Maximizing the Information Content of Ill-Posed Space-Based Measurements Using Deterministic Inverse Method.

Area of Science:

• Atmospheric remote sensing
• Inverse methods in geosciences
• Hyperspectral data analysis

Background:

• Stochastic approaches dominate operational retrievals from hyperspectral infrared sounders, often leading to ambiguities.
• A key drawback of stochastic methods is their reliance on treating error as definitive information.
• Deterministic inverse methods offer a path to unambiguous retrievals.

Purpose of the Study:

• To apply and evaluate the regularized total least squares (RTLS) deterministic inverse method for atmospheric profile and surface temperature retrievals.
• To compare the performance of RTLS against the stochastic optimal estimation method (OEM) using simulated and real-world data.
• To introduce and utilize novel concepts for in-depth analysis of ill-posed inversions and information content.

Main Methods:

• Application of regularized total least squares (RTLS) for simultaneous ozone (O3) profile and surface temperature (ST) retrieval.
• Comparative assessment of RTLS and optimal estimation method (OEM) under identical simulation settings for Cross-track Infrared Sounder (CrIS) and Tropospheric Emission Spectrometer (TES) data.
• Utilizing ozonesonde profile data for validation and incorporating sub-space analysis for information content analysis.

Main Results:

• RTLS demonstrates consistently unambiguous retrievals, overcoming limitations of the OEM.
• OEM shows a loss of information compared to a priori knowledge after incorporating measurements.
• RTLS achieves a deterministic "information gain" of approximately 40-50% from the same dataset.

Conclusions:

• Regularized total least squares (RTLS) provides a more robust and informative approach for hyperspectral data retrieval compared to traditional optimal estimation methods (OEM).
• The study highlights the importance of deterministic methods for overcoming ambiguities and maximizing information extraction in atmospheric remote sensing.
• RTLS offers a significant advancement in retrieving atmospheric variables like ozone and surface temperature with enhanced accuracy and clarity.