• Cotton, W. R., and Coauthors, 2003: RAMS 2001: Current status and future directions. Meteor. Atmos. Phys., 82 , 529.

  • Deeter, M., , and K. F. Evans, 1998: A hybrid Eddington-single scattering radiative transfer model for computing radiances from thermally emitting atmospheres. J. Quant. Spectrosc. Radiat. Transfer, 60 , 635648.

    • Search Google Scholar
    • Export Citation
  • Derber, J. C., , and W. Wan-Shu, 1998: The use of TOVS cloud-cleared radiances in the NCEPSSI analysis system. Mon. Wea. Rev., 126 , 22872299.

    • Search Google Scholar
    • Export Citation
  • Errico, R. M., 1997: What is an adjoint model? Bull. Amer. Meteor. Soc., 78 , 25772591.

  • Errico, R. M., , and T. Vukicevic, 1992: Sensitivity analysis using an adjoint of the PSU–NCAR Mesoscale Model. Mon. Wea. Rev., 120 , 16441660.

    • Search Google Scholar
    • Export Citation
  • Evans, K. F., 1998: The spherical harmonics discrete ordinate method for three-dimensional atmospheric radiative transfer. J. Atmos. Sci., 55 , 429446.

    • Search Google Scholar
    • Export Citation
  • Greenwald, T. J., , R. Hertenstein, , and T. Vukicevic, 2002: An all-weather observational operator for radiance data assimilation with mesoscale forecast models. Mon. Wea. Rev., 130 , 18821896.

    • Search Google Scholar
    • Export Citation
  • Greenwald, T. J., , T. Vukicevic, , L. D. Grasso, , and T. H. Vonder Haar, 2004: Adjoint sensitivity analysis of an observational operator for visible and infrared cloudy-sky radiance assimilation. Quart. J. Roy. Meteor. Soc., 130 , 685705.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., 2003: Atmospheric Modeling, Data Assimilation and Predictability. Cambridge University Press, 341 pp.

  • Kopken, C., , G. Kelly, , and J-N. Thepaut, 2004: Assimilation of Meteosat radiance data within the 4D-Var system at ECMWF: Assimilation experiments and forecast impact. Quart. J. Roy. Meteor. Soc., 130 , 22772292.

    • Search Google Scholar
    • Export Citation
  • McMillin, L. M., , L. J. Crone, , M. D. Goldberg, , and T. J. Kleespies, 1995: Atmospheric transmittance of an absorbing gas. 4. OPTRAN: A computationally fast and accurate transmittance model for absorbing gases with fixed and variable mixing ratios at variable viewing angles. Appl. Opt., 34 , 62696274.

    • Search Google Scholar
    • Export Citation
  • Menzel, W. P., , and J. F. W. Purdom, 1994: Introducing GOES-I: The first of a new generation of geostationary operational environmental satellites. Bull. Amer. Meteor. Soc., 75 , 757781.

    • Search Google Scholar
    • Export Citation
  • Menzel, W. P., , F. C. Holt, , T. J. Schmit, , R. M. Aune, , A. J. Schreiner, , G. S. Wade, , and D. G. Gray, 1998: Application of GOES-8/9 soundings to weather forecasting and nowcasting. Bull. Amer. Meteor. Soc., 79 , 20592077.

    • Search Google Scholar
    • Export Citation
  • Meyers, M. P., , R. L. Walko, , J. Y. Harrington, , and W. R. Cotton, 1997: New RAMS cloud microphysics parameterization. Part II: The two-moment scheme. Atmos. Res., 45 , 339.

    • Search Google Scholar
    • Export Citation
  • Stokes, G. M., , and S. E. Schwartz, 1994: The Atmospheric Radiation Measurement (ARM) Program: Programmatic background and design of the cloud and radiation test bed. Bull. Amer. Meteor. Soc., 75 , 12011221.

    • Search Google Scholar
    • Export Citation
  • van de Hulst, H. C., 1957: Light Scattering by Small Particles. Dover, 470 pp.

  • Vukicevic, T., , T. Greenwald, , M. Zupanski, , D. Zupanski, , T. Vonder Haar, , and A. S. Jones, 2004: Mesoscale cloud state estimation from visible and infrared satellite radiances. Mon. Wea. Rev., 132 , 30663077.

