Durango RICD, Supplement to My Expert Report of December 29, 2006

Home Browse Search

Durango RICD D. Randolph Seaholm Supplemental Report February 20, 2007 • Page 2 of 4 solution to a mathematical problem of fitting a distribution to data." In other words, it has general application, but as with any distribution technique it must be applied properly. The CWCB did not use the generalized skew factors in Bulletin 17 -B for annual peak flows in their computations for this RICD as suggested by Mr. Thompson in his expert report, rather they were calculated in the appropriate manner based on the gaged data. That is probably the reason that he cannot duplicate our results. One of the benefits of the log- Pearson type III distribution is that it is not restricted to an assumption of a log - normal distribution. It provides for an adjustment to the distribution to better fit skewed data, data that does not reflect a, "normal bell shaped distribution." When skew is observed in the data, it is not prudent to ignore it and continue with the assumption of a log- normal distribution. For several time periods described in this RICD filing, the skew of the data was noticeable. Therefore, the skew of the data was hand calculated and used in computing the distribution so that the calculated distribution would better fit the observed data. However, the effect on the calculation of the 40 percentile flows is relatively small, which is expected, since the median, middle or 50 percentile value from the data is an important value for calculating the entire distribution and does not change much no matter what distribution procedure is used. Therefore adjustments made for skew have little effect near the median value. The larger effects of adjustments for skew are on the tails of a distribution (for example, the 90 and 10 percentile flows). This is expected, since skew is a measure of the non - normal shape of the tails of the . distribution. The log- Pearson type III distribution was used to allow better fit to skewed data and because of its acceptance by Federal agencies and others for flood -flow and low -flow frequency analysis has made it a well -known distribution technique and an often -used distribution technique for all purposes. However, as with any distribution technique it must be properly applied. The CWCB has at times compared the results of the log- Pearson type III distribution to other distributions and as expected, the 40 percentile value, because it is relatively close to the median value changes little between distribution techniques. In any event, the arguments advanced by Mr. Thompson relative to the data distribution technique used are for practical purposes very small if not insignificant. The significant difference in our results with those of Mr. Thompson is due to the fact that we "back cast" current depletions or uses across the entire period of historic hydrology. RICD's and Compact Development Opportunities: In Durango's rebuttal reports, it is implied that ample opportunities exist to develop Colorado's remaining apportionment in other areas of the state. While it is true that Colorado's compact apportionment can be developed anywhere in Colorado, the argument that those opportunities remain plentiful is somewhat misleading. RICD's decreed to date have been for very large flow rates and in cases where they have been extremely large have caused the basin above to become overappropriated. For example, the RICD at Gunnison has caused the basin upstream of it to become overappropriated resulting in the need to augment new water rights. This condition has resulted in Gunnison • seeking a reduction in their RICD flow amounts in water court to lessen this augmentation requirement and burden. The same situation occurred with the Steamboat RICD resulting in an apparent "selective subordination" of that RICD to fixed amounts of future upstream Flood Protection • Water Project Planning and Finance • Stream and Lake Protection Water Supply Protection • Conservation Planning