Tropospheric Chemistry
MOPITT: Measurement of Pollution in The Troposphere
The Measurement Of Pollution In The Troposphere (MOPITT) experiment (John Gille, David Edwards, Charles Cavanaugh, Jarmei Chen, Cheryl Craig, Merritt Deeter, Louisa Emmons, Gene Francis, David Grant, Shu-Peng Ho, Boris Khattatov, Jean-Francois Lamarque, Deborah Mao, Daniel Packman, Barbara Tunison, Juying Warner and Daniel Ziskin) is a joint Canadian-U.S. effort to measure and interpret the global distributions of tropospheric carbon monoxide (CO) and methane (CH4), thus supporting studies of the oxidizing capacity of the lower atmosphere on large scales. The experiment, which was launched on board the NASA Earth Observing System (EOS) Terra spacecraft in December, 1999, represents a significant advancement in the application of space-based remote sensing to global tropospheric chemistry research.
During its planned 5-year mission, MOPITT will continuously scan the atmosphere to provide the first long term, global measurements of CO and CH4 levels in the troposphere. These will be essential to studies of pollution sources and transport. Measurements of CO provide a "window" into troposphere chemistry. As CO levels rise, they affect the abundance of atmospheric OH which usually functions to clean the atmosphere of many pollutants. CO and CH4 oxidation processes are also very important in determining the tropospheric O3 budget. CO has a limited lifetime of about two months and can be used as a regional tracer of pollution and transport. It is one of the few tropospheric gases that can be successfully monitored from space.
Before MOPITT, there had been no long-term global measurements of CH4. Methane is an important greenhouse gas whose atmospheric concentration has changed in recent times, but for which the chemical processes governing its creation and consumption are not well understood. Global measurements are essential to improving our understanding of its future evolution and climactic impact.
The instrument is a joint Canadian Space Agency/NASA project led by Instrument Principal Investigator James Drummond, University of Toronto. The U.S. Principal Investigator is John Gille who leads the group in NCAR/ACD. NCAR is responsible for the continued development of the algorithms and for operational data processing at every stage from instrument counts at Level 0 through calibrated radiances at Level 1 to retrieved carbon monoxide and methane products at Level 2. These are: the global vertically resolved distributions of tropospheric carbon monoxide (CO), total column of CO, and total column of CH4. More information can be found at http://www.eos.ucar.edu/mopitt/ and http://www.atmosp.physics.utoronto.ca/MOPITT/home.html.
MOPITT Data Processing and Retrieval Algorithm Development
In the years leading up to launch, the ACD MOPITT group developed the tools necessary to receive and process the instrument measurements. After launch, additional effort in this area was required in order to properly understand the complexity of the instrument and to fully characterize the measurements which showed some unexpected artifacts. The elements of this processing capability and the people responsible are: the data handling interfaces and protocols between NCAR and the NASA centers which receive and archive the satellite data and the ancillary meteorological data (Ziskin and Chen); the Level 0-1 processor which calibrates the instrument counts to produce geolocated radiances (Ziskin, Mao, and Chen); and the Level 1-2 processor which comprises a forward model which provides a full simulation of the MOPITT measurement, a cloud detection algorithm, and a retrieval algorithm (Edwards, Francis, Warner, Deeter, Grant, Ho, and Gille). The retrieval combines information from the measurements, the forward model, and previous measurements that define the current understanding of the atmosphere, to obtain the most likely CO profile or CH4 column consistent with the measured MOPITT signal.
Work has continued on the characterization of the MOPITT radiances and the development of improved algorithms. Ho has worked on issues related to instrument noise, and recent improved understanding in this area will hopefully allow recovery of 40% of the MOPITT data which had to be removed from the processing stream due to excessive noise. The problem of radiance biases at 4.7 micrometers noted soon after launch has been successfully resolved by Deeter, Francis, Edwards and Gille, with a redefinition of the modeling of the optical filters. This work has been verified by comparing MOPITT radiances with forward model predictions for well-characterized scenes (Deeter et al., J. Geophys. Res., submitted).
Recently, there have been significant advances in the understanding of the high noise levels that were observed in the 2.2 micron solar reflectance channels that are used to give information about the CO and CH4 total atmospheric columns. The identification of two contributing effects to the noise, one instrumental and the other geophysical, should enable a useful CH4 product to be obtained in the near future.
Algorithm development has continued at level 2, with enhancements to the forward model, retrieval, and cloud detection codes. Work is now underway to examine the possibility of making CO retrievals under partly cloudy conditions. This work makes use of the cloud products from the Terra/Moderate Resolution Imaging Spectrometer (MODIS) instrument.
The instrument has experienced two anomalies this year. In May, MOPITT experienced a cooler failure which resulted in the loss of 4 of the 8 channels. In August, there was a chopper failure for 2 of the remaining channels, although fortunately the chopper stuck in the open position, allowing the data to be used with a modified calibration. The instrument is now back in science mod, and a CO product is being retrieved with reduced vertical sensitivity. This has required a new version of the retrieval scheme to be developed which uses data from a smaller number of channels.
