywhere in the mesosphere).
This suggests that gravity waves are important agents through which
the data assimilated below the mesosphere is able to impose a constraint
on mesosphere.
--------------------------------------------------------------------------------
New 3D-Var Dynamical Constraints at
Environment Canada
by
Matt Reszka
Environment Canada
Coauthors: Saroja Polavarapu, Luc Fillion
New dynamical constraints are investigated in the context of the
3D-Var global data assimilation system used by Environment Canada.
Flow dependence in the mass-wind balance is introduced by replacing
statistical, time-averaged covariances with the Charney and hydrostatic
balances, linearized about the background state. The Charney balance
performs well in the extra-tropics and has some value in the tropics.
A new constraint is also imposed on the velocity potential by employing
the quasigeostrophic omega equation and continuity equation. This
is done in order to limit spurious mixing caused by the insertion
of observations which are not in balance with the background and
with each other. The GCM used in the assimilation system is a variant
of the operational model at Environment Canada (GEM), which has
a relatively high lid (0.1 hPa) and includes a comprehensive online
chemistry package (BIRA). The new constraints have been implemented
using the hybrid vertical coordinate for consistency with the model.
Resulting balanced increments are shown to compare favorably with
output from a free-running model. The utility of the new constraints
is examined in online 3D-Var experiments, comparing with the previous
statistical approach, and focusing in particular on the tropical
region. The impact on forecast scores is also discussed.
--------------------------------------------------------------------------------
Assimilation of MIPAS chemical constituents
during a major EPP-NOx event over Antarctic winter 2003
by
Alain Robichaud
ENVIRONNEMENT CANADA
Coauthors: Richard Ménard, Jean de Grandpré, Yves
Rochon and Yan Yang
The stratospheric chemical constituents retrieved from MIPAS (Michelson
Interferometer for Passive Atmospheric Sounder) instrument has been
assimilated using the Canadian 3D-VAR system in a new coupled dynamical-chemical
stratospheric model(GEM-BACH). The MIPAS data products onboard the
ENVISAT satellite, one of the most largest observation platform
ever launched to space, include temperature and various gas vertical
profiles from limb sounding in the mid-infrared part of the spectrum.
The period of study is austral winter 2003 where a considerable
flux of NOx was reported to descend to the stratosphere linked with
an EPP event (Energetic Particle precipitation). Such phenomenon
can modify substantially the NOx budget of upper stratosphere and
participate in catalytic processes controlling ozone of polar regions.
The goal of the study is to evaluate how the assimilation system
can handle the phenomenon given that there is no provision for the
model to simulate EPP (or similar geomagnetic events) and its indirect
effects in the stratosphere.
--------------------------------------------------------------------------------
3D-FGAT assimilation of MIPAS-IMK and
GOMOS chemical data
by
Yves J. Rochon
Environment Canada
Coauthors: Simon Chabrillat, Richard Ménard, Yan Yang, Alain
Robichaud, Cécilien Charette
In partnership with the Belgium Institute for Space Aeronomy, Environment
Canada has developed a coupled chemistry-dynamics model and assimilation
system for addressing chemical weather issues. This system has been
tested and developed using observations from MIPAS and GOMOS, two
limb-scanning instruments aboard ENVISAT. The main dataset used
in this study is MIPAS-ESA. Here we present the assimilation of
the stratospheric chemistry observations from two other datasets:
MIPAS-IMK (O3, NO2, HNO3, ClONO2) and GOMOS (O3, NO2). After a brief
description of the datasets and set-up of the assimilation experiments,
we will present the background and observation error statistics
obtained by the Hollingsworth-Lonnberg method from a first-pass
MIPAS-IMK assimilation. We will compare the MIPAS-IMK and GOMOS
analyses with the MIPAS-ESA analyses, to show that GOMOS observations
bring useful information to complement the MIPAS-ESA dataset, while
MIPAS-IMK observations are an interesting alternative for assimilation
of chemically active species.
--------------------------------------------------------------------------------
Stratospheric influences on surface
winter climate and prospects for seasonal forecasting
by
Adam Scaife
Hadley Centre, Met Office, UK.
