Laserfiche WebLink
<br />OD1353 <br /> <br />24 <br /> <br />Measures of fit - In the published studies, the various models were <br />usually qualitatively analyzed. If the simulated hydrograph "looked" <br /> <br />1 ike the observed hydrograph then the model was said to predict the <br /> <br />real world event "fairly well." No precise measure was used to <br /> <br />establ ish "goodness of fit." In this examination, the writers used <br /> <br />hydrograph "goodness of fit" measures suggested by Sarma (79) and <br />modified by McCuen (67), together with the peak discharge measure <br />presented by Marsalek (64). These measures of predictive capability <br /> <br />are: <br /> <br />1) The Ratio (QpRATIO) of Simulated Peak Runoff Rate (Qps) to <br />the Observed Peak Runoff Rate (Qpo). This ratio has the following <br />properties: <br />a) If Qp RATIO is greater than 1.0, the model overpredicts <br />the peak runoff rate. <br />b) If Qp RATIO equals 1.0, the model accurately predicts <br />the peak runoff rate. <br />c) If QpRATIO is less than 1.0, the model underpredicts <br />the peak runoff rate. <br />2) The Modified Correlation Coefficient (RMOD) as defined by <br />McCuen (67). This coefficient is the linear correlation coefficient <br />between the simulated and the observed hydrograph (as described by <br />Sarma) adjusted for hydrograph size. It is defined by: <br /> <br />RMOD = (af) x (R) <br /> <br />where af is the adjustment factor defined by: <br />