This technology uses a two-phase method to automatically correct temperature measurement errors from ground-based microwave radiometers, improving accuracy in all weather conditions without needing to modify existing retrieval algorithms, and can be extended to other remote sensing data.
Background: Ground-based Microwave Radiometers (MWRs) are essential tools in atmospheric science, providing continuous, high-resolution temperature profiles that are critical for weather forecasting, climate monitoring, and research. These instruments are particularly valuable for filling observational gaps left by radiosondes and satellites, offering real-time data on atmospheric thermodynamics. However, the accuracy of MWR temperature retrievals is paramount, as even small systematic errors can propagate through weather models and impact forecasts, climate analyses, and decision-making processes. As the demand for precise, reliable atmospheric data grows—especially for operational networks and research testbeds—there is an increasing need to address and minimize sources of error in MWR-derived temperature profiles. Despite their utility, MWR temperature retrievals are prone to systematic biases, most notably a persistent cold bias that can vary with weather conditions and atmospheric structures such as temperature inversions. Traditional bias correction approaches often require access to, or modification of, the retrieval algorithm or its input parameters—an impractical solution for many operational networks that rely on proprietary or closed-source retrieval software. Furthermore, existing correction methods may not adequately account for weather-dependent variations in bias, leading to residual errors, especially near inversion layers where sharp temperature gradients occur. These limitations hinder the ability to deliver consistently accurate temperature profiles, reducing the effectiveness of MWRs in both research and operational meteorology.
Technology Overview: This technology is a novel bias correction scheme designed to improve the accuracy of temperature measurements obtained from ground-based Microwave Radiometers (MWRs), such as those used in the New York State Mesonet Profiler Network. The approach operates in two main phases: a training phase and an application phase. During training, the system uses multiple linear regression to model the relationship between observed MWR brightness temperatures and systematic temperature biases, referencing high-resolution atmospheric analyses from sources like the HRRR. In the application phase, it leverages the observed brightness temperatures to estimate and remove bias directly from the MWR temperature retrieval outputs. This method is particularly effective at reducing cold bias throughout the lower and mid-troposphere, significantly lowering the standard deviation of temperature errors by 10-15%. It also adapts to changing weather conditions by using brightness temperature as a dynamic predictor and has demonstrated robust performance in real-time scenarios for up to several weeks. What sets this solution apart is its ability to apply bias correction directly to the output of temperature retrievals, rather than requiring modifications to the retrieval algorithms or their inputs. This makes it uniquely suitable for use with proprietary or unmodifiable retrieval systems, which are common in operational networks. The scheme’s dynamic, weather-adaptive correction—enabled by using brightness temperature as a predictor—addresses both persistent and weather-dependent biases, including those associated with temperature inversion layers that often challenge traditional correction methods. Its extensibility to other atmospheric variables and remote sensing platforms further enhances its value, making it a versatile tool for improving the reliability of ground-based remote sensing data in research, forecasting, and climate monitoring applications.
https://suny.technologypublisher.com/files/sites/adobestock_211250077.jpeg Photo for reference only, not a depiction of the invention.
Advantages: • Effectively removes systematic cold bias in temperature retrievals from ground-based Microwave Radiometers (MWRs). • Reduces temperature error standard deviation by 10-15%, improving measurement accuracy. • Applies bias correction directly to output, enabling use with proprietary or unmodifiable retrieval algorithms. • Accounts for weather-dependent variations in bias by using observed brightness temperatures as predictors. • Significantly reduces errors near and above temperature inversion layers, enhancing profile accuracy. • Robust for real-time application on unseen data for up to two to three weeks, supporting operational use. • Extensible to bias correction of other retrieved atmospheric variables and remote sensing observations. • Validated with independent radiosonde data and field campaigns, ensuring reliability and broad applicability.
Applications: • Weather forecasting network data correction • Climate monitoring station calibration • Operational remote sensing quality control • Atmospheric research field campaign support • Commercial MWR system enhancement
Intellectual Property Summary: Patent pending
Stage of Development: TRL 2
Licensing Status: This technology is available for licensing.