Once again, DFO shows how well they can estimate. Sad.

OldBlackDog

Well-Known Member
Management Group

Pre-Season P50 Run Size

End-of-season Run Size

Difference

Early Stuart

99,000

47,000

-53%

Early Summer Run

343,000

163,000

-52%

Summer Run

3,407,000

1,062,000

-69%

Late Run

583,000

210,000

-64%

Total Sockeye

6,788,000

2,120,000

-67%

Total Pink Salmon

8,693,000

3,616,000

-58%
 
Well before you start criticizing forecasting you should first try to understand it, how to best interpret it, how it fits into the Fraser Sockeye assessment cycle, and the challenges as well as limitations forecasters face. The best place to start is the actual document:

http://www.dfo-mpo.gc.ca/csas-sccs/Publications/ScR-RS/2017/2017_016-eng.pdf

A probability of p50 basically assumes average survival. It’s the probability that the media, environmentalist and even fishery managers quote, but it’s a fallacy to see forecasts like this as a single point estimate. It happens every year. Although the modelling can get quite technical and confusing the uncertainty that surrounds these forecasts shouldn’t be. These forecasts need to be viewed as a “range” of probabilities. These reflect the variability in marine survival as well as the what we know about freshwater survival which also has challenges. For instance, see the bottom of page 16 in the document I attached. The inability to get a reliable smolt outmigration at Chilko in 2015 due to extremely high water levels in the spring represented a big obstacle in forecasting this season. Summers are generally a main contributor to escapement with Chilko making up most of this.

Secondly, the end of season run size are PSC estimates at Mission which also have their own uncertainty and challenges, so they should not be viewed as final run size. Also note that “total run size” and “total escapement” are not the same thing, so I hope the are not being used synonymously. We don’t know final 2017 escapements or final run size for 2017 yet.

So, before ragging on estimates take the time to understand.
 
Thank you Shuswap for that last document. I do appreciate the efforts in assessing and forecasting salmon stocks.
Please correct me if I am wrong but it is appearing that all escapement assessment past smolt enumeration is considered "ocean survival"? Is there no "migratory corridor" or "marine transition" escapement area? There is a large area of migration in stream and transition area from fresh to salt that does not seem to be included with stock assessment formulas. Is this area indeed grouped with "ocean survival"?
 
Interesting, but missing the mark by 50% or more is a huge concern.
So if you are going to defend this, explain what is needed to got the forecast within 15%?

I am guessing that they just don’t know and throw numbers around.

Why not just say they have no ideal and need millions of dollars to get a handle on it.




Well before you start criticizing forecasting you should first try to understand it, how to best interpret it, how it fits into the Fraser Sockeye assessment cycle, and the challenges as well as limitations forecasters face. The best place to start is the actual document:

http://www.dfo-mpo.gc.ca/csas-sccs/Publications/ScR-RS/2017/2017_016-eng.pdf

A probability of p50 basically assumes average survival. It’s the probability that the media, environmentalist and even fishery managers quote, but it’s a fallacy to see forecasts like this as a single point estimate. It happens every year. Although the modelling can get quite technical and confusing the uncertainty that surrounds these forecasts shouldn’t be. These forecasts need to be viewed as a “range” of probabilities. These reflect the variability in marine survival as well as the what we know about freshwater survival which also has challenges. For instance, see the bottom of page 16 in the document I attached. The inability to get a reliable smolt outmigration at Chilko in 2015 due to extremely high water levels in the spring represented a big obstacle in forecasting this season. Summers are generally a main contributor to escapement with Chilko making up most of this.

Secondly, the end of season run size are PSC estimates at Mission which also have their own uncertainty and challenges, so they should not be viewed as final run size. Also note that “total run size” and “total escapement” are not the same thing, so I hope the are not being used synonymously. We don’t know final 2017 escapements or final run size for 2017 yet.

So, before ragging on estimates take the time to understand.
 
