Grays Harbor Fall Chum Abundance and Distribution, 2018

Background

Escapement estimates for the Grays Harbor Chum Salmon (Oncorhynchus keta) population have been derived by the Washington Department of Fish and Wildlife (WDFW) each year for the past fifty years. Even at its inception in the 1970s, the current method for estimating Chum escapement was considered hastily put together out of necessity and required updating as soon as possible (Brix 1978). The current method appears to underestimate the number of Chum returning to Grays Harbor and lacks information on their distribution, but WDFW has been unable to dedicate the resources necessary for assessing the degree of underestimation. In 2014, the Aquatic Species Enhancement Plan (ASRP) Technical Committee of the Chehalis Basin Strategy began conducting studies on the habitat and aquatic species within the Chehalis Basin to understand the impacts of potential flood reduction and restoration actions. The accurate abundance and distribution of Chum returning to Grays Harbor were identified by WDFW and the ASRP as key information gaps for restoration, and work began in 2015 to fill in these gaps.

In 2015, a pilot study in Grays Harbor tributaries informed a survey frame in sub-basins to identify index reaches with high densities of Chum spawners. In 2016, a survey design was developed and implemented that utilized carcass tagging, live/dead counts, and peak spawning surveys to determine the abundance and distribution at a sub-basin level. The new survey design was first implemented in the Wynoochee and Satsop sub-basins in 2016 and 2017 (Edwards and Zimmerman 2017). In 2018, the new methodology was applied to the Humptulips sub-basin, and data from 2017 and 2018 were used to calculate abundance estimates for the Humptulips, Wynoochee, and Satsop sub-basins. The 2018 analysis resulted in a combined abundance estimate of 48,046 Chum within the tributaries of the Humptulips, Wynoochee, and Satsop. Grays Harbor tributaries with significant Chum spawning (Hoquiam, Wishkah, etc.) have yet to be assessed with the new methodology and were not included in the abundance estimate. Regardless, the escapement estimates calculated with the new methodology were consistently higher than escapement estimates for the entire Grays Harbor Chum population using the current methodology. The 2017 Wynoochee and Satsop abundance estimates alone were 50% greater than the total Grays Harbor estimate. In 2018, the Humptulips, Wynoochee, and Satsop abundance estimates were 69% greater than the total Grays Harbor estimate for the year. These findings justified the purpose of this study as well as further efforts to improve spawning escapement information on Grays Harbor Chum.

Methods

Data collected for this study include distribution inside versus outside index reaches, area-underthe- curve estimates within index reaches, carcass tagging estimates of abundance in select index reaches, survey life estimates, and total spawner abundance on Chum salmon. Distribution inside versus outside index reaches was based on live counts during a one-time survey conducted throughout the Chum survey frame during the peak spawning period. Area-under-the-curve (AUC) estimates within the index reaches were based on live counts obtained during weekly surveys. Carcass tagging estimates of abundance were based on a Jolly-Seber abundance estimator for open populations. Survey life was calculated in selected index reaches from the combination of area-under-the-curve and carcass tagging estimates of abundance. The index reaches that were used to estimate survey life represented a range of stream size classes – side channel, small, medium, and large. Abundance in all index reaches was based on AUC calculations and the survey life of the corresponding stream size classification. Total spawner abundance was the abundance in index reaches expanded by the proportion of spawning that occurred inside versus outside index reaches. Live count data used in the analysis were partitioned between ‘spawners’ (i.e., actively spawning) and ‘holders’ (i.e., holding in pools and potentially passing through the spawning area) to ensure we understood the sensitivity of the final estimate to these two different types of live counts. This distinction will be important when considering how to apply the results of this work to historical live counts from the index reaches. In 2018, the focus was the Humptulips sub-basin but abundance estimates were also calculated for the Wynoochee and Satsop using 2018 index live counts and 2017 distribution data.