    • Search Google Scholar
    • Export Citation
  • Vukicevic, T., , M. Sengupta, , A. S. Jones, , and T. Vonder Haar, 2006: Cloud-resolving satellite data assimilation: Information content of IR window observations and uncertainties in estimation. J. Atmos. Sci., 63 , 901919.

    • Search Google Scholar
    • Export Citation
  • Walko, R. L., , W. R. Cotton, , M. P. Meyers, , and J. Y. Harrington, 1995: New RAMS cloud microphysics parameterization. Part I: The single moment scheme. Atmos. Res., 38 , 2962.

    • Search Google Scholar
    • Export Citation
All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 10 10 3
PDF Downloads 4 4 0

Analysis of Information Content of Infrared Sounding Radiances in Cloudy Conditions

View More View Less
  • 1 Cooperative Institute for Research in the Atmosphere, Colorado State University, Fort Collins, Colorado
© Get Permissions
Full access

Abstract

Information content analysis of the Geostationary Operational Environmental Satellite (GOES) sounder observations in the infrared was conducted for use in satellite data assimilation. Information content is defined as a first-order response of the top-of-atmosphere brightness temperature to perturbations of simulated temperature and humidity profiles, obtained from a cloud-resolving model, both in the presence and absence of clouds. Sensitivity to the perturbations was numerically evaluated using an observational operator for visible and infrared radiative transfer developed within a research satellite data assimilation system. The vertical distribution of the sensitivities was analyzed as a function of cloud optical thickness covering the range from a cloud-free scene to an optically thick cloud. The clear-sky sensitivities to temperature and humidity perturbations for each channel are representative of the corresponding channel weighting functions for a clear-sky case. For optically thin–moderate ice clouds, the vertical distributions of the sensitivities resemble clear-sky results, indicating that the use of infrared sounding observations in data assimilation can potentially improve temperature and humidity profiles below those clouds. This result is significant, as GOES infrared sounder data have until now only been used in cloud-cleared scenes. It is expected that the use of sounder data in data assimilation, even in the presence of optically thin to moderate high clouds, will help reduce errors in temperature and water vapor mixing ratio profiles below the clouds.

Corresponding author address: Tomoko Koyama, Cooperative Institute for Research in the Atmosphere, Colorado State University, Foothills Campus, Fort Collins, CO 80523-1375. Email: tomoko.kd@gmail.com

Abstract

Information content analysis of the Geostationary Operational Environmental Satellite (GOES) sounder observations in the infrared was conducted for use in satellite data assimilation. Information content is defined as a first-order response of the top-of-atmosphere brightness temperature to perturbations of simulated temperature and humidity profiles, obtained from a cloud-resolving model, both in the presence and absence of clouds. Sensitivity to the perturbations was numerically evaluated using an observational operator for visible and infrared radiative transfer developed within a research satellite data assimilation system. The vertical distribution of the sensitivities was analyzed as a function of cloud optical thickness covering the range from a cloud-free scene to an optically thick cloud. The clear-sky sensitivities to temperature and humidity perturbations for each channel are representative of the corresponding channel weighting functions for a clear-sky case. For optically thin–moderate ice clouds, the vertical distributions of the sensitivities resemble clear-sky results, indicating that the use of infrared sounding observations in data assimilation can potentially improve temperature and humidity profiles below those clouds. This result is significant, as GOES infrared sounder data have until now only been used in cloud-cleared scenes. It is expected that the use of sounder data in data assimilation, even in the presence of optically thin to moderate high clouds, will help reduce errors in temperature and water vapor mixing ratio profiles below the clouds.

Corresponding author address: Tomoko Koyama, Cooperative Institute for Research in the Atmosphere, Colorado State University, Foothills Campus, Fort Collins, CO 80523-1375. Email: tomoko.kd@gmail.com

1. Introduction

The direct assimilation of satellite radiances has been shown to significantly improve numerical weather prediction (NWP) and analysis (Derber and Wan-Shu 1998; Kalnay 2003; Kopken et al. 2004). However, only clear-sky data are typically used for forecast systems because current operational data assimilation schemes cannot handle cloudy cases. This is primarily because of two reasons (i) in cloudy conditions, large uncertainties in the cloud model make direct data assimilation impractical, and (ii) the complexity of the radiative transfer process in clouds leads to unfeasible computational requirements that cannot be handled by current operational data assimilation systems. This use of only cloud-cleared data ultimately leads to a large amount of spatial information, regarding temperature and moisture distributions, being excluded from atmospheric analysis. Additionally, initialization of cloud properties in the forecast systems is disregarded, leading to the possibility of a considerable reduction in forecast skill. Also, current models used in atmospheric analysis do not fully represent dynamical interactions between the clouds and their environment, and consequently, this leads to a limited usefulness of the information content of weather analysis in climate and meteorological research.

