300
Air Pollution and Turbulence: Modeling and Applications
et al. (2003). Rayner (2004) has extended network design calculations through
optimization in the presence of model error, this being based on results from
TransCom.
Two areas of ongoing research where trace inversions are being used for “experimental design” described in the following sections are
Analyzing the utility of continuous data (hourly or better time-resolution)
•
Analyzing the utility of satellite data
•
Most global-scale inversions of CO 2 have used monthly mean data, obtained
mainly from fl ask sampling. However, there are long-term continuous measurements from the NOAA CMDL (NOAA GMCC) program at the South Pole, Samoa,
Mauna Loa, and Point Barrow. Other national programs also maintain continuous
analyzers.
The development of the LoFlo analyzer (Francey and Steele, 2003) potentially
allows wider deployment of continuous analyzers by reducing the demands for
calibration gases. Law and collaborators have undertaken a series of experimental
design studies, exploring the utility of various modes of deployment (Law et al.,
2002, 2003b, 2004).
Inversions using continuous data (often inverting hourly values) place severe
demands on the modeling:
1. The transport modeling needs to be based on analyzed winds in order to
match specifi c events, in contrast to low-resolution inversions that often use
transport fi eld generated by GCMs.
2. The statistical modeling needs to capture a complicated space–time structure of the transport fi elds.
For experimental design studies, point (1) is not a problem, since the analysis can be
done using synthetic data. Point (2) remains a challenge to researchers. Additional
research is being conducted through the TransCom continuous data activity.
The potential for measuring CO 2 concentrations from satellites brings the prospect of global-scale coverage at high spatial resolution and regular time sampling.
Against this stand the technical diffi culties of achieving reliable measurements from
space. As well as future missions targeted at CO 2 (OCO and GOSAT) there have also
been efforts to obtain CO 2 information from existing instruments (e.g., AIRS and
SCIAMACHY) that were designed for other purposes.
This creates a valuable role for “experimental design” studies that assess the
potential utility of satellite data. Initial studies (Rayner and O’Brien, 2001a,b;
O’Brien and Rayner, 2002) suggested that a useful reduction in uncertainty in estimated fl uxes could be achieved with column-integrated values with an 8° × 10°
footprint and 2.5 ppm uncertainty on monthly mean values. Pak and Prather (2001)
compared the relative utility of various cases such as tropospheric versus whole
column and spatial resolution versus coverage. They noted in particular that the
most useful data would be for tropical regions, that is, those that currently have
© 2010 by Taylor and Francis Group, LLC
Air Pollution and Turbulence: Modeling and Applications
et al. (2003). Rayner (2004) has extended network design calculations through
optimization in the presence of model error, this being based on results from
TransCom.
Two areas of ongoing research where trace inversions are being used for “experimental design” described in the following sections are
Analyzing the utility of continuous data (hourly or better time-resolution)
•
Analyzing the utility of satellite data
•
Most global-scale inversions of CO 2 have used monthly mean data, obtained
mainly from fl ask sampling. However, there are long-term continuous measurements from the NOAA CMDL (NOAA GMCC) program at the South Pole, Samoa,
Mauna Loa, and Point Barrow. Other national programs also maintain continuous
analyzers.
The development of the LoFlo analyzer (Francey and Steele, 2003) potentially
allows wider deployment of continuous analyzers by reducing the demands for
calibration gases. Law and collaborators have undertaken a series of experimental
design studies, exploring the utility of various modes of deployment (Law et al.,
2002, 2003b, 2004).
Inversions using continuous data (often inverting hourly values) place severe
demands on the modeling:
1. The transport modeling needs to be based on analyzed winds in order to
match specifi c events, in contrast to low-resolution inversions that often use
transport fi eld generated by GCMs.
2. The statistical modeling needs to capture a complicated space–time structure of the transport fi elds.
For experimental design studies, point (1) is not a problem, since the analysis can be
done using synthetic data. Point (2) remains a challenge to researchers. Additional
research is being conducted through the TransCom continuous data activity.
The potential for measuring CO 2 concentrations from satellites brings the prospect of global-scale coverage at high spatial resolution and regular time sampling.
Against this stand the technical diffi culties of achieving reliable measurements from
space. As well as future missions targeted at CO 2 (OCO and GOSAT) there have also
been efforts to obtain CO 2 information from existing instruments (e.g., AIRS and
SCIAMACHY) that were designed for other purposes.
This creates a valuable role for “experimental design” studies that assess the
potential utility of satellite data. Initial studies (Rayner and O’Brien, 2001a,b;
O’Brien and Rayner, 2002) suggested that a useful reduction in uncertainty in estimated fl uxes could be achieved with column-integrated values with an 8° × 10°
footprint and 2.5 ppm uncertainty on monthly mean values. Pak and Prather (2001)
compared the relative utility of various cases such as tropospheric versus whole
column and spatial resolution versus coverage. They noted in particular that the
most useful data would be for tropical regions, that is, those that currently have
© 2010 by Taylor and Francis Group, LLC