The FGE's were calibrated with data from fish collected and PIT tagged in the Snake River trap and recovered at LGR, LGS, LMO (1993-1995 only), and MCN dams. Data was used from the years 1990 through 1995. The numbers marked and recaptured are given in Table 42. The data was extracted from the PTAGIS data base. All recoveries were used, which means that in 1993-1995 significant numbers of fish were detected more than once due to slide-gate operations for NMFS survival studies.
The FGE's estimated from the fitting process are given in Table 43 and comparisons to FGE's determined by fyke net and PIT tag studies are given in Table 44. The CRiSP-based estimates are within a standard deviation for the PIT tag derived values for Lower Granite an Little Goose dams and for the fyke net derived value for McNary dam. The fyke net derived value for Lower Granite and Little Goose dams are higher than the CRiSP based estimates. This difference suggests that point-estimated fyke net-based FGE values may not be representative of seasonal passage conditions at upriver dams, due to changing fish condition, smoltification, and other factors.
Results for 1993-1995 are complicated by slide-gate operation at some of the projects; while this was included in modeling FGE, it adds a degree of uncertainty to our estimates. Clearly, FGE at McNary Dam was not 100% in 1994 - although a remarkably large fraction of the release was detected there as compared to other years (27% in 1994; average of other years with slide gates = 16.5%). Similarly, detections at Lower Monumental in 1995 were surprisingly high. Moreover, system operations in 1989 allowed many tagged fish to escape collection at Lower Granite Dam, despite entering the bypass system; this would lead us to underestimate FGE at Lower Granite and to overestimate at downstream projects, since more fish were in the system than should have been (G. Matthews, NMFS Seattle, pers. comm.). We consider these numbers anomalous and recommend using mean values that exclude the outliers (see Table 46).
Note that our estimates are consistent with estimates of FGE and collection efficiencies determined during the NMFS survival studies of 1993-1995 (Table 45)(Iwamoto et al. 1994, Muir et al. 1995, 1996). While collection efficiencies represent minimum estimates of FGE due to loss of smolts via spillway passage, they provide another check on CRiSP calibration.
The FGE's developed using CRiSP are consistent with the observed passage numbers and the more robust estimates based on PIT tag studies. The suggested FGE means and ranges for uses in modeling passage conditions for the period 1996-1998 are given in Table 46. The range was derived from the range observed in MCN fyke net studies in 1992 which was from 37 to 91 with a mean of 61.
Note in Table 44 that the fyke net FGE estimates are generally greater than the CRiSP-derived estimates. Since this may be from bias in the FGE studies we suggest lowering all fyke net FGE's by some amount. The average CRiSP derived FGE is about 80% of the fyke net-derived estimates from three dams. We therefore suggest lowering the FGE to 80% of the fyke estimates.
| 1Stuehrenberg and Johnson (1990) 2Cramer, Willis and Witty (1995) |
| aBased on estimate from coordination process and adjusted downward to account for bias in fyke net studies (provided by D. Askren of BPA) |
PIT tag data from wild-caught Snake River subyearling chinook were used to estimate actual FGE during 1993 through 1995. Previous years' data suffered from small sample sizes and poor collection rates. The approach is to create a release of fish in CRiSP that has the same properties as the actual tagging groups and then to adjust FGE at Lower Granite, Little Goose, and McNary dams to obtain good agreement in number of fish collected. The data are given in Table 47 and the resulting FGE calibration is given in Table 48 below, along with the NMFS Coordination estimates for mean, low, and high FGE's at the relevant projects. For 1994 a considerable mortality rate must be assumed following tagging given the extremely low rate of recapture; also note that in 1995 slide gate operations resulted in considerable rates of return to the river for PIT-tagged fish during the fall chinook outmigration season. Note that estimates of FGE are close to the specified range of coordination FGE values.
| Table 47 Fall chinook PIT tag release and recapture information for 1993-1995 | |||||
|---|---|---|---|---|---|
| Year | Release | LGR | LGO | LMO | MCN |
| 1993 | 1509 | 251 | 52 | 10 | 5 |
| 1994 | 2776 | 201 | 50 | 50 | 33 |
| 1995 | 1652 | 440 | 238 | 198 | 147 |
A similar approach was taken for juvenile steelhead PIT tagged from the Dworshak hatchery. These fish were detected at Lower Granite, Little Goose, and McNary Dams, and assuming travel time and mortality algorithms were calibrated, estimates of FGE could obtained for these projects. FGE was estimated using data from 1989-1995 inclusive (Table 49). Because of the variation in year-to-year fits, the average of these years' FGE values was used. Note that the PIT tag-calibrated FGE value is close to that estimated by NMFS for coordination purposes, but at McNary, for spring chinook, the calibrated value is about 5/6 that of the coordination value in the System Operation Review. This makes sense in the context of the fyke net argument made above. Also note that 1994 and 1995 observations are complicated by the fact that slide gates were in operation at all three upper projects; this led, for example, to an astonishingly high collection rate at Lower Monumental Dam in 1995.
Nearly all Federal projects on the Columbia and Snake Rivers have undergone considerable change since their initial construction. Most have added bypass systems or other mechanisms to provide improved juvenile passage; consequently, FGE has improved over time. We have used current estimates of FGE and scaled them for historical patterns of screen addition, gate raises, and other operational changes that alter FGE. Estimated historical FGE values for CRiSP 1.5 for all species are given in Table 50. A comparison of the CRiSP1.5 FGE values to values used in CRiSP1.4 and in the Previous Coordinations are given in Table 51.
The calibration of time varying FGE is not available for CRiSP1.5.
Fish guidance goes to zero when the surface elevation drops below the bypass orifice elevation (Fig. 57). This parameter, designated bypass_elevation, is set in the columbia.desc file. If bypass_elevation is missing or commented out (with #) the bypass elevation is set to the pool floor_elevation and bypass will occur for all reservoir elevations. This function applies with or without selection of age dependent fge.
The bypass elevations and forebay elevations in feet above sea level (obtained from the Army Corps of Engineers) are set in the columbia.desc file for each dam where a bypass system exists.