 |
|
||
|
|
Global Chemical Tropospheric Modeling (GCTM) GroupGCTM Group Members:
Summary of Activities:1. Regional and Global Air Quality
1.1 Impact of meteorology on global chemistry simulations
Using the recently finalized Version 4 of the Model for Ozone and Related chemical Tracers (MOZART-4), the impact of analyzed meteorological fields on global chemical simulations has been studied ( Sensitivity of chemical budgets to meteorology in MOZART-4, Emmons et al, in preparation 2006). Simulations with MOZART-4 driven with NCEP/NCAR Reanalysis and ERA-40 (ECMWF Reanalysis) meteorological fields show significant differences in the chemical budgets of a number of compounds. The diagnosis of clouds, and therefore convection and precipitation vary significantly in the 2 reanalyses, resulting in significant differences in OH concentrations and washout of aerosols and trace gases. Figure 1 compares the annual average of the aerosol optical depth from the 2 MOZART simulations compared to the observations from the MODIS instrument on the Terra satellite. The most significant differences are seen away from source regions, where the transport and washout of aerosols have a strong influence on the distributions.
Figure 1: Aerosol optical depth from a MOZART-4 simulation using NCEP/NCAR reanalysis meteorology, ECMWF ERA-40 reanalysis meteorology and from the MODIS instrument on the Terra satellite.
1.2 Constraining surface emissions
Constraining surface emissions of atmospheric species is crucial in constraining their budgets and is a critical first-step in understanding the impacts of emissions on global air-quality. One common methodology in constraining emissions is the use of “top-down” emission estimates from satellites. However, estimates of CO sources derived from inversions using satellite measurements still exhibit persistent discrepancies. Arellano and Hess (Sensitivity of top-down estimates of CO sources to GCTM transport, submitted to GRL) carry out controlled inverse analyses to elucidate the influence of model transport on the robustness of regional estimates of CO sources. We utilized 2 widely-used GTCMs (MOZART and GEOS-Chem) driven by 3 currently available meteorological datasets (NCEP, ECMWF and GMAO) to generate response functions for prescribed regional CO sources. We find that inter-model differences in CO due to differences in transport are within 10-30% of the inter-model mean CO concentration (see Figure 2). However, these differences can translate to regionally significant spread in the source estimates. While we find that CO source estimates for East Asia and North Africa are reasonably robust, we find inconsistencies and inter-model spread of greater than 40% in our source estimates for Indonesia , South America , Europe and Russia . This clearly indicates that the current estimation of top-down emission uncertainty fails to account for transport bias.
Figure 2. Prior (black) and posterior source and error (2) estimates (in Tg CO) derived from the 3 sets of inverse analysis using MZ4 NCEP in red, MZ4 ERA40 in blue, and GEOS-Chem in green. The estimates are divided into 3 categories depending on the degree of consistency between estimates: A. consistent, B. moderately consistent, C. inconsistent. Regions are define as: NAM, North America; EUR, Europe; RUS, former Soviet Union; NAF, Northern Africa; CAM/NSAM, Central America/Northern South America; OCN, Oceania; IND, Indonesia and Malay archipelago; SAS, South Asia; SEA, Southeast Asia; EAS, East Asia; MDE, Middle East; SAF, Southern Africa; SAM, South America
MOZART-4 has also been used to constrain sources of CO during 2004 over South America from biomass burning, anthropogenic emissions, biogenic CO emissions and isoprene oxidation (Pfister et al., in preparation). The biomass burning source is estimated as 80 Tg CO yr -1 which accounts to close to 30% of the global biomass burning emissions and reflects the importance of South American biomass burning sources (Figure 3). No robust constraint could be put on biogenic, anthropogenic and isoprene sources, because the contributions from these sources are within the range of uncertainties associated with observations and model results. However, the study reveals the importance of CO produced from isoprene oxidation in CO budgets on both the regional and global scale and thus the need for improved techniques for constraining isoprene emission inventories.
Figure 3. Comparison of priori (black), posteriori (red) and GFED-2 biomass burning emissions over South America for 2004.