To enable the assimilation of satellite radiances under all weather conditions, the Regional Atmospheric Modeling and Data Assimilation System (RAMDAS) has been developed at the Cooperative Institute for Research in the Atmosphere at Colorado State University (CSU). RAMDAS includes a four-dimensional variational data assimilation algorithm with a cloudresolving model and an observational operator for visible (VIS) and infrared (IR) satellite radiances (VISIROO). VISIROO has the capability to address the question of the potential benefit of satellite data assimilation under all weather conditions.

VISIROO was initially used in the work of Greenwald et al. (2002, 2004) to evaluate the sensitivity of top-of-atmosphere measured radiances from the Geostationary Operational Environmental Satellite-9 (GOES-9) imager to cloud water and ice mixing ratio at 0.63, 3.9, and 10.7 μm (channels 1, 2, and 4 of GOES-9 as shown in Table 1) for several cases including a warm continental stratus case, a thunderstorm case, and a winter storm case. Although smaller liquid drops and ice particles for optically thin clouds generally contributed to the largest sensitivities, spectrally distinctive sensitivity features were observed, namely that (i) the sensitivity at 0.63 μm was the largest within optically thin water and ice clouds, (ii) the 3.9-μm band was the most sensitive to thick water clouds and to ice clouds, and (iii) the 10.7-μm band was the most sensitive to thinner ice clouds.

Following the results of the sensitivity analyses demonstrating the potential for the use of VISIROO in NWP, Vukicevic et al. (2004, 2006) performed numerical data assimilation experiments with GOES-9 imager observations. In the first study (Vukicevic et al. 2004), only 0.63- and 10.7-μm data were used for a warm continental stratus case. It was shown that observations using the two channels positively influenced the mesoscale model cloud forecast in the assimilation. However, the assimilation was not as successful in regions where clouds were present in the observations but not in the model.

In a follow-up study (Vukicevic et al. 2006), the GOES-8 imager data at 6.7, 10.7, and 12.0 μm were assimilated for a case of overcast mid- to high-level clouds. The study period also contained a warm boundary layer cloud below the high cloud. The experiment was performed for two periods using the original forecast as the background for (i) a case with a thin ice model cloud and (ii) no model clouds. The satellite observations showed overcast mid to high clouds for both periods, but the clouds dissipated quickly in the second period. In the first period, the modeled high clouds after the assimilation of satellite data were in very good agreement with observations, as independently verified by a ground-based cloud profiling radar located within the domain. However, in the second period, insufficient information about the moisture and temperature profiles in the observations led to a partial success. Also, the warm boundary layer clouds were not formed, as the imager channels were unable to sense below the high cloud and provide information. A comparison of moisture profiles from the model and soundings showed that relative humidity in the lower layers was too low to form a cloud, thereby displaying the need for information about temperature and moisture profiles in both clear and cloudy cases. Therefore, it is essential to introduce observations containing information about the atmospheric temperature and moisture profiles in cases where the forecast model underestimates clouds in order to improve the results from data assimilation.

The current study focuses on the possibility of using GOES sounder data to provide a moisture and temperature profiling capability under cloudy conditions. Information content analysis with VISIROO provides an estimation of the impact of each channel within the data assimilation system for both clear and cloudy conditions.

