U.S. Dept Commerce/NOAA/NMFS/NWFSC/Publications
Tech Memo-22: Status Review for Mid-Columbia River Summer Chinook Salmon


Historic Abundance Estimates and Trends

Historic abundance estimates for chinook salmon in the upper and middle Columbia River Basin are in the hundreds of thousands (Mullan 1987). However, with the advent of commercial fisheries in the 1800s, these populations were declining by 1900. A well- documented example of this pattern of change concerns stocks in the Yakima River. Estimates of salmon returns (all species) to the Yakima River in the mid-1800s range from 160,000 to 250,000 (Robison 1957). With agricultural and fisheries development, the run size (all species) fell to 20,000 from 1875 to 1905, and then dropped to 1,000 around 1930 (Robison 1957). Summer run chinook salmon were last seen in the Yakima River in 1970 (YFP 1992). Estimates of fall chinook salmon from 1983 to 1987 averaged 570 (range 221 to 1,332; Fast et al. 1989).

There is some historical evidence that chinook salmon upstream from the MCR underwent periodic fluctuations in abundance. Native Americans in this portion of the Columbia River Basin traditionally depended on salmon for subsistence. Yet episodes of starvation among Native American tribes were recorded in 1811, from 1826 to 1829, and in 1831 (Mullan et al. 1992b). These episodes appear to be linked to the failure of chinook salmon to return to Kettle Falls, the primary aboriginal fishing area on the upper-Columbia River (Fig. 1), in sufficient numbers to sustain the local population (Mullan et al. 1992b). In 1831, salmon had to be procured from Fraser River tribes to sustain the traders at Fort Walla Walla through the winter, since few Columbia River chinook salmon had been caught the previous summer. Assuming that spawning and rearing habitat were still pristine at the time, it is likely that variations in ocean conditions were responsible for the extreme fluctuations of chinook salmon populations of the upper Columbia River observed by the earliest settlers (Mullan et al. 1992b). Similarly, chinook salmon abundance in the Fraser River--an undammed river system--has fluctuated substantially over time, apparently due to variations in ocean conditions (Richards and Olsen 1993).

Present Abundance Estimates and Trends

In considering whether ocean-type chinook salmon of the MCR are threatened or endangered according to the ESA, we evaluated both qualitative and quantitative information. Qualitative evaluations considered recent, published assessments of the status of chinook salmon stocks within the petitioned area by agencies or conservation groups (Nehlsen et al. 1991, WDF et al. 1993). These assessments are summarized in Table 4. Nehlsen et al. (1991) considered summer chinook salmon to be of "special concern" in the Okanogan River and at "moderate risk of extinction" in the Methow River. Summer chinook salmon in the Entiat River and spring chinook salmon in the Okanogan River were considered to be extinct. No other chinook salmon populations in the MCR were reported at risk. Washington Department of Fisheries (WDF et al. 1993) considered summer chinook salmon in the Okanogan and Methow Rivers to be "depressed" but rated all other existing summer and fall chinook salmon stocks in this region as "healthy."

Quantitative assessments were based on time series of adult counts at Columbia River dams and redd counts in the tributary rivers. However, counts from lower Columbia River dams provided little direct information about MCR ocean-type fish. First, these counts included summer-run chinook salmon from the Snake River. Second, there is abundant evidence that inflexible cutoff dates between spring-, summer-, and fall-run chinook salmon can lead to erroneous conclusions based on dam counts alone, particularly in years with unusual environmental and oceanic conditions, as discussed above. This second caveat also applies to attempts to separately enumerate summer- and fall-run chinook salmon at upriver dams. Therefore, we evaluated passage of summer- and fall-run chinook salmon as a single unit.

Table 4. Recent qualitative assessments of status of ocean-type chinook salmon stocks in the mid-Columbia region (Nehlsen et al. 1991, WDF et al. 1993).

