TOWARDS A NEW CLASSIFICATION OF RIVERS BASED UPON GENERIC STAGE-DISCHARGE RATING CURVES

The objective of this study is to explore the possibility of classifying rivers based upon their stage-discharge rating curves. The current practice is to create a rating curve for each stream site by carrying out thirty or more simultaneous measurements of stream stage and stream discharge. However, if a river could be assigned a generic rating curve with a small number of adjustable parameters, the cost and manpower of developing rating curves could be vastly reduced. The motivation for this study arose from Utah Valley University’s long-term program of groundwater development in the Trans-Mexican Volcanic Belt, which requires hydrographs for rivers near remote mountain villages. Previous research showed that the generic rating curve for rivers identified as bedrock step-pool rivers, which are common in the Trans-Mexican Volcanic Belt, was only a marginal improvement over the generic rating curve for rivers chosen randomly. This study is taking the reverse approach of first identifying classes of generic rating curves and then identifying the types of rivers that share those generic rating curves. Generic rating curves are being developed by importing the USGS National Water Information System database of about three million simultaneous measurements of stream stage and stream discharge at about 15,000 active and historic stream gaging sites into a custom Python-driven data manipulation script. As a first step, linear relationships between the z-score of the logarithm of stage and the z-score of the logarithm of discharge are being developed for each site. Frequency spectra of the slopes and intercepts of the linear relationships are being created by summing normal distributions, in which each site is represented by a normal distribution with mean equal to slope (or intercept) and standard deviation equal to the uncertainty in the slope (or intercept). Spectral peaks are being identified statistically so that rivers are classified by their proximity to a spectral peak in slope and/or intercept. The common geomorphological characteristics of rivers with a common spectral peak are being considered. The process is being repeated with increasingly complex functional forms of rating curves. Results will be reported at the meeting.