A significant amount of further work is required to reach meaningful conclusions about the effects of plants on wave transmission. The study I did could certainly be improved with longer data sets and more data points, in the hope of confirming the preliminary conclusions from the limited data I was able to collect. There are, however, things that would not be accounted for even if this study were carried out more rigorously:

1) Elasticity. The modulus of elasticity of the modeled plants would need to be compared to that of real plants, in order to determine the validity of the results. It seems obvious that  my modeled plants are considerable more rigid than real ones; it is less clear what effect this would have on the results. 

2) Roughness: Like the elasticity, it is clear that roughness values would be quite different for real plants than they are for PVC pipe. I could imagine this change being quite significant. 

3) Plant types: My models were based on roughly cylindrical plants, such as cattails; the natural environment contains many different kinds of plants which would have to be identified and dealt with individually. If one were to establish a study area, the types could be narrowed.

4) Spacing and layout: My model was built on a regular grid pattern, this is not at all like natural plantings. Particularly with periodic phenomena like waves, such patterns can yield surprising  results. It would be best to develop a protocol for randomization that was based on natural plantings.

5) Dimensional analysis: In order to be truly useful in a real world setting, the study would need to be designed around a dimensional analysis, accounting for elasticity, roughness, density, plant morphology, interaction length, stem diameter, wave parameters (length, height) and fluid characteristics. This would allow the study to give attenuation estimates for particular shorelines, or for proposed plantings.