2. GOES sounder data

The GOES series has played an important role for weather analysis and forecasting since the mid-1970s. GOES-8 was launched on 13 April 1994 as the first three-axis stabilized spacecraft of the GOES series. Prior to GOES-8, the GOES instruments were the Visible and Infrared Spin Scan Radiometer (VISSR) used as an imager and its advanced sensor, the VISSR Atmospheric Sounder (VAS), used both for imaging and sounding. Although the imager data were successfully used for operational purposes, the VAS sounding data on GOES-4, -5, -6, and -7 were used only for special experiments. In July 1995, the National Environmental Satellite, Data, and Information Service began producing operational hourly GOES-8 temperature and moisture soundings. The GOES-8 sounder had higher temporal and spatial resolution than its previous counterpart with a better signal-to-noise ratio. These improved instrument capabilities enabled the depiction of atmospheric conditions (e.g., vertical variations of temperature, moisture, and ozone) for nowcasting severe weather and for NWP. The sounder measures radiances from 8-km footprints (with a spacing of 10 km) in 1 visible and 18 infrared spectral bands (Table 2), as detailed in Menzel and Purdom (1994) and Menzel et al. (1998).

3. Cloud-resolving forecast model in RAMDAS

The Regional Atmospheric Modeling System (RAMS) was developed at CSU by combining a cloud model and two hydrostatic mesoscale models. An overview of RAMS including recent developments and applications is presented by Cotton et al. (2003). The most important RAMS feature for this research is its bulk microphysics parameterization, allowing data assimilation in cloudy regions. The sophisticated microphysics of RAMS considers seven types of hydrometeors. Of those, a single-moment bulk scheme is applied to predict the cloud water mixing ratio (Walko et al. 1995). For pristine ice, aggregates, snow, graupel, hail, and rain, a two-moment bulk scheme is used to compute both the mixing ratio and the number concentration (Meyers et al. 1997). The frequency distribution of particle sizes is represented by a generalized gamma distribution function. A more sophisticated microphysical parameterization, the bin-resolving multimoment scheme, is also available. However, that scheme is computationally expensive and impractical for this work.

4. Observational operator in RAMDAS (VISIROO)

The VISIROO is a set of forward and adjoint models for visible and infrared channels of general satellite instruments. It was developed by Greenwald et al. (2002, 2004) as a major component of the RAMDAS algorithm. The forward observational operator uses the model state variables to compute quantities that can be compared directly with observations (radiances and brightness temperatures for comparison with satellite data in our case). The corresponding adjoint operator performs a sensitivity analysis of cloud state quantities under both clear and cloudy conditions, including multilayer clouds, under the plane-parallel assumption as shown in Fig. 1.

a. Forward observational model

The forward part of the system includes a cloud property model, a gas extinction model, and two different radiative transfer (RT) models depending on the radiation source. The two RT models used are the spherical harmonic discrete ordinate method (SHDOM; Evans 1998) for visible wavelengths and a delta-Eddington two-stream approach (e.g., Deeter and Evans 1998) for the infrared. It should be noted that both the models account for multiple scattering. Compared with other RT models available for visible computations, SHDOM is more accurate and can accommodate complicated scattering phase functions within practical computational times. A modified version of anomalous diffraction theory (ADT) developed by van de Hulst (1957) is used to calculate the scattering properties of the seven hydrometeor types that can occur in the model clouds. Gaseous extinction is computed using a single-band regression model called the Optical Path Transmittance (OPTRAN) method (McMillin et al. 1995). OPTRAN is used in operational data assimilation systems and depends on the model state variables of atmospheric temperature, pressure, and water vapor mixing ratio.

b. Adjoint observational model

The main function of an adjoint model is to evaluate the sensitivity of model output with respect to input (e.g., Errico and Vukicevic 1992; Errico 1997). By definition, the adjoint model is an adjoint of the linearized forward nonlinear observational operator. To verify the adjoint model in VISIROO, Greenwald et al. (2004) numerically tested the degree of linearity of the radiative response computed using the observational operator with respect to cloud mixing ratio perturbations and compared the tangent linear and adjoint computations. The results show that the observational operator response is mostly linear over a large range of perturbations and imply validity of the adjoint model in sensitivity computations.

