Resumen
The characterization of temporal and spatial patterns in phosphorus (P) loading in Lake Erie is essential in order to continue monitoring the excessive P condition that comes from the western Lake Erie Basin. This study aims to perform such a characterization using the continuous wavelet transform (CWT) and the discrete wavelet transform (DWT) methods. These wavelet transformations were conducted on streamflow data, TP loads, and soluble reactive phosphorus (SRP) of six stations located near Lake Erie of Northern Ohio. These stations are located near the outlet of Cuyahoga, Grand, Maumee, Vermilion, Raisin, and Sandusky watersheds. Long-term continuous P loading data, in which some dated back to 1970, were used in the analysis. The results obtained from the CWT and DWT approaches were found to complement each other. Streamflow had significant mixed variability at 1, 2, and 4 years. The variability for SRP was limited to 1 and 2 years while the TP variability was only seen at the 1-year scale. It was interesting to find that strong temporal patterns of SRP were observed in most of the watersheds only after the mid-1990s. The CWT wavelet spectra also reflected the land use characteristics of each watershed. For example, the wavelet spectra of surface runoff and TP for the agricultural watersheds (i.e., Raisin, Sandusky, and Maumee Rivers) were similar and characterized by significant variability primarily at the annual scale and at the two to four-year periodicities. The urbanized watershed (i.e., Cuyahoga River) did not show any association between either phosphorus (TP or SRP) with surface runoff and baseflow, which indicates that P in the urbanized watershed was not driven by the flow.