Coauthors: Jeff Knight, Sarah Ineson and Andrew Marshall
The influence of stratospheric variability on surface winter climate
is investigated in modelling experiments, observational datasets
and seasonal hindcasts. Stratospheric changes appear to be important
for the very rapid warming of Europe in winter between the 1960s
and 1990s and associated changes in the frequency of climate extremes.
The winter of 2005/6 is used as a case study to illustrate how this
influence occurs in individual years. Finally, hindcasts with general
circulation models are used to assess potential improvements in
seasonal forecasts.
--------------------------------------------------------------------------------
Cross validation of MIPAS/SAGE II
and MIPAS/HALOE trace gas observations by means of four dimensional
variational assimilation
by
Jörg Schwinger
Rhenish Institute for Environmental Research at the University of
Cologne, Aachener Straße 201, 50931 Köln, Germany
Coauthors: Hendrik Elbern
One of the objectives of four dimensional variational data assimilation
is the possibility to cross validate retrieved profiles from different
space borne instruments. Usually this task is accomplished by comparing
an ensemble of collocated profiles, which coincide within some predefined
maximum distance and a maximum time separation. This technique is
common practise, but the selection of collocation criteria is somewhat
arbitrary and problematic: Either the statistical basis for cross
validation remains poor (strict selection) or the large distances
in space and time (lax selection) introduce unquantifiable errors.
The use of chemically consistent constituent fields obtained by
4D-var data assimilation offers a more satisfying approach to the
cross validation challenge. Without applying any artificial criteria,
a global analysis produced by processing observations of one instrument
can be compared to profiles retrieved from another sensor, which
have been withheld from the assimilation procedure.
We have assimilated MIPAS observations from several periods in
2002/2003 using the stratospheric chemistry assimilation system
SACADA. The resulting global fields of trace gas distributions are
used to compare them to SAGE II and HALOE retrievals. For this comparison
it is essential to have an error estimate for the 4D-var analyses.
As error margins are not delivered by the 4D-var system, we use
an estimate of the analysis error covariance matrix in observation
space. As a prerequisite, special care has been taken to specify
the observation and background error covariance matrices sufficiently
realistic. The results of conventional cross validation studies
are -where available- compared to the results of our 4D-var based
cross validation study.
--------------------------------------------------------------------------------
Investigating Middle Atmospheric Chemistry
at the Polar Environment Atmospheric Research Laboratory (PEARL)
by
Kimberly Strong
Department of Physics, University of Toronto
Coauthors: J. Drummond, H. Fast, A. Manson, T. McElroy, G. Shepherd,
R. Sica, J. Sloan, K. Strawbridge, K. Walker, W. Ward, J. Whiteway,
J. McConnell, P. Bernath, T. Shepherd, R. Batchelor, P. Fogal, A.
Fraser, D. Fu, A. Harrett, M. Harwood, T. Kerzenmacher, A. Khmel,
R. Lindenmaier, C. Midwinter, R. Mittermeier, P. Loewen, O. Mikhailov,
M.A. Okraszewski, H. Popova, K. Sung
The recently established Polar Environment Atmospheric Research
Laboratory (PEARL) is located in the Canadian high Arctic at Eureka,
Nunavut (80?N). It is being equipped with a suite of instrumentation
to investigate chemical and physical processes in the atmosphere
from the ground to 100 km. One of four research themes being pursued
at PEARL is that of Arctic Middle Atmosphere Chemistry, which is
focussed on the question of “What is the composition of the
Arctic atmosphere above the site and how is it changing with time?”
The overall goal of this theme is to improve our understanding of
the processes controlling the Arctic stratospheric ozone budget
and its future evolution, using measurements of the concentrations
of stratospheric constituents, in conjunction with dynamical, radiative,
aerosol/PSC, and meteorological observations also made at PEARL.
The complexity of the atmosphere and the different spectroscopic
signatures of its many chemical constituents make it impossible
to measure all relevant species using any one remote sounding technique.
Rather, these measurements will be made using the complementary
capabilities of several of the PEARL instruments, including an ozone
lidar, a Fourier transform infrared spectrometer, a UV-visible grating
spectrometer, and an Atmospheric Emitted Radiance Interferometer.