5. Information content analysis

Solutions from VISIROO are used to study the information content of GOES sounding measurements in the presence of ice and liquid clouds of varying optical depths. The response of equivalent brightness temperature Tb to all perturbations of each input state variable Xi is represented by
i1520-0493-134-12-3657-e1
a Taylor’s series expansion. This relationship is called the tangent linear approximation, and the model state variables (Xi’s) used are temperature T (K), pressure p (hPa), cloud water mixing ratio rc (kg kg−1), and water vapor mixing ratio rυ (kg kg−1). The gradient of the cost function J, or the sensitivity of J to the input variables, is derived by applying the chain rule
i1520-0493-134-12-3657-e2
When the adjoint model forcing is set to unity, the cost function gradients become the Jacobians. These Jacobians can be derived at each vertical level for every channel wavelength. The linear responses of satellite measured brightness temperatures to perturbations in temperature and vapor mixing ratio at a model level l and wavelength λ can be defined as
i1520-0493-134-12-3657-e3a
i1520-0493-134-12-3657-e3b
respectively, and used as measures of information content. It is therefore obvious that VISIROO can be used to compute the brightness temperature sensitivities to perturbations in model state variables such as temperature and water vapor mixing ratio. In this study, the temperature and layer-averaged humidity of each model layer was perturbed by 1 K and 10% of the layer, respectively.

6. Results

The RAMS model integration domain is in northern Oklahoma where the Atmospheric Radiation Measurement Program’s central facility is located within a 600 km × 600 km area (Stokes and Schwartz 1994). The model horizontal grid spacing is 6 km with 84 vertical levels over a depth of 17 km, while the spatial resolution of the GOES sounder is approximately 10 km at the subsatellite point. Lateral boundary and initial condition data were obtained from the Eta Data Assimilation System daily weather analysis at 80-km horizontal resolution.

The forecast model was first integrated for 15 h starting at 0000 UTC 21 March 2000. Then data from all IR channels (channels 2–6 and 10–17) of the GOES-8 sounder were assimilated after 1200 UTC. During the experimental period, the domain was covered with disperse clouds that dissipated in association with the passage of a surface cold front. In the results shown here, the computed brightness temperatures using VISIROO correspond to 1500 UTC.

The experiment design of this research follows that of Greenwald et al. (2004) with the GOES imager observations. We studied the information content of GOES sounding measurements for clear sky and in the presence of ice and liquid clouds of varying optical depth. The experiments with liquid clouds (i.e., low-level warm stratus) show similar results to the previous study with the GOES imager. Therefore, we only discuss the results for the clear-sky and ice-cloud cases in this paper.

Information content computed using positive perturbations of temperature and humidity from Eqs. (3a) and (3b) should have positive and negative values, respectively. This can be physically explained by the fact that an increase in temperature in an atmospheric layer leads to a higher amount of radiative energy reaching the sensor, resulting in a positive impact. On the other hand, an increase in relative humidity in a layer leads to higher emission from that layer, but a larger corresponding attenuation of the radiation coming from lower warmer layers, resulting in a reduction of radiation reaching the sensor and causing a negative impact. Given these sensitivities, many interesting physical feedback mechanisms can be studied with a full data assimilation experiment (not just the sensitivity analysis presented here) and will be the focus of future work.

a. Clear sky

The response to perturbations in temperature and vapor mixing ratio in different model layers for clear sky in terms of equivalent brightness temperature change is shown in Figs. 2 and 3. For these computations, we used the profiles from our grid that were designated as clear and averaged the responses. While Figs. 2a and 3a represent the actual sensitivity to the perturbations, the normalized results scaled by the maximum value of the sensitivity in each channel are shown in Figs. 2b and 3b. Table 2 provides a list of the actual wavelengths corresponding to the channel numbers and is useful in interpreting the results. The maximum sensitivities of the moisture-sensitive channels 10, 11, and 12 are seen to be larger than 0.02 K, while the other channels show minimal sensitivity. Channel 12 lies in a region of high water vapor absorption with the sensitivity peaking in the upper troposphere (8–10 km) and becoming saturated below 4 km. The sensitivity of channel 11 shows peaks between 5 and 8 km with the variation being a representation of the variability in the vapor mixing ratio profile. Channel 10 lies in a weaker part of the water vapor absorption band and therefore the peak sensitivity lies lower, at around 4 km. As a comparison, channel 6, used in the retrieval of total precipitable water, shows much less sensitivity compared with these 3 midwave moisture-sensitive channels. It is interesting to note that channels 10, 11, and 12 are also highly sensitive to temperature perturbations. The weighting functions computed for the U.S. Standard Atmosphere, 1976, in clear sky are shown in Fig. 4 (after Menzel et al. 1998) for comparison purposes. The normalized sensitivities in Figs. 2b and 3b have peaks that are around the same levels as the weighting functions. This is to be expected as the weighting functions represent the relative impact of change at a particular altitude on the top-of-atmosphere measured brightness temperatures. Of note is that every channel, except channels 2 and 17, shows distinct peaks located between 1- and 9-km altitude.