Pfister et al. (manuscript in preparation) examine how isoprene and differences in its emission inventory affect tropospheric composition. For this purpose they extended the current MOZART chemical scheme to track carbon-containing species produced from isoprene. From the model simulations they estimate the contribution of isoprene to the global CO burden below 100 hPa as 15% (Figure 4). Contributions for HCHO are ~20% and for PAN ~30%. Omitting isoprene in the simulations has a rather small effect on the global annual burden of ozone (~4%), but regionally the impact on ozone can be significant (e.g. changes in surface ozone levels of up to 20 ppbv occur over the Eastern US in summertime). They also look into the feasibility of using the model tracers in deriving isoprene emissions estimates from space based HCHO observations.
Figure 4: Contribution (%) of CO from isoprene oxidation to the total CO column.
1.4 Impact of Biomass burning on global air quality
MOZART-4 simulations and CO and ozone observations at the PICO-NARE station in the Azores examined the amount of ozone produced from wildfires in Alaska and Canada in summer 2004 (Pfister et al., 2006; accepted). Modeled and observed enhancement ratios on the order of 0.25 ppbv/ppbv were calculated resulting in a global net ozone production from the fires of 12.9±2 Tg O3. On average, the wildfires increased the ozone burden (surface-300 mbar) over Alaska and Canada by 7-9% during summer 2004, and over Europe by 2-3% (Figure 5).
Figure 5: MOZART estimate for the percentage change in the ozone column amount over July15-25, 2004 due to the fires in Alaska and Canada
1.5 Impact of Megacities and Intercontinental Transport on Global Air Quality
The ACD global tropospheric modeling group participated in the MILAGRO and INTEX-B measurement campaigns. These campaigns were designed to: 1) assess the impacts of megacities on regional to global atmospheric chemistry; 2) quantity the transpacific transport and evolution of Asian pollution to North America and assess its importance for regional air quality and climate. The MOZART-4 chemical transport model was used to assimilate real-time CO from the MOPITT satellite and then produce global forecasts at approximately 50 km resolution of 3 days. These forecasts were then used for flight planning purposes. Figure 6 shows a curtain plot of the model forecast CO field on May 4 along the DC-8 flight track, and a comparison of the forecast and measured CO on that day.
Figure 6: a) Curtain plot of MOZART CO forecast on May 4 and the profile of the flight track; b) comparison of modeled (red) and measured CO (blue) on May 4, with the flight altitude shown in black.
1.6 Ensemble data assimilation
The onset of near-global and long-term measurements of chemical species from space offers an opportunity to better understand changes in atmospheric composition from regional to global scales through the integration of measurements with global chemical transport models (GCTMs). The goal of this research project is to build a state-of-science chemical data assimilation system to provide a flexible platform for related studies in chemical weather forecasts, assessing the impact of global air pollution to regional air quality, and estimating anthropogenic emissions. For the past year, Arellano and Hess have focused on building an ensemble-based system (i.e. using Ensemble Kalman Filter approach) in close collaboration with Dr. Jeffrey Anderson (NCAR/DAReS). The assimilation system includes the Data Assimilation Research Testbed (DART) software developed by Dr. Jeffrey Anderson and the Community Atmosphere Model (CAM) which is the atmospheric component of the global climate model Community Climate System Model (CCSM).
The initial focus was to use the DART/CAM setup by assimilating meteorological observations and MOPITT CO retrievals. This first step builds on previous inverse modeling and assimilation studies of the CO tracer, with the added complexity of assimilating meteorological variables and making use of ensembles instead of a single representation of the atmosphere. Analyses characterizing uncertainties in modeling CO transport and its impact on emission estimates were conducted to gain more insights on the sensible estimates of CO distribution and CO emissions (Sensitivity of top-down estimates of CO sources to GCTM transport, Arellano and Hess, submitted to GRL). The outcome of the analyses provided a more quantitative means to generate initial ensembles of CO state and emissions necessary for the ensemble-based data assimilation.
