Schreck et al. (1986) and Utter et al. (1989) suggested that neither spring-, summer-, nor fall-run chinook salmon represent monophyletic lineages in the Pacific Northwest. Both reports state that, in general, geographic proximity was a more important factor than run- timing in predicting similarities between stocks. Thus, fish with different run-times from the same area were typically more similar than were fish from different areas with the same run-timing. This pattern suggests that run-time differences may have evolved independently a number of times following colonization of a new area by one form. Foote et al. (1989) concluded that a similar phenomenon--derivation of freshwater kokanee from anadromous sockeye salmon (Oncorhynchus nerka)--has occurred numerous times within the species O. nerka.
However, in spite of this general pattern, substantial differences are found between some populations having different run-times in the same geographic area. Striking examples of this are the pronounced genetic and life history differences between spring/summer and fall chinook salmon in the Snake River. Therefore, because of compelling evidence that fall chinook salmon are reproductively isolated from other chinook salmon in the Snake River, they are being considered separately in evaluating the ESA petitions for Snake River chinook salmon (see NMFS Status Review Report for Snake River Fall Chinook Salmon).
The relationship between spring- and summer-run fish in the Snake River is not so clear. The demarcation of the two forms based on time of adult passage at Bonneville Dam does separate some spring- and some summer-run populations that appear to be substantially reproductively isolated. However, this isolation may be due to geographical separation as much as to temporal differences in spawning time. Furthermore, as noted above, even in streams assumed to have only one of the forms, some fish may pass Bonneville Dam on the "wrong" side of the 1 June demarcation line. Thus, there is some overlap in migration timing of spring and summer chinook salmon in the Snake River.
The key to understanding the evolutionary significance of spring/summer run-timing is the relationship between the two forms in streams where they co-occur. As noted above, there is insufficient evidence at present to determine the nature of this relationship. Because we cannot rule out the possibility of substantial levels of gene flow between the two forms in at least some localities, it is inappropriate in ESA evaluations to treat the two forms as independent evolutionary lineages. Therefore, NMFS will consider the two forms as a unit in determining whether they are an ESU. This decision, however, does not imply that the two forms are not both important. Clearly, the presence of fish with a spectrum of run- and spawn-timing is crucial to the long-term health and viability of Snake River chinook salmon.
We next address the question whether Snake River spring and summer chinook salmon are represented by one or more ESUs. If they are not an ESU, then presumably they are part of a larger ESU that would have to be defined. To be considered an ESU, and hence a "species" under the ESA, a population (or group of populations) must satisfy two criteria: it must be reproductively isolated, and it must contribute substantially to the ecological/genetic diversity of the biological species.
The most compelling evidence in support of reproductive isolation in anadromous salmonid populations is their ability to return with high fidelity to their natal streams to reproduce. This is particularly true for upriver populations such as Snake River spring and summer chinook salmon (Chapman et al. 1991). The great distances that these fish travel to return to their natal streams tend to reduce the likelihood of straying from other major river systems outside the Snake River. All available tagging evidence supports the notion that virtually no straying of Columbia River fish occurs into areas occupied by Snake River spring and summer chinook salmon.
Several recent studies examined the genetic relationships among chinook salmon stocks from the Columbia River Basin. The studies differed in the populations sampled and the number of gene loci used, but they were consistent in finding substantial differences between Snake River spring- and summer-run fish and a) spring chinook salmon from the lower and mid-Columbia regions and b) summer chinook salmon from the upper Columbia River Basin. These data are consistent with the premise that there is at present little, if any, genetic exchange between Snake River spring and summer chinook salmon and these other groups. Differences between the Snake River fish and spring chinook salmon in the upper Columbia River are smaller, a result that may reflect the mixed origins of the Carson Hatchery stock.
The recent Snake River data also show significant allele frequency differences between streams in the same drainage, as well as between streams from different drainages. These differences do not suggest complete isolation of individual spawning units, but they do show that levels of genetic exchange, even between nearby populations, can be small enough for some level of differentiation to occur. Again, tagging data are consistent with the concept of, at most, a low level of straying between drainages within the Snake River Basin.
Phenotypic, life history, and genetic data support the conclusion that Snake River chinook salmon are distinct in an ecological/genetic sense. In a cluster analysis of environmental data (stream gradient, precipitation, elevation, vegetation type, etc.), Schreck et al. (1986) demonstrated two distinct groups of Snake River localities, with one group including those from the Imnaha and Grande Ronde Rivers and the other including those from the Salmon River. Both groups were quite distinct from other localities in the Columbia River Basin. Phenotypic data also indicate that the populations are structured geographically. The fact that juvenile migration behavior is the same for spring and summer chinook salmon in the Snake River, but different for these two forms in the upper Columbia River, strongly implies ecological/genetic differences between the regions. The precision required to migrate great distances from different natal streams and tributaries and return with high fidelity and exact timing to start the next generation 1 to 3 years later speaks of biological entities that are highly adapted to their particular environments. The differences detected by protein electrophoresis between Snake River spring/summer chinook salmon and chinook salmon in the lower and mid-Columbia River Basin may be an indication of adaptive genetic differences at parts of the genome not sampled by protein electrophoresis. By comparison, the genetic differences found between different spring and summer chinook salmon populations within the Snake River are rather modest.