b. Ice clouds

With the clear-sky case in the previous section being the standard reference case, we now investigate the sensitivities for cases in our model domain that contain cirriform ice clouds, varying in thickness throughout the domain but lying between 6 and 10 km. Figures 5a–h show the responses to temperature perturbations at different heights as a function of increasing cloud optical depth plotted on a natural logarithm scale. Figures 6a–h, on the other hand, show the responses to changes in vapor mixing ratio. From our previous discussion, it is important to recall that only channels 6, 10, 11, and 12 are significantly sensitive to moisture.

1) Sensitivity to temperature variation

We observe in Fig. 5 that the response to positive temperature perturbations is a positive change in brightness temperature, consistent with the clear-sky results. The shortwave temperature channels (channels 13, 14, 15, and 16) with wavelengths around 4 μm exhibit sensitivities similar to the longwave channels (channels 8, 10, 11, and 12) and are not shown. The results from channel 17 (not shown) show low sensitivity to perturbations of temperature and water vapor for the whole profile including areas lying within the cloud until the cloud reaches an optical depth above 3. The low sensitivity in the cloud-free area is to be expected, as channel 17 is traditionally used to retrieve surface temperatures. What is important though is the low sensitivity in the cloudy areas implying that the channel can be used even in cloud scenarios for surface temperature determination.

The results for the longwave temperature sounding channels (channels 2–5) shown in Figs. 5a–d show significant sensitivity below the cirrus clouds for optical depths below 3 (corresponding approximately to 1.1 on the x axis). For smaller optical depths, the sensitivity profiles resemble those obtained from the clear-sky case for the corresponding channel. For optical depths larger than 3 (optically thick clouds), the sensitivities expectedly are located within cloud layers with the maxima around the cloud tops, and no information from atmospheric layers below the cloud can be obtained. These results indicate that channels 2–5 contain information about the temperature profile below cirrus clouds and can be used for data assimilation in cloudy regions.

Channel 6 is typically used for the precipitable water retrieval but has peak sensitivity to temperature perturbations at about a height of 2 km even in the presence of thin cirrus clouds (Fig. 5e). This result suggests that the inclusion of channel 6 for data assimilation has the potential to add resolution in the lower troposphere to vertical temperature retrievals.

Channels 11 and 12 are classified as sensitive to mid- to upper-tropospheric water vapor and are incapable of providing any significant information below even thin cirrus clouds (Figs. 5g,h). On the other hand, channel 10, designated as sensitive to the lower troposphere, shows sensitivity to temperature perturbations below thin cirrus layers (Fig. 5f), implying its possible use for temperature retrieval in data assimilation.

2) Sensitivity to vapor mixing ratio variation

We provided a theoretical explanation behind the negative response of brightness temperature to positive perturbations of vapor mixing ratio at the beginning of section 4. Figure 6 shows that we indeed get negative changes in brightness temperature in response to positive perturbations of vapor mixing ratio for all vertical levels. The information content of the temperature-sensitive channels in Table 2 (channels 2–6; shown as Figs. 6a–d) with respect to change in the vapor mixing ratio is low, as expected. The other temperature-sensitive channels (channels 13–17) with wavelengths shorter than 5 μm display low sensitivities as well (not shown). Additionally, channel 6, one of the water vapor bands, exhibits moderate sensitivity. All of the sensitivities reported above are expected from the clear-sky results in Fig. 3a, which show that only channels 10, 11, and 12 have significant sensitivity to changes in water vapor. The clear-sky peak sensitivities in channels 10, 11, and 12 also show up in the presence of clouds, as seen in Figs. 6f–h. Channel 10 is probably the most important of the 3 channels for data assimilation, as it is sensitive to changes in water vapor in the lower troposphere.

