Input data measurement-induced uncertainty in watershed modelling

Abstract

This study presents a new methodology for estimation of input data measurement-induced uncertainty in simulated dissolved oxygen (DO) and nitrate-nitrogen (NO 3-N) concentrations using the Hydrological Simulation Program-FORTRAN (HSPF) model and data from the Amite River, USA. Simulation results show that: (1) a multiplying factor of 1.3 can be used to describe the maximum error in temperature measurements; similarly, a multiplying factor of 1.9 was estimated to accommodate the maximum of ±5% error in rainfall measurements; (2) the uncertainty in simulated DO concentration due to positive temperature measurement errors can be described with a normal distribution, N(0.062, 0.567); (3) the uncertainty in simulated NO 3-N concentration due to rainfall measurement errors follows a generalized extreme value distribution; and (4) the probability density functions can be utilized to determine the measurement-induced uncertainty in simulated DO and NO 3-N concentrations according to the risk level acceptable in water quality management. © 2012 Copyright 2011 IAHS Press.