The habitat occupied by spring/summer chinook salmon in the Snake River appears to be unique to the biological species. In contrast to coastal mountains and the Cascade Range, the Snake River drainage is typified by older, eroded mountains with high plateaus containing many small streams meandering through long meadows. Much of the area is composed of batholithic granite that is prone to erosion, creating relatively turbid water with higher alkalinity and pH in comparison to the Columbia River (Sylvester 1959). The region is arid with warm summers, resulting in higher annual temperatures than in many other salmon production areas in the Pacific Northwest. These characteristics combine to produce a highly productive habitat for these fish. As previously mentioned, the Salmon River alone once produced nearly half of the spring and summer chinook salmon returning to the Columbia River.
Chapman et al. (1991) described 10 geologic provinces in the Snake River Basin. Each is unique to some degree in the type of habitat it provides for anadromous salmonids in terms of both geology and climate. Together, these areas form an aquatic ecosystem for chinook salmon that is unique in the Columbia River Basin and, probably, the world. It seems likely that the anadromous salmonid populations that inhabit this ecosystem are unique also.
Snake River spring and summer chinook salmon as a group meet both criteria to be considered a "species" under the ESA; they are strongly isolated reproductively from other conspecific population units, and they contribute substantially to the ecological/genetic diversity of the biological species. There are indications that more than one ESU may exist within the Snake River Basin. However, we do not feel that available data are sufficient to clearly demonstrate the existence of multiple ESUs or to define their boundaries. At present, therefore, we conclude that the Snake River spring and summer chinook salmon should be a single ESU of the biological species O. tshawytscha.
This conclusion is consistent with the NMFS policy, which states that ESUs in general should correspond to more comprehensive units in the absence of clear evidence for evolutionarily important differences between smaller population segments (Waples 1991). Nevertheless, we acknowledge the geographical and ecological complexity of an area as extensive as that occupied by Snake River spring/ summer chinook salmon. In recognition of evidence for important differences between some population segments within the Snake River Basin, we emphasize that the viability of the proposed ESU is strongly dependent on the continued existence of healthy populations throughout its area. This latter provision is also consistent with published agency policy.
In determining the nature and extent of the ESU for Snake River spring and summer chinook salmon, it is also necessary to consider the effects of artificial propagation and stock transfers. In general, introduced salmon populations will not be considered for protection under the ESA (Waples 1991, Section IIIG), and changes caused by artificial propagation or hybridization may also erode qualities by which a population is recognized as distinct (Waples 1991, Section IIIC).
As discussed above and documented in more detail in the Appendix, there is a long history of human efforts to enhance production of chinook salmon in the Snake River Basin through supplementation and stock transfers. Less well understood is the extent to which these efforts have altered the genetic makeup of indigenous populations. In a recent review of studies assessing the success of efforts to supplement salmonid populations, Hindar et al. (in press) found evidence in some cases that hatchery-reared fish had interbred with native fish, but they also found cases in which repeated supplementation has had no detectable genetic effect on the indigenous population.
Considering Snake River spring and summer chinook salmon in this light, there are a number of streams in most basins without any recorded history of outplanting, and others (e.g., the Tucannon and Imnaha Rivers and Capehorn Creek in the Middle Fork of the Salmon River) that have been planted with only a minimal number of nonindigenous fish. Presumably, genetic characteristics of fish in these areas have been essentially unchanged by artificial propagation. Conversely, some streams (e.g., Catherine Creek in the Grande Ronde River drainage) have been planted with large numbers of nonindigenous hatchery fish. In many cases the hatchery stocks themselves were from mixed origins. The genetic makeup of fish in these streams may have been substantially altered by the plantings. However, more research will be necessary to conclusively demonstrate the effects of the plantings.
One area for which the evidence of stock transfers and hybridization is overwhelming is the Clearwater River. Indigenous chinook salmon populations were virtually or totally eliminated by Lewiston Dam (1927-40). Subsequent efforts to restore the runs included transfer of eggs from the Salmon River and massive outplants of juveniles from hatcheries throughout the Columbia River Basin. Descendants of these fish of mixed, nonnative origin are not considered part of the ESU for Snake River spring and summer chinook salmon. However, the habitat should be considered as part of the range of the ESU because some wild fish may persist, and the habitat contained spring and summer chinook salmon that were historically a part of the ESU as currently defined.
