Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Prof. Wayne K. Hocking


This thesis focuses on the investigation of the reliability of turbulence measurements by radars when the spectral-width method is used. This method employs the spectral widths observed by radar (experimental spectral width) to determine turbulence. However, the experimental width can be affected by non-turbulent effects, including radar geometry. Therefore, the spectral width due to non-turbulent effects (theoretical spectral width) must be removed from the experimental width. This can occasionally lead to negative values of turbulence.

It is our aim (1) to study the effects of both experimental and theoretical spectral widths on the accuracy of turbulence measurements, (2) to study the validity of “negative” values of turbulence, and (3) to compare radar-estimated values of turbulence with in-situ measurements.

This is performed by studying the factors that can contribute to errors in estimation of spectral widths, including mean wind speed, wind shear, anisotropy and length of data. In addition to that, radar-estimated turbulence is compared to turbulence measured by high-resolution research and commercial aircraft.

Significant findings include:

1. The statistical fluctuations of mean wind speed and its impact on estimation of the theoretical spectral width is the most important factor in producing errors in turbulence measurements.

2. The choice of spectrum fitting algorithm and length of data is very important in producing experimental spectral widths.

3. In order to estimate theoretical widths more accurately, a formula is developed by comparing models that calculate the theoretical spectral width.

4. It is found that both negative and positive values of turbulence need to be included in turbulence analysis. However, if the percentage of negatives exceeds 35%, the measurements are not reliable.

5. Turbulence data measured by radar agrees well with high-resolution aircraft data for weak turbulence. However, in-situ aircraft measurements show a higher probability of strong turbulence than hourly radar data.

6. The theoretical spectral width is small compared to the experimental one for strong turbulence. By discarding the theoretical spectral widths in turbulence calculations, we can measure near-instantaneous values of turbulence rather than hourly averages. Therefore, we can improve the agreement between the aircraft-estimated and radar-estimated turbulence for strong values of turbulence.

Included in

Physics Commons