Aerodynamic roughness lengths over movable beds: Comparison of wind tunnel and field data

Abstract

We evaluate the increase in apparent roughness length during aeolian saltation. Using a data set comprising 291 wind profiles measured in wind tunnels or in the field, we tested the Charnock (1955), modified Charnock (Sherman, 1992), and Raupach (1991) relationships for the entire data set and for segregated wind tunnel and field data sets. We assessed the abilities of the models to predict enhanced roughness and investigated scaling differences between wind tunnels and field environments. For the entire data set, the Charnock relationship worked best, with an r2 of 0.25. For the segregated data sets, the Charnock relationship also worked best, with an average r2 of 0.675, as opposed to 0.66 for the modified Charnock and 0.61 for the Raupach model. For conditions with shear velocity ≤1.0 ms−1, the comparable, average values of r2 are 0.54, 0.625, and 0.645, respectively. There was an order of magnitude difference in the roughness length responses for wind tunnel and field data for each model evaluated, convincing evidence of a fundamental scaling constraint in the wind tunnel data. The scaling differences are attributed to conceptual and bias errors that include the use of the law of the wall versus the wake‐defect law or differences in turbulence characteristics, especially regarding the role of large coherent structures. Raupach's (1991) A was found to be an extremely volatile constant, making the modified Charnock relationship most suitable for predicting aerodynamic roughness lengths over movable sand beds.