Transfer of flood information through multiple linear regression.

Abstract

Maximum annual flow records (floods) are used for hydrologic dimensionality of hydraulic structures designed for protection and crossing. Logically, the longer the available series, the more reliable their probabilistic estimates or predictions. Thus, this work describes in detail the procedure for flood information transfer through multiple linear regression. Short records are enlarged based on longer nearby series, examining the statistical advantage of said transfer. The mathematical formulation is presented in a simple way using a matrix solution and a numerical example is developed to enlarge the series at the Platón Sánchez hydrometric station on the Tempoal River in the state of Veracruz, using several nearby records. Lastly, the conclusions are formulated, which point out the simplicity of the procedure and suggest its systematic application.