Navajo Reservoir Operations Vol. II

Home Browse Search

Previous modeling of the San Juan River in support of project authorization and Consultation under Section 7 of the Endangered Species Act (ESA) relied on Colorado River Simulation System (CRSS) estimates of the 1929 to 1974 monthly natural flows at Archuleta, New Mexico, and Bluff. As part of the San Juan River Basin modeling exercise, an analysis of the 1929 to 1974 streamflow record was conducted to determine whether there were differences in the statistical properties of the San Juan River Basin hydrology pre- and post-I 974. Statistics were calculated using a 20-year moving window to assess changes in the mean flow and the variability and seasonality of the flows. An investigation of the impacts on reservoir storage needed to meet various target yields and yield failure was also performed. The 1974 to 1993 record was found to exhibit significant differences from the prior record in terms of these criteria. It was a relatively wet period. It was therefore determined that inclusion of the 1929- 1973 data would likely lead to more reasonable and more stringent estimates of low flows and drought conditions. Therefore, the monthly 1970 to 1993 natural flows recalculated by Reclamation as explained . above were extended from 1969 back to 1929 using a spatial disaggregation model. The particular disaggregation model used preserves the mean, standard deviation, and one-month lag statistics of the hydrologic series. The model relies on key stations with full periods ofrecord (in this case 1929 to 1993) as drivers for the record extension. The natural flows at Archuleta and Bluffwere forced, by adjusting stream reach gains and losses to exactly match the CRSS natural flows at Archuleta and Bluff for the period 1929 to 1969. The 1935 to 1993 monthly gaged record for the San Juan River at Pagosa Springs, Colorado, served as the key station for stations, including all tributaries, above Navajo Reservoir. The gaged record at Pagosa Springs was extended back to 1929 using the spacial disaggregation method with the 1929 to 1934 CRSS natural flow for the San Juan River near Archuleta as its key station. For stations in the Animas drainage, the Animas River at Durango, Colorado, was the key station for 1929 to 1993. The tributaries entering the San Juan River below Farmington (La Plata, Mancos, and McElmo) were disaggregated using the La Plata River at Hesperus, . Colorado, as the key station. From the full set of natural flows (the 1929 to 1969 extension and the 1970 to 1993 Reclamation natural flows) the gains and losses were calculated for each reach by subtracting the upstream stations from the downstream station. However, for stations along the San Juan River (Farmington, Shiprock, and Four Comers, New Mexico), another method was used to find the gain and loss files. There were two reasons for the change: (I) for this study monthly natural flows at these stations needed to be further disaggregated into daily values, and (2) the daily gage error at these stations could be suppressed by using a different method to find gains and losses. For these stations along the mains tern of the San Juan River, the monthly natural flows for 1929 to 1969 were estimated by distributing gains and losses between Archuleta and Bluff (Mexican Hat). The method consisted of subtracting the monthly natural flows of the La Plata River, the Mancos River, McElmo Creek, and the CRSS San Juan River near Bluff from the CRSS natural flow at Archuleta. The net gains and losses in this reach were then distributed among the intermediate stations along the mainstem of the San Juan River. The distribution for each reach was calculated as the mean annual gain or loss using the 1970 to 1993 natural flows for the A-4 .... I"'