Laserfiche WebLink
<br />Chapter II <br /> <br />Development of a Physically Based Distributed <br />Parameter Rainfall Runoff Model <br /> <br />This algorithm adjusts (either up or down) the value at nearby sites using the ratio of monthly <br />means. The distance factor in the equation results in stations closer to the site of interest being <br />weighted higher. When one or more values for the five nearest sites are also missing, the terms in <br />the equation for these sites are left out. This algorithm was also used to synthesize missing <br />minimum or maximum daily temperatures in the Gunnison Model. <br /> <br />MODEL CALIBRA nON <br /> <br />In the initial phases of Gunnison Model development, separate PRMS models were developed for <br />many of the subbasins in the Upper Gunnison River Basin. Distinct PRMS models were created <br />for Taylor Park (nine HRU's), East (nine HRU's), Quartz (nine HRU's), Upper Tomichi (nine <br />HRU's), Curecanti (six HRU's), Soap (six HRU's) and Lake Fork (nine HRU's) Basins. This was <br />done because, at the time, the standard version of PRMS allowed for only one outflow point per <br />model and did not support routing of water between nodes in the system. <br /> <br />Calibration of these distinct models was performed using the following general sequence. <br /> <br />First, model simulations were performed over a 38-year period (1952-89), and the major <br />components in the water balance were studied. These major components include precipitation, <br />evapotranspiration, runoff, and change in storage in the PRMS conceptual reservoirs. Results of <br />these initial simulations revealed discrepancies between the simulated volumes and virgin volumes <br />over the 38-year period. To rectify these differences, precipitation correction factors (the PRMS <br />parameter which accounts for the difference between observed precipitation at a climate recording <br />station and the amount of precipitation that the model distributes to model HRU's) were adjusted. <br />The amount of simulated runoff either increased or decreased as these correction factors were <br /> <br />adjusted. This adjustment was made in an iterative fashion, until simulated volumes matched <br />virgin volumes over the 38-year period6 <br /> <br />h Initial differences between simulated and virgin volumes for the 16 subbasins in the upper Gunnison River <br />Basin varied. In two of the subbasins (Taylor Purk and Quartz). the initial simulated volumes nearly matched virgin <br />volumes (less than I percenl difference). In most of the subbasins, the initial discrepancy was about 10 percent. The <br />highest initial discrepancy was in the Cebolla Basin. where initial simulated volumes had to be reduced by 17 percent. <br /> <br />20 <br />