We have concluded that, at this time, Snake River spring/summer chinook salmon will be considered a single ESU for purposes of the ESA. As more data become available, smaller units may be defined. The next step, then, is to determine the level of risk faced by the ESU. As noted previously, factors relevant to this determination include historical and current abundance, population trends, the distribution of fish in space and time, and other information indicative of the health of the population.
During this century, man's activities have resulted in a severe and continued decline of the once robust runs of Snake River spring and summer chinook salmon. Nearly 95% of the total reduction in estimated abundance occurred prior to the mid-1900s. Over the last 30-40 years, the remaining population was further reduced nearly tenfold to about 0.5% of the estimated historical abundance. Over the last 26 years, redd counts in all index areas combined (excluding the Clearwater River) have also shown a steady decline. This is in spite of the fact that all in-river fisheries have been severely limited since the mid-1970s (Chapman et al. 1991). The 1990 redd count represented only 14.3% of the 1964 count.
To obtain insight into the likely persistence times of the ESU given present conditions, we applied the stochastic extinction model of Dennis et al. (1991) to a 33-year record of redds counted in index areas. The 33-year period is the longest possible, as redd counting in the Snake River began in 1957. We examined both sets of redd counts described previously: a 33-year series excluding the Grande Ronde River and a 26-year series that began with the first count of redds in the Grand Ronde River in 1964. We feel it is prudent to include the Grande Ronde River in at least part of the analysis because it has contributed between 10 and 20% of the total number of redds in the Snake River since 1964. Five-year running sums of redd counts (hereafter referred to as the "index value") were used to approximate the number of redds in single generations. These index values were the input data for the Dennis model; output was the probability that the index value would fall below a threshold value in a given time. An "endangered" threshold was defined as the index value at which the probability of reaching extinction (index value < 1) within the next 100 years is 5%; a "threatened" threshold was defined as the index value at which the probability of reaching the "endangered" threshold within the next 10 years is 50%.
Results of the analyses are shown in Table 1. For the 33-year time series (excluding the Grande Ronde River), the current index value of 8,456 redds is well below the threatened index value of 15,474 redds and only slightly above the endangered index value of 7,065 redds. According to the model, the probability of extinction in 100 years is 0.032, and the probability of reaching the endangered threshold in 10 years is 0.943. For the 26-year time series (including the Grande Ronde River), the current index value of 10,258 redds is somewhat above the threatened index value of 7,730 redds. According to the model, the probability of extinction in 100 years is < 0.001, and the probability of reaching the endangered threshold in 10 years is 0.270. The different results are primarily attributable to the fact that the initial index value was higher and the current index value lower in the former analysis. As previously discussed, the use of redd counts means that results of the model provide a conservative perspective of the rate of decline in abundance of adult salmon; hence, the model predictions are also conservative.
The results from the Dennis model should be regarded as rough approximations, given that the model's simplicity undoubtedly fails to consider all of the factors that can affect population viability. In particular, the model does not consider compensatory or depensatory effects that may be important at small population sizes. Nevertheless, considered together, results of the two analyses suggest that the ESU is at risk of extinction.
Other factors besides total abundance are also relevant to a threshold determination. Although the most recent data suggest that several thousand wild spring and summer chinook salmon currently return to the Snake River each year, these fish are thinly spread over a large and complex river system. In many local areas, the number of spawners in some recent years has been low. For example, in the small index area of upper Valley Creek, redd counts averaged 215 (range 83 to 350) from 1960 through 1970 (White and Cochnauer 1989). However, from 1980 through 1990, redd counts averaged only 10 (range 1 to 31) (M. White) (7). Similarly, in the large index area of the entire Middle Fork of the Salmon River, redd counts averaged 1,603 (range 1,026 to 2,180) from 1960 through 1970 but only 283 (range 38 to 972) from 1980 through 1990. If significant population subdivision occurs within the Snake River Basin (as evidence discussed above suggests may be the case), the size of some local populations may have declined to levels at which risks associated with inbreeding or other random factors become important considerations. As numbers decline, fish returning to spawn may also have difficulty finding mates if they are widely distributed in space and time of spawning.
Short-term projections for spring and summer chinook salmon in the Snake River are not optimistic. The recent series of drought years undoubtedly impacted the number of outmigrating juveniles that will produce returning adults in the next few years. The very low number of jacks returning over Lower Granite Dam in 1990 provides additional reason for concern for the ESU.
Collectively, these data indicate that spring and summer chinook salmon in the Snake River are in jeopardy: Present abundance is a small fraction of historical abundance, the Dennis model provides evidence that the ESU is at risk, threats to individual subpopulations may be greater still, and the short-term projections indicate a continuation of the downward trend in abundance. We do not feel the evidence suggests that the ESU is in imminent danger of extinction throughout a significant portion of its range; however, we do feel it is likely to become endangered in the near future if corrective measures are not taken.
Table of Contents