7. Summary

Information content was defined as the response to changes in top-of-atmosphere brightness arising from perturbations in temperature or water vapor in a particular atmospheric layer for a particular wavelength. Information content in the GOES sounder channels was analyzed using VISIROO to assess the potential for its use in data assimilation under both clear and cloudy conditions. Channels 2–6, 10, and 11–17 of the GOES sounder were used for the analysis, as they are designed for temperature and moisture sounding. The sensitivities at each vertical level were computed as a response to perturbations in temperature and water vapor mixing ratio using the definitions in Eqs. (3a) and (3b). The temperature perturbation for a model layer was taken to be 1 K while the vapor mixing ratio was perturbed by 10% of the layer average for the whole grid. The information content analysis shows the following:

  1. In clear-sky conditions, the shape of the vertical distribution of the sensitivity to temperature perturbations is similar to the weighting function for the corresponding channel. Channels 10, 11, and 12, the moisture-sensitive channels, additionally display significant sensitivity to temperature changes. Again, only these three channels have a significant response to vapor mixing ratio perturbations.
  2. Sensitivity profiles of temperature perturbations in the presence of thin cirrus are shaped similar to those observed for clear sky. The moisture-sensitive channels also show significant sensitivity to temperature perturbations in the presence of thin cirrus. Therefore, results from channel 6, a weakly absorbing water vapor channel, indicate its potential use for temperature sounding in the lower troposphere for both clear-sky and thin cirrus cases.
  3. As expected, only the water vapor–sensitive channels 10, 11, and 12 show any significant sensitivity to water vapor perturbations in the presence of thin cirrus. Channel 10, with a weighting function peak in the lower troposphere, provides the most important result, as it can sense variations below the cloud. The sensitivity below the cloud declines with an increase in cloud optical depth.

These results suggest that the temperature and moisture-sensitive channels of the infrared sounder are useful for the data assimilation not only in clear sky but also in the cloudy regions. Factors such as the model, background errors, observation errors, and characteristics of the given atmospheric state will influence final assimilation results; the quantitative impact of sounding observations on data assimilation cannot be determined using information content analysis alone. However, this study shows that sounding observations do have the potential to improve NWP not only in clear-sky situations where they are currently used, but also in cases with optical thin high clouds. Therefore, our results provide encouragement for assimilating satellite sounder observations under all weather conditions.

Acknowledgments

This work was supported by the Department of Defense Center for Geosciences/Atmospheric Research at Colorado State University under Cooperative Agreement DAAD19-02-2-0005 with the Army Research Laboratory. The Japanese Ministry of Education, Culture, Sports, Science and Technology is also acknowledged, which provided funding for Ms. Koyama’s research training at Colorado State University.

REFERENCES

  • Cotton, W. R., and Coauthors, 2003: RAMS 2001: Current status and future directions. Meteor. Atmos. Phys., 82 , 529.

  • Deeter, M., , and K. F. Evans, 1998: A hybrid Eddington-single scattering radiative transfer model for computing radiances from thermally emitting atmospheres. J. Quant. Spectrosc. Radiat. Transfer, 60 , 635648.

    • Search Google Scholar
    • Export Citation
  • Derber, J. C., , and W. Wan-Shu, 1998: The use of TOVS cloud-cleared radiances in the NCEPSSI analysis system. Mon. Wea. Rev., 126 , 22872299.

    • Search Google Scholar
    • Export Citation
  • Errico, R. M., 1997: What is an adjoint model? Bull. Amer. Meteor. Soc., 78 , 25772591.

  • Errico, R. M., , and T. Vukicevic, 1992: Sensitivity analysis using an adjoint of the PSU–NCAR Mesoscale Model. Mon. Wea. Rev., 120 , 16441660.

    • Search Google Scholar
    • Export Citation
  • Evans, K. F., 1998: The spherical harmonics discrete ordinate method for three-dimensional atmospheric radiative transfer. J. Atmos. Sci., 55 , 429446.

    • Search Google Scholar
    • Export Citation
  • Greenwald, T. J., , R. Hertenstein, , and T. Vukicevic, 2002: An all-weather observational operator for radiance data assimilation with mesoscale forecast models. Mon. Wea. Rev., 130 , 18821896.