Figure 1: Example of the combined MOPITT 700 mb data for the first three days of August, 2000. Global coverage is almost complete in three days, although no retrieval is possible in those regions of persistent cloud cover, and retrievals are not presently performed at latitudes higher than 65 degrees. Elevated CO levels associated with industrial pollution are evident over the U.S. and China, with significant Asian outflow into the Pacific Ocean. Biomass burning provides an important source of CO in Africa, South America, and India at this time of year.
Modeling and Data Assimilation
The MOPITT modeling team (Khattatov, Lamarque, and Yudin) has continued work on the assimilation of MOPITT data, profiles and total column independently, in the MOZART-2 global chemistry-transport model for the first 10 months of MOPITT data (year 2000). This assimilation takes into account the full information from averaging kernels. This in turn allows retrievals to be mapped to a regular time/space grid using sequential assimilation, and is very important in facilitating comparison of MOPITT measurements with other correlative data, especially when there is not an exact coincidence in time and location. Use of the ACD/HANK model is currently being investigated for assimilating MOPITT retrievals on the regional scale.
In collaboration with graduate student Gabrielle Petron and Claire Granier (University of Paris), Lamarque and Yudin developed a prototype inversion scheme for deriving surface fluxes of carbon monoxide using MOPITT data and the MOZART-2 model.
MOPITT Airborne Test Radiometer
During 2001, the MOPITT Airborne Test Radiometer (MATR) team (Hills [RAP], Jianguo Niu, and Deeter) operated the instrument in two series of flights: (1) during April and May over the Oklahoma CART site, and (2) during November over three Western U.S. cities (Los Angeles, Las Vegas, and Denver). These flights were supported by in-situ measurements by Paul Novelli (NOAA CMDL).
Recently, emphasis has been placed on quantifying sources of radiance bias and noise for both the longwave LMC and PMC channels. One technique found to be useful has been to fly the instrument at low altitude and use the observed 'D' signal to calculate a systematic radiance offset. Multi-level flight data has also been helpful in characterizing radiance biases. Currently, LMC and PMC retrieval results agree fairly well, although PMC retrieval results appear much noisier. Analysis of retrieval results for recent MATR measurements over western cities is underway.
MOPITT Data Validation and Science Activities
In the months since launch, the group has been heavily involved in validation activities. These are essential at each level of the data processing to ensure full understanding of the in-flight MOPITT performance to allow characterization of measurement accuracy, precision, and resolution, and to point the way to needed improvements. In most of this validation activity, a step-by-step approach has been used starting with the simpler situations and learning from the results before moving on to more complicated cases.
At Level 0-1 there has been a careful examination of the instrument engineering data on a daily basis. The resulting calibrated radiances are being extensively compared with model calculations for special cases where the measurement scenes are believed to be well characterized. This work has led to studies by Deeter to characterize the way in which retrieval errors are related to radiance uncertainties.
Warner developed a MOPITT/MODIS collocation algorithm that enables the use of the MODIS cloud mask in the MOPITT cloud algorithm. This allowed validation of the MOPITT Cloud Algorithm (MOPCLD) output by comparing to MODIS products and has resulted in algorithm improvements based on the comparison.
At level 2, Emmons, Yudin and Jean-Luc Attie, (Laboratoire d'Aerologie, Toulouse, France), have continued the work of comparing individual retrieved CO profiles with in-situ measurements taken by Novelli's CMDL group. Comparisons have also been made with CO total column retrievals from ground based spectrometers, coordinated by Nikita Pougatchev (Christopher Newport University). These show good agreement in terms of seasonal trends.
Several investigations and collaborations are currently underway to validate aspects of the data under a variety of situations and to show the reliability of the data for scientific studies. Science questions under investigation include examination of global and seasonal CO and CH4 distributions; improved estimation of CO and CH4 sources using MOPITT data and inverse modeling; examination of biomass burning sources and transport of pollutants; the role of mega-cities in pollutant production and a potential involvement of MOPITT in the ACD Megacity Impacts (MCI) Initiative; the impact of the recent U.S. western fires on tropospheric chemistry; and the use of elevated CO in mid/upper troposphere as a signature of convective processes. MOPITT data was provided to the Transport and Chemistry Evolution over the Pacific (TRACE-P) team looking at pollution outflow over the Pacific from Asia for use in aircraft flight planning in the Spring of 2001. Scientific collaboration with TRACE-P will be pursued.
Data at both Level 1 and Level 2 are currently being delivered to the NASA Langley Data Archive Analysis Center for distribution to the community. To assist users of MOPITT data, the MOPITT web page has been redesigned with links to available documentation, data and information essential to data users such as description of the proper technique for calculating and applying the MOPITT CO averaging kernels.