CLIMCAPS Optimal Estimation retrieval

CLIMCAPS retrieves multiple Earth surface, cloud and atmospheric state variables from infrared and microwave measurements using Optimal Estimation inversion (Rodgers, 2000). We describe CLIMCAPS in detail elsewhere, with specific reference to uncertainty (Smith and Barnet, 2019) and information content (Smith and Barnet, 2020). Figure 1 outlines the main retrieval steps. Each box with a variable name represents an Optimal Estimation retrieval.

Figure 1: Flow diagram of the CLIMCAPS sequential retrieval algorithm. This gives a broad overview of the main retrieval steps and their logical flow towards two final products files, CLIMCAPS retrievals (CLIMCAPS RET) and cloud cleared radiances (CLIMCAPS CCR). Note that we discuss different aspects of CLIMCAPS algorithm flow also in (Smith and Barnet, 2019, 2020). See Table 1 for a description of the acronyms and symbols used here.

Table 1: List of acronyms and symbols used in Figure 1.

 About the flowchart

The flowchart in Figure 1 outlines the main CLIMCAPS retrieval steps. We highlight only those steps that result in the primary retrieval products, namely cloud cleared radiances (CCR) and the nine profile retrievals – T(p), H₂O, O₃, CO, CO₂, CH₄, SO₂, HNO₃, N₂O.

CLIMCAPS profiles are available on 100 pressure coordinates, with T(p) represented on 100 pressure levels (air_pres) and the gas species on 100 pressure layers (air_pres_lay). For the sake of simplicity, we omit reference to those steps where (i) diagnostic metrics and quality control indices are calculated, and (ii) quantities are derived from the main retrieval variables, such as outgoing long-wave radiation (OLR) and relative humidity.

CLIMCAPS output variables are saved to two product files in netCDF format:

1. *.L2_CLIMCAPS_CCR.*: Cloud cleared radiances.
2. *.L2_CLIMCAPS_RET.*: Retrieved and derived geophysical variables. This file is organized into two tiers, with the main group of variables as tier 1 (T(p), surface and cloud variables, relative humidity, etc.) and the remaining retrieved variables, derived quantities and diagnostic metrics organized into four subgroups on tier 2. We refer to three of these subgroups in Figure 1:
   • AUX: collection of auxiliary variables, diagnostic metrics as well as minor gas detection flags.
   • MOL_LAY: collection of the eight primary trace gas profile retrievals, H₂O, O₃, CO, CO₂, CH₄, SO₂, HNO₃, N₂ With the exception of CO₂ that is retrieved as volume mixing ratio, all trace gases are retrieved as layer column densities. The averaging kernel matrices for all nine retrieval variables (eight gases and T(p)) are contained in the AVE_KERN subgroup.
   • MW: collection of MW-only retrieval variables using the algorithm developed by (Rosenkranz, 2001, 2006)

CLIMCAPS uses several diagnostic metrics to derive a single, final quality control flag (0 = best, 1 = good, 2 = bad). Some reasons for rejection include:

• The footprint is covered with precipitating clouds such that T_MW(p) fails
• The retrieved cloud fraction exceeds 80%
• Large differences exist between the clear radiance estimate (from the MW-only step) and the cloud clear estimate (from the cloud clearing step)
• The RMS of the observed minus calculated brightness temperature exceeds 1.75 K for specific window channels.
• The boundary layer estimate of dT(p)*** > 1.5 K

None of the above checks alter the retrieved values, but instead are used to derive a final quality flag. CLIMCAPS, therefore, does not have quality control flags tailored for each individual retrieved variable, but instead applies the logic that if T(p) and H₂O fail, all subsequent trace gas retrievals also fail.

CLIMCAPS and Microwave

CLIMCAPS does not use any of the MW channels in cloud clearing or cloud parameter retrievals (cloud fraction and cloud top pressure).

The MW-only retrievals of liquid water path (LIQ) and emissivity (\(\varepsilon\)_MW) are used to define the background atmospheric state in subsequent IR+MW retrievals, while the MW-only temperature and water vapor retrievals are written to the output RET file as is; they are not used in subsequent IR+MW retrievals.

References

• GMAO: MERRA-2 inst3_3d_asm_Nv: 3d,3-Hourly,Instantaneous,Model-Level,Assimilation,Assimilated Meteorological Fields V5.12.4, https://doi.org/10.5067/WWQSXQ8IVFW8, 2015.

• Hook, S.: Combined ASTER and MODIS Emissivity database over Land (CAMEL) Coefficient Monthly Global 0.05Deg V002, https://doi.org/10.5067/MEASURES/LSTE/CAM5K30CF.002, 2019a.

• Hook, S.: Combined ASTER and MODIS Emissivity database over Land (CAMEL) Emissivity Monthly Global 0.05Deg V002, https://doi.org/10.5067/MEASURES/LSTE/CAM5K30EM.002, 2019b.

• Hook, S.: Combined ASTER and MODIS Emissivity database over Land (CAMEL) Uncertainty Monthly Global 0.05Deg V002, https://doi.org/10.5067/MEASURES/LSTE/CAM5K30UC.002, 2019c.

• Rodgers, C. D.: Inverse methods for atmospheric sounding: theory and practice, World Scientific, Singapore; Hackensack, N.J., 2000.

• Rosenkranz, P. W.: Retrieval of temperature and moisture profiles from AMSU-A and AMSU-B measurements, 39, 2429–2435, https://doi.org/10.1109/36.964979, 2001.

• Rosenkranz, P. W.: Cloud liquid-water profile retrieval algorithm and validation, 111, https://doi.org/10.1029/2005JD005832, 2006.

• Smith, N. and Barnet, C. D.: Uncertainty Characterization and Propagation in the Community Long-Term Infrared Microwave Combined Atmospheric Product System (CLIMCAPS), 11, 1227, https://doi.org/10.3390/rs11101227, 2019.

• Smith, N. and Barnet, C. D.: CLIMCAPS Observing Capability for Temperature, Moisture and Trace Gases from AIRS/AMSU and CrIS/ATMS, Gases/Remote Sensing/Data Processing and Information Retrieval, https://doi.org/10.5194/amt-2020-71, 2020.

• Wu, X. and Smith, W. L.: Emissivity of rough sea surface for 8–13 µm: modeling and verification, 36, 2609, https://doi.org/10.1364/AO.36.002609, 1997.