    • Search Google Scholar
    • Export Citation
  • Greenwald, T. J., , T. Vukicevic, , L. D. Grasso, , and T. H. Vonder Haar, 2004: Adjoint sensitivity analysis of an observational operator for visible and infrared cloudy-sky radiance assimilation. Quart. J. Roy. Meteor. Soc., 130 , 685705.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., 2003: Atmospheric Modeling, Data Assimilation and Predictability. Cambridge University Press, 341 pp.

  • Kopken, C., , G. Kelly, , and J-N. Thepaut, 2004: Assimilation of Meteosat radiance data within the 4D-Var system at ECMWF: Assimilation experiments and forecast impact. Quart. J. Roy. Meteor. Soc., 130 , 22772292.

    • Search Google Scholar
    • Export Citation
  • McMillin, L. M., , L. J. Crone, , M. D. Goldberg, , and T. J. Kleespies, 1995: Atmospheric transmittance of an absorbing gas. 4. OPTRAN: A computationally fast and accurate transmittance model for absorbing gases with fixed and variable mixing ratios at variable viewing angles. Appl. Opt., 34 , 62696274.

    • Search Google Scholar
    • Export Citation
  • Menzel, W. P., , and J. F. W. Purdom, 1994: Introducing GOES-I: The first of a new generation of geostationary operational environmental satellites. Bull. Amer. Meteor. Soc., 75 , 757781.

    • Search Google Scholar
    • Export Citation
  • Menzel, W. P., , F. C. Holt, , T. J. Schmit, , R. M. Aune, , A. J. Schreiner, , G. S. Wade, , and D. G. Gray, 1998: Application of GOES-8/9 soundings to weather forecasting and nowcasting. Bull. Amer. Meteor. Soc., 79 , 20592077.

    • Search Google Scholar
    • Export Citation
  • Meyers, M. P., , R. L. Walko, , J. Y. Harrington, , and W. R. Cotton, 1997: New RAMS cloud microphysics parameterization. Part II: The two-moment scheme. Atmos. Res., 45 , 339.

    • Search Google Scholar
    • Export Citation
  • Stokes, G. M., , and S. E. Schwartz, 1994: The Atmospheric Radiation Measurement (ARM) Program: Programmatic background and design of the cloud and radiation test bed. Bull. Amer. Meteor. Soc., 75 , 12011221.

    • Search Google Scholar
    • Export Citation
  • van de Hulst, H. C., 1957: Light Scattering by Small Particles. Dover, 470 pp.

  • Vukicevic, T., , T. Greenwald, , M. Zupanski, , D. Zupanski, , T. Vonder Haar, , and A. S. Jones, 2004: Mesoscale cloud state estimation from visible and infrared satellite radiances. Mon. Wea. Rev., 132 , 30663077.

    • Search Google Scholar
    • Export Citation
  • Vukicevic, T., , M. Sengupta, , A. S. Jones, , and T. Vonder Haar, 2006: Cloud-resolving satellite data assimilation: Information content of IR window observations and uncertainties in estimation. J. Atmos. Sci., 63 , 901919.

    • Search Google Scholar
    • Export Citation
  • Walko, R. L., , W. R. Cotton, , M. P. Meyers, , and J. Y. Harrington, 1995: New RAMS cloud microphysics parameterization. Part I: The single moment scheme. Atmos. Res., 38 , 2962.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Schematic of the VISIROO [after Vukicevic et al. (2004)].

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3254.1

Fig. 2.
Fig. 2.

Vertical profiles of (a) sensitivity to perturbations of temperature and (b) the normalized results.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3254.1

Fig. 3.
Fig. 3.

Vertical profiles of (a) sensitivity to perturbations of humidity and (b) the normalized results.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3254.1

Fig. 4.
Fig. 4.

Weighting functions for GOES sounder: (a) longwave, (b) midwave, and (c) shortwave.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3254.1

Fig. 5.
Fig. 5.

Sensitivity to perturbations of temperature of channels (a) 2, (b) 3, (c) 4, (d) 5, (e) 6, (f) 10, (g) 11, and (h) 12.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3254.1

Fig. 6.
Fig. 6.

Same as in Fig. 5 but for humidity.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3254.1

Table 1.

GOES imager channels [after Menzel and Purdom (1994); Greenwald et al. (2004)].

Table 1.
Table 2.

GOES sounder channels [after Menzel and Purdom (1994)].

Table 2.
Save