Andy Hrycyna
Mystic River Watershed Association
February 9, 2022
Main summary
- The 2021 estimated total of fish passing through the Mystic Lakes Dam was 553,746 (+/- 47,774).
- This estimate is based on volunteer counts exclusively!
- Probable decline, then recovery. The main story is that the number of river herring migrating up the Mystic River seems to have bounced back from a decline in 2020. (We did not monitor in person because of COVID in 2020, but video evidence suggested lower numbers). In 2020 much reduced river herring runs were seen around the state, and this is blamed on drought conditions 3 years prior. Juvenile survival was probably much lower in that drought year, and so we see the effect 3-4 years later, when the (smaller than usual) group of fish born in 2016, which would have reached reproductive age in 2020, returned to the Mystic for the first time.
- Still more recovery to go. The 2021 total (~550K) is still lower than the previous high of 2019 (~790K). You would expect this: it takes time for a population to fully recover.
- Horn Pond. The Horn Pond count was 43,307 +/- 11,930. This total is higher than estimated in previous years, probably because of the work that DMF to make the existing outlet of the pond more passable for fish. Still, we know that many more fish get as far as Horn Pond but are not able to enter. The fish ladder coming through NAWCA and NRD and other money will fix this.
- Video discrepancy. The estimate from the video count this year was much lower than the official DMF estimate based on in-person counts. In the last section below, I make the argument that this is because of 2 issues: the system down for a few key days; and throughout the season the water clarity was low. In 2017, low water clarity had almost exactly the same effect on the video count.
Mystic Herring Run 2012-2021
This is the new updated graphic for the Mystic River herring migration for all years.
Daily estimates for 2021
We can also look at the estimates of total fish passing through the dam for each day in the 2021 season, based on in-person monitoring. 13 days had more than 20,000 fish pass the dam. 20K/day is a rate of more than 1500/hr for 12 hours.
Daily estimates, all years
One cool new graph possible this year is this one, which shows daily estimates for all the DMF-modeled years.
As you scan across, you can see the growth of the run over time. And you can see that busier days have gotten much busier. In the first four years the number of days with more than 25K fish passing was 6 or 7. In the four years since 2017 there have been 32 days with 25K+ fish (>2000/hr x 12 hours).
Video discrepancy
The video counts are never the same as the DMF counts. Sometimes they are notably higher and sometimes notably lower.
But this year the divergence was especially large: ~280K video vs. ~~550K in-person. Why?
If you compare the difference between the video count and the DMF count in all the years for which we have data, you can see that 2021 stands out:
| 2017 |
630098 |
410,667 |
219431 |
35% |
| 2018 |
589924 |
756798 |
-166874 |
-28% |
| 2019 |
787222 |
635015 |
152207 |
19% |
| 2021 |
553746 |
283131 |
270615 |
49% |
But if I remove just three days this year–when I know from our video data that the camera was out of commission (May 11, May 15, May 16)–the difference goes down to 39%, which is getting close to the difference in 2017. (If I include all the blackout periods, it’s even closer.)
| 2017 |
630098 |
410,667 |
219431 |
35% |
| 2018 |
589924 |
756798 |
-166874 |
-28% |
| 2019 |
787222 |
635015 |
152207 |
19% |
| 2021 |
466545 |
283131 |
183414 |
39% |
In 2017, we believe that we had a very accurate picture of what was going on. Murky water and other visibility issues made video counts and in-person counts for the exact same periods systematically lower. Video observers just couldn’t see the fish, or the murkiness reduced contrast to the point where fish passing through did not trigger the system’s motion sensors, so fish passed through without videos being recorded.
2021 was also a year of very murky conditions throughout the viewing period. So I believe this is the best explanation for the lower video numbers: low visibility leading to undercounts, plus equipment failure on a handful of days.
2018 we had the opposite problem. The water was exceptionally clear that season, AND we had problems filling slots in the in-person program. In that year, we know that we had gaps in the in-person count on very busy fish migration days. The daily discrpencies were sometimes even greater than this year, but in the opposite direction.
Takeaways:
- The video estimate is affected greatly by water clarity.
- Staying on top of the equipment condition is much harder during COVID when I live 1/2 away from the dam, rather than working 8 minutes away. Wi-fi was also down this year, making remote inspection of equipment impossible. We need to fix the wi-fi problem this year, and maybe assign the job of frequent checks on the equipment to a super volunteer who lives close.
- If we can get excellent compliance from our volunteer in-person monitoring team–which we did this year–the results will likely be more reliable than the video system, no matter how many videos we get counted.
