Start Date
5-2020 12:00 AM
Description
The Rayleigh-scatter lidar (RSL) system at the Atmospheric Lidar Observatory at Utah State University (ALO-USU) provided a rich database of absolute temperatures throughout the mesosphere from 45 km to above 90 km between 1993 and 2004. Recently, a new method for retrieving absolute temperatures from RSL observations has been developed by a group at the University of Western Ontario (UWO), Canada. The Optimal Estimation Method (OEM) uses machine learning to minimize a cost function by optimizing the temperature parameter in a forward model, in our case the lidar equation, to RSL data. This optimization provides some benefits over the existing method through a robust uncertainty budget and a quantitative determination of the cut-off altitude, or the topmost altitude in the temperature profile. Using this method also provides a slight increase in the top observable altitude and does not have a large dependence on the initial temperature. The OEM procedure was converted from MATLAB, which is used by the UWO group, into Python, which is used at ALO-USU. The temperatures were then reduced using the OEM from observations made between 1993 and 2004. Initial results obtained using the Python version of OEM were compared with those using MATLAB showing good agreement. More observations from ALO-USU were then reduced using OEM and compared with the original reduction method. The results show good agreement between the two methods until higher altitudes. These differences can be attributed to dependence on initial conditions in the original method or over-constraining from overestimating the altitude range to be used in the OEM retrieval. At higher altitudes, however, the temperatures tend to agree within the given uncertainties. Further work with this method is being done to generate a temperature climatology using ALO-USU observations and developing a method to retrieve absolute neutral densities using a modification of the forward model in the OEM.
Included in
Applying the Optimal Estimation Method for Retrieving Rayleigh-Scatter Lidar Temperatures in the Mesosphere
The Rayleigh-scatter lidar (RSL) system at the Atmospheric Lidar Observatory at Utah State University (ALO-USU) provided a rich database of absolute temperatures throughout the mesosphere from 45 km to above 90 km between 1993 and 2004. Recently, a new method for retrieving absolute temperatures from RSL observations has been developed by a group at the University of Western Ontario (UWO), Canada. The Optimal Estimation Method (OEM) uses machine learning to minimize a cost function by optimizing the temperature parameter in a forward model, in our case the lidar equation, to RSL data. This optimization provides some benefits over the existing method through a robust uncertainty budget and a quantitative determination of the cut-off altitude, or the topmost altitude in the temperature profile. Using this method also provides a slight increase in the top observable altitude and does not have a large dependence on the initial temperature. The OEM procedure was converted from MATLAB, which is used by the UWO group, into Python, which is used at ALO-USU. The temperatures were then reduced using the OEM from observations made between 1993 and 2004. Initial results obtained using the Python version of OEM were compared with those using MATLAB showing good agreement. More observations from ALO-USU were then reduced using OEM and compared with the original reduction method. The results show good agreement between the two methods until higher altitudes. These differences can be attributed to dependence on initial conditions in the original method or over-constraining from overestimating the altitude range to be used in the OEM retrieval. At higher altitudes, however, the temperatures tend to agree within the given uncertainties. Further work with this method is being done to generate a temperature climatology using ALO-USU observations and developing a method to retrieve absolute neutral densities using a modification of the forward model in the OEM.
Comments
Due to COVID-19, the Symposium was not able to be held this year. However, papers and posters were still submitted.