The installation and testing of DART/CAM setup included the following activities: 1) addition of CO tracer routines to CAM3.1 (emissions, and chemistry), 2) modification of DART to support the model horizontal grid based on finite volume dynamical core of CAM3.1, 3) addition of CO tracer and observations (MOPITT CO) to DART, and 4) creation of initial ensembles of CO based on perturbed dynamical states and perturbed emissions. The setup was first tested using pseudo-data experiments or observing system simulation experiments (OSSEs). Results show that assimilating simultaneous synthetic observations of radiosonde temperature (T), radiosonde and satellite-derived horizontal wind velocity (U,V) and MOPITT CO retrieval for 700 hPa together provide better constraints for the CO distribution. Results using real data also show the ability of the current assimilation system to constrain the model state variables. An ensemble of dynamical states to drive the chemical evolution of CO not only produces consistent analyses but more importantly it enables to better capture the variability of the CO fields needed for successful assimilation.
1.7 Ozone Variability
Ozone acts as both a radiatively important gas and a primary pollutant adversely affecting human health and the environment. Extensive emission controls have been implemented in the U.S. and other countries to reduce surface ozone concentrations, with a significant improvement noted in U.S. emissions. On the other hand significant growth in surface ozone has been reported on the west coast of the U.S. and at Mace Head Ireland with substantial interannual variability. Northern mid-latitude ozone data between 1970 and 1996 also suggests substantial interannual variability with temporal trends varying by region. The combination of variability in ozone precursors and meteorological variability superimposed on climate change makes changes in the ozone record particularly difficult to interpret. Hess and Lamarque (Ozone source attribution and its modulation by the Arctic Oscillation during the spring months, in revision, JGR, 2006) show substantial ozone variability, up to 5 ppbv, can be explained by meteorological variability in association with the Arctic Oscillation (AO). Further this paper shows that much of this variability is explained by variations in stratosphere-troposphere exchange. Figure 7 shows measurements and simulated differences in the ozone mixing ratio between positive and negative phases of the AO.
Figure 7: Vertical profiles of measured and modeled ozone at North American Stations a) The green symbols give the 1990 minus 1980 (90M80) measurements at the various stations examined (Goose Bay: plus; Resolute: asterisk; Churchhill: diamond; Edmonton: triangle; Alert: square; Boulder, cross), where the dotted green line gives the median measured value. The black symbols give the measured ozone difference at these stations between the 6 years with the highest AO signal and the 6 years with the lowest AO signal in the 21 years examined, where the dotted black line gives the median value. Not all stations recorded measurements for 1980 and 1990. b) The solid lines give the 90M80 signal in the model and the dashed lines the regression between the model and the AO over all 21 simulations. The model signal is partitioned into anthropongenic O 3 (red), stratospheric O 3 (blue) and their sum (black). The diamonds on the dotted lines indicate those heights where the signal is significant at the 90% confidence level.
The CAM model with chemistry (see below) is being used to investigate the seasonal cycle of tropical tropospheric chemistry (Figure 8). In the left panel, all the ozonesondes between the Equator and 30 o N for the period 1998-2005 are used to represent the mean observed seasonal cycle in ozone. In the middle panel, the model, interpolated to the location of the stations, shows that, while the overall structure is similar, there are striking differences in the mid-troposphere seasonal cycle between the model and the observations. The right panel shows a similar simulation, but in which the lightning source is reduced by a factor of 2. This latter simulation results in a similar pattern, but with perhaps a better timing of the maximum in June.
Figure 8: Height-time sections of ozone between the Equator and 30 o N for the period 1998-2005. Time is given in months. See text for details.
2. Tropospheric Chemistry-Climate Studies
2.1 Development of the Community Atmosphere Model (CAM) with chemistry
The incorporation of interactive chemistry in the Community Atmosphere Model (CAM) has seen considerable progress during the last year. Many of the processes included in the chemistry-transport model MOZART have now been transferred to CAM . These include interactive photolysis rates, interactive dry deposition and interactive biogenic emissions (in collaboration with C. Heald, University of California , Berkeley ). A number of different options exist for simulating aerosols and chemistry to facilitate using the model in the optimal configuration for extended chemistry-climate studies. Aerosols can either be prescribed, simulated using simple input oxidant fields, or simulated using the full MOZART-4 aerosol parameterization. Lamarque et al. (in preparation) developed a reduced mechanism in which the number of hydrocarbons is quite reduced from the MOZART-4 mechanism (leading to a speed-up of a factor of 2. This reduction has been analyzed and evaluated against measurements and has it reproduces many the main features of atmospheric chemistry relevant for climate studies, e.g. the ozone distribution and the methane lifetime. The indirect effect has been incorporated through collaboration with S. Ghan.
