Search Results
You are looking at 11 - 12 of 12 items for
- Author or Editor: Chong-Jian Qiu x
- Refine by Access: All Content x
Abstract
The simple adjoint method of Qiu and Xu is extended and used to retrieve the low-altitude horizontal wind field from single-Doppler radial wind data measured during the Phoenix II field experiment. Since the extended method uses only the radial momentum equation on a low-altitude horizontal plane with a weak nondivergence constraint for the horizontal winds, the pressure gradient and vertical advection are treated as an unknown residual forcing. The test results show that (i) the method can retrieve the low-altitude time-mean (or running mean) horizontal winds (averaged over a period of several sequential radar scans) from single-Doppler radial wind data; (ii) retrieving the time-mean part of the unknown residual forcing term improves the wind retrieval; (iii) the detailed data treatments and proper settings of the weights considered in Xu et at. remain useful for improving the retrieval.
Abstract
The simple adjoint method of Qiu and Xu is extended and used to retrieve the low-altitude horizontal wind field from single-Doppler radial wind data measured during the Phoenix II field experiment. Since the extended method uses only the radial momentum equation on a low-altitude horizontal plane with a weak nondivergence constraint for the horizontal winds, the pressure gradient and vertical advection are treated as an unknown residual forcing. The test results show that (i) the method can retrieve the low-altitude time-mean (or running mean) horizontal winds (averaged over a period of several sequential radar scans) from single-Doppler radial wind data; (ii) retrieving the time-mean part of the unknown residual forcing term improves the wind retrieval; (iii) the detailed data treatments and proper settings of the weights considered in Xu et at. remain useful for improving the retrieval.
Abstract
Although the simple adjoint (SA) method for retrieving the low-altitude winds from single-Doppler scans was previously examined with the Phoenix II data collected for rather uniform wind fields on nonstorm days with aluminum chaff dispensed from an aircraft (to enhance the reflectivity), the method has not been tested with storm data for complex flow fields. To examine this problem, the SA method is further developed and tested with the Denver airport microburst data through a series of numerical experiments. In addition to the earlier upgrading of the SA method, three new objectives are fulfilled to improve the retrievals. In particular, it is found that by imposing a weak vorticity constraint, by using the previous time-level retrieval as the initial guess of the retrieval at the current time level, and finally by incorporating the surface anemometer data into the method, the averaged rms difference between the retrieved (from FL-2 radar) and dual-Doppler observed winds reduces from 3.87 to 2.89 m s−1, then to 2.66 m s−1, and finally to 2.55 m s−1, while the averaged correlation coefficient increases from 88% to 94%, then to 95%, and finally to 96%. The surface anemometer data can make a significant improvement, especially when the retrieval purely based on radar data is relatively poor. Factors responsible for the retrieval error are, also examined and discussed with physical interpretations.
Abstract
Although the simple adjoint (SA) method for retrieving the low-altitude winds from single-Doppler scans was previously examined with the Phoenix II data collected for rather uniform wind fields on nonstorm days with aluminum chaff dispensed from an aircraft (to enhance the reflectivity), the method has not been tested with storm data for complex flow fields. To examine this problem, the SA method is further developed and tested with the Denver airport microburst data through a series of numerical experiments. In addition to the earlier upgrading of the SA method, three new objectives are fulfilled to improve the retrievals. In particular, it is found that by imposing a weak vorticity constraint, by using the previous time-level retrieval as the initial guess of the retrieval at the current time level, and finally by incorporating the surface anemometer data into the method, the averaged rms difference between the retrieved (from FL-2 radar) and dual-Doppler observed winds reduces from 3.87 to 2.89 m s−1, then to 2.66 m s−1, and finally to 2.55 m s−1, while the averaged correlation coefficient increases from 88% to 94%, then to 95%, and finally to 96%. The surface anemometer data can make a significant improvement, especially when the retrieval purely based on radar data is relatively poor. Factors responsible for the retrieval error are, also examined and discussed with physical interpretations.