Laserfiche WebLink
streamflow gage located in the SPDSS study area as a potential independent gage. <br />Include the historical period of record for each gage, with a maximum of 100 years per <br />gage (the MSM program limit). <br />2. Use the SEP selection criterion, allowing MSM to develop annual and monthly <br />regressions with multiple independent gages for each dependent gage. Review the <br />independent gages selected by MSM for reasonableness in terms of similar period of <br />record, location, and effects of storage or large diversions on the dependent gage. <br />3. Rerun MSM using dependent gage-specific input files, excluding any "unreasonable" <br />gages from those that MSM identified in the initial run. <br />4. As a result of MSM treating each gage as both an independent and dependent gage, the <br />MSM output file contains filled data sets for every gage in the input files (including the <br />independent gages with any missing historical data filled). While the extraneous data <br />will not affect future analyses, the user may wish to bring the MSM output files back into <br />TSTooI and edit the output to retain only the gages that were intentionally being filled as <br />dependent gages. <br />Results <br />Table 1 provides a statistical comparison of regression results using TSTooI and MSM to fill <br />missing data for three representative South Platte streamflow gages. The MSM correlation <br />coefficients (r) are provided in Table 1 for comparison with the TSTooI results. Through <br />regressing with multiple independent gages, MSM generally produces better (higher) ranges of <br />monthly and annual correlation coefficients. <br />Figure 1 shows a comparison of the filled results for the Big Dry Creek at Mouth Near Fort <br />Lupton (06720990) gage, using both MSM and TSTooI (the graph provided for this gage <br />illustrates the general procedure used for all three representative gages). After running MSM <br />with all of the SPDSS study area streamflow gages and then limiting the input file to those <br />determined to be reasonable as described above, MSM selected the South Platte River at Denver <br />(06714000), Tarryall Creek Below Rock Creek Near Jefferson (06699005), and Bear Creek at <br />Morrison (06710500) gages. The South Platte River at Denver and Bear Creek at Morrison <br />gages were also run with TSTooI, using all four combinations of monthly and annual correlations <br />with linear and logarithmic transformations. The combinations producing the best (highest) <br />correlation coefficients are shown in Figure 1. Tarryall Creek Below Rock Creek Near Jefferson <br />was not run with TSTooI because it only has nine years of data within the Big Dry Creek gage <br />missing period (1950 - 1991) while the other gages are complete. The TSTooI results using the <br />Bear Creek at Morrison gage with a logarithmic data transformation look most similar to the <br />MSM results, however, the linear monthly regression produced the best correlation. The South <br />Platte River at Fort Lupton (06710500) gage was also tested with TSTooI because it is the closest <br />streamflow gage in proximity. Winter flows for the filled period were significantly higher than <br />the historical period of record or those produced using the other streamflow gages. Therefore, <br />the Bear Creek at Morrison gage was determined to be the best option using TSTooI. <br />Task2.doc 4 of 10 <br />