Thank you Shuswap for that last document. I do appreciate the efforts in assessing and forecasting salmon stocks.
Please correct me if I am wrong but it is appearing that all escapement assessment past smolt enumeration is considered "ocean survival"? Is there no "migratory corridor" or "marine transition" escapement area? There is a large area of migration in stream and transition area from fresh to salt that does not seem to be included with stock assessment formulas. Is this area indeed grouped with "ocean survival"?
Sorry I don’t really understand your question entirely. However, in the case of Chilko, ocean survival would include the freshwater outmigration to the ocean. That’s why Chilko is an important indicator stock for Fraser Sockeye. There is an estimate of outmigrating smolts and adult escapement which assists forecasters. Unfortunately, fence installation, maintenance and operation is dependent on water levels. In the case of Chilko, water levels were far too high too early to install the smolt fence.
 
Interesting, but missing the mark by 50% or more is a huge concern.
So if you are going to defend this, explain what is needed to got the forecast within 15%?

I am guessing that they just don’t know and throw numbers around.

Why not just say they have no ideal and need millions of dollars to get a handle on it.
Did you read the document I provided? That’s the first thing you need to do. When you say “get the forecast within 15%” you are not understanding or appreciating the complexities in factoring the variability of marine survival. Fraser Sockeye predominantly return as 4 year olds with half of that spent in the marine environment so to think one can forecast to within 15% is not living in reality. That’s why forecasts are presented as a probability distribution. A probability of 50% assumes average survival which I think most here can agree is not advisable given what we have seen, especially from 2009 to 2010. No, they don’t just throw numbers around, so I’m assuming you haven’t read the document. There is no key indicator in the marine environment that can consistently produce what critics want, but it doesn’t mean there’s isn’t work being done to make forecasts better. Again, you also need to understand how forecasting fits into the WHOLE Fraser Sockeye assessment cycle. For that read Cohen Volume 1.
 
No clue on run size so open it to full blown netting from april till december..
Determining Final run size is a formal process post season where the PSC and DFO get together with applicable researchers to see what the return was. Specially they talk about DBEs (Difference between estimates), enroute mortality and environmental conditions in the Fraser. What you are seeing now are estimates.
 
Sorry I don’t really understand your question entirely. However, in the case of Chilko, ocean survival would include the freshwater outmigration to the ocean. That’s why Chilko is an important indicator stock for Fraser Sockeye. There is an estimate of outmigrating smolts and adult escapement which assists forecasters. Unfortunately, fence installation, maintenance and operation is dependent on water levels. In the case of Chilko, water levels were far too high too early to install the smolt fence.
Thanks Shuswap. You fully answered my question. I will start another thread to express my concerns.
 
Most interesting.

Seems that other people have problems with DFO estimates.
 
I am sorry but I just had to chime in here @OldBlackDog. Either you didn't read or didn't understand the document that was posted by @shuswap. I know it is very easy (and often fun) to blame DFO for complete incompetence but they are working with the tools and resources they have been provided. The document does a pretty good job of explaining forecast methodology as well as explaining what the ranges of estimates indicate. It is a very simplistic mentality to say "why can't they just get it within 15%"? It is just marginally more difficult to look into the forecasts with a little more depth to appreciate how they are estimated. There are lots of issues with DFO that are worthy of civil discussion but the "DFO is a bunch or morons" line doesn't help with anything.

By the way, you can update your signature/quotation after watching this 2 minute video below:




Interesting, but missing the mark by 50% or more is a huge concern.
So if you are going to defend this, explain what is needed to got the forecast within 15%?

I am guessing that they just don’t know and throw numbers around.

Why not just say they have no ideal and need millions of dollars to get a handle on it.
 
Sorry missed putting this with my post.


http://steelheadvoices.com/?p=590#more-590

Small section for you.
The Weight of Evidence
UncategorizedComments: 1

Over many years in the fisheries business I’ve witnessed more predictions regarding salmon returns than I care to remember. One thing I’ve not forgotten, however, is my overriding impression – if you want to be wrong, just make a prediction. British Columbia is replete with examples of that this year. I don’t think I’m overstating it to say the batting average was so low there wouldn’t be a glimmer of hope for a new contract for any of the predictors if they were ball players. I don’t see that as a lack of available information as much as I do failure to look for it and/or apply it.

The business of predicting steelhead returns involves much greater uncertainty than for salmon. Steelhead have so many combinations of freshwater and ocean ages and repeat spawning frequencies that confidence in anything other than broad scale forecasts is probably not something most managers want to think about. That aside, here’s some information that should be top of mind.