Nehlsen et al. WDF et al. (SASSI)b
River Basin Run Risk Levela Origin Production Status
Similkameen Summer -- -- -- --
Okanogan Summer Special Concern Native Wild Depressed
Methow Summer Moderate
Mixed Wild Depressed
Lake Chelan Fall -- Nonnative Wild Healthy
Entiat Summer Extinct -- -- --
Wenatchee Summer -- Mixed Wild Healthy
Hanford Fall -- Native Wild Healthy
Yakima Fall -- Unknown Composite Healthy
Marion Drain Fall -- Native Wild Healthy

a Nehlsen et al. (1991) identified stocks as extinct, at high risk of extinction, at moderate risk of extinction, or of special concern. Their list did not include stocks that did not qualify for one of these categories. The dash indicates that the stock(s) was not mentioned.

b SASSI ("Salmon and Steelhead Stock Inventory," WDF et al. 1993) classified stocks by apparent genetic origin (native, nonnative, or mixed), by current production type (wild, hatchery, or composite), and by apparent stock status (critical, depressed, or healthy). These are preliminary classes, subject to later refinement, and reflect the consensus opinion of state and tribal biologists contributing to the report. Dash indicates that the stock(s) was not mentioned.

The petitioners provided only selected information about escapement in the MCR, and we consulted other sources to provide a more complete picture. Redd counts (including both summer- and fall-run fish) for the Wenatchee, Methow, Okanogan, and Similkameen Rivers all show large fluctuations, with very low points in the early 1980s (Fig. 4). Since that time, Wenatchee River redd counts have shown a substantial increase, while those for the other three rivers have shown no discernible trend. Over the entire available data series (1956-93), only the Methow River redd counts show a substantial downward trend (Table 5), although both the Methow and Okanogan River counts are substantially below peak counts from the late 1960s and early 1970s. Both the Similkameen and Wenatchee River counts show substantial upward trends over the full data series.

Historic counts of adult salmon ascending fish ladders at dams provide additional assessment of population abundance and trends. The longest record for the mid-Columbia River is from Rock Island Dam. Because of its location, counts from Rock Island Dam provide an index primarily of the petitioned stocks, not of the entire ESU. Counts of adult late-run ocean-type chinook salmon at this dam (Fig. 5) showed a decline in the late 1930s, followed by a substantial increase during the 1940s and 1950s. Since the late 1950s, abundance has fluctuated over about a three-fold range with no substantial trend. However, over this period, the summer component has been a decreasing proportion of the total run.

Figure 4
Figure 4. Ocean-type chinook salmon redd counts (expanded aerial counts) for the Wenatchee, Methow, Okanogan, and Similkameen Rivers, 1956-93. Data from Chapman et al. (1994, appendix D-2).

Table 5. Summary of recent 5-year average (1989-93) and percent annual change of ocean-type chinook salmon redd counts (expanded aerial counts) in the Columbia River Basin above Rock Island Dam (RIS). Percent annual change is based on exponential regression of passage counts against time for two periods: the entire data series (1956-93), and the most recent 20 years (1974- 93). Based on data from Chapman et al. (1994; appendix D-2)
Average Percent annual change
Location redd count Full series Last 20 years
Similkameen 191
+2.8 -0.3
Okanogan 112
-0.1 -3.1
Methow 202
-2.9 -6.4
Wenatchee 2,530
+1.5 +1.6
Total above RIS* 3,064
+0.5 +0.3
* Includes a small number of redds counted in the Entiat River, the Chelan River, and the Columbia River between Rock Reach Dam and Wells Dam.
Figure 5
Figure 5. Counts of adult ocean-type chinook salmon passing Rock Island Dam, 1933-93. Based on data from Mullan (1987; counts for 1933-59) and the Coordinated Information System Computerized Distributed System (Anderson 1994; counts for 1960-93).