2.2 Biomass burning
Biomass burning has an important impact on climate. The interactions between climate change and biomass burning are not well understood. In a continuing investigation the MOZART and the CAM models have been used to investigate the radiative forcing of ozone and carbon aerosols from wildfires in Alaska and Canada . Some of these results have been incorporated into an exhaustive study of the impact of biomass burning on the boreal forest (Randerson et al., Boreal forest fires cool climate, Science, in revision 2006). This paper shows that despite the release of large amounts of carbon dioxide, aerosols and other trace species into the atmosphere by fires, the ability of fire to change surface albedo dominates the net climactic affect of fires, leading to cooling.
2.3 Climate Change and Air-Quality
The impact of climate change on U.S. surface ozone levels is investigated in Murazaki and Hess (JGR, 2006). Murazaki and Hess simulated two 10 year periods using the global chemical transport model MOZART-2 (Model of Ozone and Related chemical Tracers version 2): 1990–2000 and 2090–2100. In each case, MOZART-2 is driven by meteorology from the National Center for Atmospheric Research (NCAR) coupled Climate Systems Model (CSM) 1.0 forced with the Intergovernmental Panel on Climate Change (IPCC) Special Report on Emissions Scenarios (SRES) A1 scenario. During both periods the chemical emissions are fixed at 1990s levels, so that only changes in climate are allowed to impact ozone. The impact of climate change is calculated separately for background ozone and for the ozone generated through U.S. NO x emissions. The results show that the response of ozone to climate change in polluted regions is not the same as in remote regions. MOZART-2 predicts a 0–2 ppbv decrease in background ozone in the future simulation over the United States but an increase in ozone produced internally within the United States of up to 6 ppbv. The decrease in background ozone is attributed to a future decrease in the lifetime of ozone in regions of low NO x . Over the western United States the decrease in background ozone approximately cancels the increase in locally produced ozone. As a result, the main impact of future climate change on ozone is centered over the eastern United States , where future ozone increases up to 5 ppbv. We predict that in the future over the northeast United States , up to 12 additional days each year will exceed the maximum daily 8-hour averaged ozone limit of 80 ppbv (Figure 8). This is an approximate increase of 50%. Various climatic factors are identified which impact the net future increase in ozone over the United States including changes in temperature, water vapor, clouds, transport, and lightning NO x . Significant future changes are generally not found in planetary boundary layer height and precipitation.
Figure 9: Average difference between the future and control simulations of the number of days in a year the maximum daily 8 hour averaged ozone concentration is greater than 80 ppbv
2.4 Paleoclimate Studies
Application of atmospheric chemistry to paleoclimate studies is still very much a developing field. Lamarque et al ( Modeling the response to changes in tropospheric methane concentration: Application to the Permian-Triassic boundary, Paleoceanography, 2006) performed a set of simulations (with a version of WACCM (Whole Atmosphere Community Climate Model) relevant to the Permian-Triassic boundary (about 250 Ma). It is frequently assumed that during this time large amounts of methane (most likely from methane clathrates at the bottom of the ocean) were released. In the simulations, the amount of methane that reaches the atmosphere was varied over a range of possible values. Methane at concentrations larger than 1000 times pre-industrial levels result in a collapse of the total ozone column (Figure 9). This leads to a very large increase in the amount of UV radiation that reaches the ground. Lamarque et al. postulate that this could be a mechanism for the mass extinction that occurred at the Permian-Triassic boundary .
Figure 10: Column ozone against the increase of the surface methane over its pre-industrial level.
|
|