This presentation will provide an overview of the Arctic Middle
Atmosphere Chemistry theme, including its scientific motivation,
objectives, and planned measurements and science activities. Activities
in the first year will be discussed, along with some of the early
measurements.
--------------------------------------------------------------------------------
The Atmospheric Chemistry Experiment
(ACE): Mission Overview and Recent Results
by
Kaley A. Walker
Department of Physics, University of Toronto, 60 St. George Street,
Toronto, Ontario, Canada M5S 1A7
Coauthors: Chris Boone, Peter F. Bernath, C. Thomas McElroy, Sean
D. McLeod, and Ryan Hughes
The Atmospheric Chemistry Experiment (ACE), also known as SCISAT-1,
is a Canadian scientific satellite to perform remote sensing measurements
of the Earth's atmosphere. It was launched on August 12, 2003 and
has been operational for over 3.5 years. The primary instrument
on-board SCISAT-1 is a high-resolution (0.02 cm-1) Fourier Transform
Spectrometer (ACE-FTS) operating between 750 and 4400 cm-1. It also
contains two filtered imagers to measure atmospheric extinction
due to clouds and aerosols at 0.525 and 1.02 microns. The secondary
instrument is a dual UV-visible-NIR spectrophotometer called MAESTRO
(Measurements of Aerosol Extinction in the Stratosphere and Troposphere
Retrieved by Occultation) which extends the wavelength coverage
to the 280-1030 nm spectral region.
The primary measurement technique for both instruments is solar
occultation. From these measurements, altitude profiles of atmospheric
trace gas species, temperature and pressure are obtained. The 650
km altitude, 74 degree circular orbit provides global measurement
coverage with a focus on the Arctic and Antarctic regions. The primary
goal of the ACE mission is to measure and to understand the chemical
and dynamical processes that control the distribution of ozone in
the upper troposphere and stratosphere, with a particular emphasis
on the Arctic region. The mission status, current science results
and validation program will be reviewed in this paper.
--------------------------------------------------------------------------------
Tidal Signatures in the Extended Canadian
Middle Atmosphere Model
by
W.E. Ward
University of New Brunswick
Coauthors: J. Du, D. MacKenzie and D.Y. Wang
Migrating and non-migrating tidal components are among the dominant
dynamical features in the mesosphere and lower thermosphere in model
runs of the Extended Canadian Middle Atmosphere Model (CMAM). Although
the migrating diurnal tide is the dominant component, other components
are significant and in mid and high latitudes are as large or larger
than this component. Spectral analyses of the diurnal, semidiurnal
and terdiurnal signatures of the wind and temperature fields (sampled
every three hours) from one year of a multiple year run show eastward
and westward propagating components with wavenumbers as high as
5 to be present. Comparison of these results with satellite analyses
show the tidal field in the extended CMAM to be reasonably realistic.
Analysis of a separate run with a sampling interval of 5 minutes
indicates that tidal harmonics up to at least 4 /day contribute.
Appropriate treatment of these components will be one challenge
that data assimilation in the mesosphere will need to meet.
--------------------------------------------------------------------------------
Waves and Coupling Processes at the
Polar Environment Atmospheric Research Laboratory (PEARL)
by
William Ward
University of New Brunswick
Coauthors: Alan Manson, Young-Min Cho, Tatyana Chshyolkova, Dragan
Veselinovic, Ding Yi Wang, Tom Duck, Gordon Shepherd, Marianna Shepherd,
Robert J. Sica, Kimberly Strong, Jim Whiteway
Waves are the primary means through which various regions of the
atmosphere couple. At the Polar Environment Atmospheric Research
Laboratory several instruments are being installed to investigate
the nature of these coupling processes in polar regions. These instruments
include the E-Region Wind Interferometer, the meteor radar, the
Spectral Airglow Temperature Imager the PEARL All-Sky Imager, the
ozone and Rayleigh/Mie/Raman lidar, the VHF and cloud radar, the
Fourier Transform Spectrometer and the Atmospheric Emitted Radiance
Interferometer. Together these instruments provide the means to
determine the mean fields, and wave signatures associated with tides,
planetary waves and gravity waves from the stratosphere to the mesopause
region. Interpretation of these results will be supported with satellite
observations, model results and analyses from data assimilation.