Results

• Distribution inside versus outside index reaches: In the Humptulips tributaries, 44% and 46% of Chum spawning occurred within the index reaches with the highest densities observed in Stevens, O’Brien, and Donkey creeks. In the Humptulips mainstem, 32% and 40% of Chum spawning occurred within the index reaches with the highest densities of Chum observed between river miles 13.9-28.1 and 33.9-43.0. The two different proportions represent values provided by the different types of live counts (spawners only versus total live). Proportions in the Humptulips were calculated using counts from 2018 index reaches and one-time peak distribution surveys. In the Wynoochee 11% of spawning occurred within tributary index reaches and 11% occurred in mainstem index reaches. In the Satsop, 24% of spawning occurred within tributary index reaches and 40% occurred in mainstem indexes. The distribution proportions were derived from total live counts and did not distinguish spawner status. Proportions in the Wynoochee and Satsop were calculated using live/dead counts from 2017 surveys, but distribution inside and outside index reaches was based on 2018 surveys.
• Area-under-the-curve in index reaches: In the Humptulips sub-basin, fish-day calculations summed across 16 index reaches ranged between 70,173 (spawners only) and 87,619 (total live). In the Wynoochee sub-basin, fish-day calculations summed across six index reaches equaled 22,819 (total live). In the Satsop sub-basin, fish-day calculations summed across 14 index reaches equaled 18,938 (total live).
• Abundance in carcass tagging index reaches: Chum spawner abundance was estimated to be 164 in the small index (Elwood Creek) and 4,229 in the medium stream index (Stevens Creek).
• Survey life: In this study, survey life (days ± SD) represented BOTH the number of days a live Chum is present AND the observer efficiency within an index reach. For 2018, small channel index (Elwood Creek) survey life was 9.5 days (±0.9) using spawners only and 13.4 days (±1.3) using total live counts. For the 2018 medium channel index (Stevens Creek), survey life was 7.5 days (±0.2) using spawners only and 10.3 days (±0.22) using total live. For the 2017 side channel index (Satsop Tributary 0462), survey life was 8.98 days (±0.43) using counts of spawners only and total lives. The estimate did not differ by count type because no ‘holders’ were observed in the side channel index.
• Abundance in all index reaches: In the Humptulips sub-basin, abundance within the 16 index reaches was estimated between 9,011 (spawners only) and 8,204 (total live counts). In the Wynoochee sub-basin, abundance within the six index reaches was estimated to be 2,171 (total live counts). In the Satsop sub-basin, abundance within the 14 index reaches was estimated to be 1,903 (total live counts).
• Spawner abundance: The 2018 Chum spawner abundance (±SD) for the Humptulips sub-basin was estimated to be 22,328 (±345) using spawner counts only and 20,258 (±297) using total live counts. Chum spawner abundance for the Wynoochee sub-basin was estimated to be 19,964 (±956) using total live counts. Chum spawner abundance for the Satsop sub-basin was estimated to be 7,824 (±309) using total live counts.

Conclusions

The overall estimates of Chum spawner abundance differed slightly based on the type of counts (spawners only or total live counts including spawners and holders) used in the analysis. Both provided estimates that were consistently higher than those derived using the existing methodology for Grays Harbor Chum. All together, we estimated a 2018 Chum spawner abundance of approximately 48,000 (total live counts) or 50,000 (spawner only counts) for the sub-basins included in the study. Our estimate in the Humptulips, Wynoochee, and Satsop sub-basins was 20,000 and 22,000 more than the number of Chum estimated for the entire Grays Harbor basin using the existing methodology (n = 28,413). Similar to our findings in 2016 and 2017, these results suggest that the existing methodology likely underestimates the abundance of Grays Harbor Chum salmon.

A Chum abundance estimate using the new methodology was derived for the first time in the Humptulips sub-basin. Abundance estimates were also derived for the Wynoochee and Satsop sub-basins using live counts collected by multiple survey crews in 2018 and distribution data from 2017, which was not previously calculated. The abundance estimates appear to be sensitive to how live counts are collected. Designating spawner status for live Chum to account for fish that are passing through, but not spawning within the reach eventually leads to two different survey life estimates separated by as much as 3.9 days. Survey crew also appears to affect Chum live counts, with different crew counts varying as much as 30%. In addition to the importance of consistent live counts, distribution and index selection potentially greatly affects the final abundance estimate. The Satsop sub-basin abundance estimate was less than half the estimate for either the Humptulips or Wynoochee. Two possibilities for this discrepancy were identified: low Chum spawner abundance and/or low flows reduced use of some index spawning habitats. Overall, this work demonstrates that the current method of estimating escapement, underestimates Chum abundance. Moreover, further work needs to be done to better understand basin wide distribution, account for inter-annual variation, and refine methodologies.