Counts of late-run ocean-type adult chinook salmon in 1991 and 1992 at mid-Columbia River dams were well below the 10-year average, and at some dams the 1992 counts were record lows. In 1993, counts at all dams were well above the 1992 low point, and most were near or above the recent (1984-93) 10-year average. While the low counts in 1991 and 1992 are of concern, they are not unprecedented. Similar low counts occurred in 1982 and 1983, after which counts increased to reach record highs at Priest Rapids and Rock Island Dams in the late 1980s. Average run-size and trend estimates for these dams are summarized in Table 6.

Total abundance of MCR ocean-type chinook salmon is relatively large, with a recent (1989-93) 5-year average annual estimate of 22,000 adults passing Priest Rapids Dam. An additional 42,000 adults spawned in the Hanford Reach and Yakima River, for a total 5-year average annual run of approximately 64,000 MCR ocean-type chinook salmon Fig. 6). The number of fish spawning in Hanford Reach and Yakima River was estimated from McNary Dam counts, with counts from Priest Rapids Dam, Ice Harbor Dam, and various hatcheries subtracted. Note, however, that these estimates may be subject to large errors.

NMFS has not attempted to estimate extinction probabilities for these stocks and in previous status reviews has viewed such estimates with caution. However, two reports submitted to the administrative record (ADFG 1993, Chapman et al. 1994) have attempted such estimates. Both applied the model of Dennis et al. (1991) to an aggregate stock of mid-Columbia River chinook salmon. The Alaska Department of Fish and Game (ADFG 1993) estimated a probability of less than 0.001% that Priest Rapids Dam counts of adult summer-run chinook would decline to one fish by the year 2093.

Table 6. Summary of recent 5-year average (1989-93) and percent annual change of ocean-type chinook salmon passing four mid-Columbia River dams. Percent annual change is based on exponential regression of passage counts against time for two periods: the entire data series (varies by dam), and the most recent 20 years (1974-93). Based on data from the Coordinated Information System Computerized Distributed System (Anderson 1994).
Wells Rocky Reach Rock Island Priest Rapids
Average Percent annual change
Location run run size Full series Last 20 years

Figure 6
Figure 6. Estimated total adult run size for the ESU, 1964-93. Run size was estimated from total summer- and fall-run chinook salmon counts at McNary Dam with Snake River fish (Ice Harbor Dam counts) and returns to Priest Rapids and Wells Hatcheries subtracted. Data from the Coordinated Information System Computerized Distributed System (Anderson 1994), ODFW and WDF (1993), and Chapman et al. (1994).

However, the use of a threshold of one individual is inconsistent with the methods described by Dennis et al. (1991), and no confidence intervals were provided for the ADFG result. Therefore, we concluded that the reliability of this estimate is questionable. Chapman et al. (1994) applied the model to 5-year running sums of estimates of summer/fall chinook salmon spawner abundance based on redd counts above Rock Island Dam. They found that the probability of decline to 5,000 fish in 10 years was 0.02% (95% confidence interval extends from 0-0.1%) and in 50 years was 9% (confidence interval from 0-50%). Probabilities of decline to 10,000 fish were estimated to be 3% (95% confidence interval 0-15%) in 10 years, and 26% (confidence interval l0-93%) in 50 years.

These results should be viewed with caution: they are based on a very simple model with unrealistic assumptions regarding salmon population age structure, they may not include all sources of variation in stock abundance, they ignore subpopulation structure, and the estimates had wide confidence intervals. However, they suggest that the near-term risk of extinction of this stock is low.

We consider the petitioners' assessment of hatchery composition of the stocks to be accurate. Although over 200 million late-run hatchery chinook salmon have been introduced into the MCR since 1941, their influence on naturally-spawning ocean-type chinook salmon in the tributaries above Rock Island Dam has been relatively small and does not appear to pose a significant risk factor to these stocks. There has been substantial hatchery production of fall-run chinook salmon at Priest Rapids Hatchery, and upriver bright fish from this and other hatcheries have been planted in the Yakima River.

Table of Contents