Collaborations are being developed with other polar observatories
so that a global view of these processes in the Arctic middle atmosphere
can be developed. This effort will peak during International Polar
Year.
--------------------------------------------------------------------------------
Variability of Assimilated Ozone in
the Upper Troposphere and Lower Stratosphere
by
Kris Wargan
Global Modeling and Assimilation Office
Coauthors: I. Stajner, M. Sienkiewicz, S. Pawson, L. Froidevaux,
N. Livesey, P. K. Bhartia
Ample evidence suggests that constraining constituent models with
observations through data assimilation typically improves values
of constituent concentrations in comparison with independent measurements.
However, insertion of analysis increments into a model can introduce
spurious variability or smooth out fine scale features. The question
that we attempt to address is: How well and over what spatial scales
can assimilated ozone reproduce the variability of ozone fields?
This study uses level 2 ozone data from the EOS Aura satellite.
Ozone Monitoring Instrument (OMI) total column data and Microwave
Limb Sounder (MLS) ozone profiles are assimilated into the GEOS-4
General Circulation Model at NASA’s Global Modeling and Assimilation
Office (GMAO). Independent aircraft data were obtained from the
MOZAIC program. Statistical comparisons are presented for the Upper
Troposphere – Lower Stratosphere (UTLS) region in the Northern
Hemisphere for selected months in 2005. Distributions of ozone mixing
ratios and their two-point differentials are shown. Power spectra
of ozone from the aircraft data are compared to those obtained from
the assimilation and model simulations. These comparisons will provide
information about the scales at which ozone distribution in the
UTLS and the transport of ozone through the UTLS are well represented.
As a further test of assimilation results, impacts of assimilated
ozone on the numerical weather prediction skill are investigated.
In these experiments ozone data from OMI and MLS are assimilated
into the GEOS-5 system at the GMAO. The analyzed ozone is then used
in the forward model for assimilation of infrared radiances (e.g.
from the AIRS instrument) as well as in the radiative heating rate
computations in the general circulation model. We demonstrate large
improvements in the GEOS-5 ozone fields and modest impacts on the
weather prediction skill.
--------------------------------------------------------------------------------
Constraining Zonal Mean Flow and Diurnal
Tide by Space-borne Data
by
Valery A. Yudin
NCAR/ACD, PO Box 3000, Boulder, Colorado 80305-3000 USA
This paper examines feasibility and limitations to use the satellite-based
horizontal winds and temperatures to evaluate and constrain the
balanced and forced components of model predictions in the upper
atmosphere. The motivation of this study is to formulate the family
of statistical estimation schemes that can provide constraints on
the “missing” momentum tendencies, frequently associated
with gravity effects. I discuss “technical” similarities
between the data assimilation of wind data and application of sub-grid
GW schemes in the numerical models. Several GW parameterization
schemes have been considered to characterize their effects in the
“windless atmosphere”. The sensitivity of their numerics
to specification of launch level parameters and background fields
are discussed. The brief overview of the recent global HIRDLS temperature
and tracer retrievals for studies of short-scale vertical waves
will be presented.
--------------------------------------------------------------------------------
Intercomparison of equatorial waves in
the stratosphere in three data assimilation systems
by
Nedjeljka Zagar
NCAR/ASP
Coauthors: T. Shepherd, S. Polavarapu, M. Reszka and A. Jonsson
A significant part of the observed large-scale tropical variability
is represented by equatorially trapped Rossy, Kelvin, mixed Rossby-gravity
and equatorial inertio-gravity waves. Because of their vertical
propagation, these waves are a driving mechanism for the tropical
middle atmosphere and thus important to include in data assimilation
procedures.
This study compares stratospheric variance distribution among various
tropical wave motions in three global data assimilation systems
(CMAM, CMC and ECMWF models). Analysis is performed by projecting
dynamical fields (analysis increments and background errors) onto
the analytical equatorial wave solutions. Obtained spectra are used
as a forcing for an idealized stratospheric model to compare horizontal
correlation structures and balance properties of data assimilation
in various systems.
Back to Top
