No abstract

Basin scale seiches in lakes are important elements of the total energy budget and are a driver of fluxes of important ecological parameters, such as oxygen, nutrients, and sediments. At present, the extraction of the damping ratios of surface seiches, which are directly related to the capacity of seiches to drive these fluxes through the increased mixing of the water column, is reliant on spectral analysis which may be heavily influenced by the transformation of water level records from the time domain to the frequency domain, and which are sensitive to the level of noise present within the data. Existing spectral-based methods struggle to extract the periods of surface seiches which are of similar magnitude due to the overlap between their spectral responses. In this study, the principles of operational modal analysis, through the random decrement technique (RDT), currently used primarily in the analysis of high rise structures and in the aeronautical industry and not previously applied within the fields of limnology or ecology, are applied to barotropic seiches through the analysis of water level data for Lake Geneva, Switzerland, and Lake Tahoe, USA. Using this technique, the autocorrelation of the measurements is estimated using the RDT and modal analysis can then be carried out on this time-domain signal to estimate periods of the dominant surface seiches and the corresponding damping ratios. The estimated periods show good agreement with experimental results obtained through conventional spectral techniques and consistent damping ratios are obtained for the dominant surface seiche of Lake Tahoe. The effect of input parameters is discussed, using data for the two lakes, alongside discussion of the application of RDT to the study of internal seiches and current barriers to its application. RDT has great potential for the analysis of both surface and internal seiches, offering a method through which accurate damping ratios of seiche oscillations may be obtained using readily available data without necessitating spectral analysis.

No abstract

No abstract