The warm water intrusion (“the blob”) that befell virtually the entire ocean migration and rearing area for steelhead, especially those originating from British Columbia, during 2015 and well into 2016 clearly had major impacts on the returns to our streams in 2017. That was 100% predictable but I see no evidence of anything but surprise on the part of fisheries managers when it happened. Lots of head scratching after the fact but no acknowledgement of the science that predicted it. Now there is even more science that is sounding alarms for 2018 and even beyond





I am sorry but I just had to chime in here @OldBlackDog. Either you didn't read or didn't understand the document that was posted by @shuswap. I know it is very easy (and often fun) to blame DFO for complete incompetence but they are working with the tools and resources they have been provided. The document does a pretty good job of explaining forecast methodology as well as explaining what the ranges of estimates indicate. It is a very simplistic mentality to say "why can't they just get it within 15%"? It is just marginally more difficult to look into the forecasts with a little more depth to appreciate how they are estimated. There are lots of issues with DFO that are worthy of civil discussion but the "DFO is a bunch or morons" line doesn't help with anything.

By the way, you can update your signature/quotation after watching this 2 minute video below:

 
... if you want to be wrong, just make a prediction ...The warm water intrusion (“the blob”) that befell virtually the entire ocean migration and rearing area for steelhead, especially those originating from British Columbia, during 2015 and well into 2016 clearly had major impacts on the returns to our streams in 2017. That was 100% predictable but I see no evidence of anything but surprise on the part of fisheries managers when it happened. Lots of head scratching after the fact but no acknowledgement of the science that predicted it. ..
Not sure what "Fisheries Managers" Bob claims he communicated with - or how many - but I never experienced any "surprise" from ANY Fisheries Manager I know that the blob had a (largely) negative effect on salmon. The surprise might be on how negative it was, instead...
 
Well, you can always send him an e-mail on the site and ask him.
I tend to agree with him on this and question if the people doing the projections are capable.



QUOTE="agentaqua, post: 853488, member: 470"]Not sure what "Fisheries Managers" Bob claims he communicated with - or how many - but I never experienced any "surprise" from ANY Fisheries Manager I know that the blob had a (largely) negative effect on salmon. The surprise might be on how negative it was, instead...[/QUOTE]
 
Thanks for the invite, OBD - if I was worried about Bob's opinion - I might. Guess he doesn't circulate in the same circles as where most Fisheries Managers communicate.
 
I come from a chinook stock assessment background so not familiar with what they use to project pink and sockeye returns. For chinook they look at a number of factors including brood size of of all age classes, jack returns, previous 2 year returns.

There is so much going on in the ocean that it is near impossible to take all the factors into account. Factors such as environmental condition as time the fish leave the freshwater is very important but with so many different life histories even within a single system it is so hard to know what the affect will be.

In many of the systems off the north Thompson this fall the returns were higher than projected so it goes both ways
 
Again, what do they need to get closer to the mark?
I am guessing that more ocean information would be right up there?

So, that needs to be addressed.



I come from a chinook stock assessment background so not familiar with what they use to project pink and sockeye returns. For chinook they look at a number of factors including brood size of of all age classes, jack returns, previous 2 year returns.

There is so much going on in the ocean that it is near impossible to take all the factors into account. Factors such as environmental condition as time the fish leave the freshwater is very important but with so many different life histories even within a single system it is so hard to know what the affect will be.

In many of the systems off the north Thompson this fall the returns were higher than projected so it goes both ways
 
Again, what do they need to get closer to the mark? I am guessing that more ocean information would be right up there? So, that needs to be addressed.
Well, more stock assessment - that's for sure, ODB.

How, where, how often, how much are the words in the debate over more stock assessment.

Certainly wrt understanding ocean survival rates - we have few smolt fences on the coast. That hinders trying to estimate ocean survival rates and apportion out early life history mortality (for anadromous salmonids anyways).

I think we are going to need a much longer thread in discussing it...
 
Again, what do they need to get closer to the mark?
I am guessing that more ocean information would be right up there?

So, that needs to be addressed.
They can get all the data about ocean conditions that they want, the issue is applying that data to chinook survival. There is also the difficalty between establishing cause and effect.

For example the huge pink run a few years back (2010 I believe) has papers suggesting the large return was the result of iron and phosphorus from a volcanic eruption causing an algae bloom that generated large amounts of food availability. There are many papers disproving that each showing a different reason for the large run. The truth likely lies somewhere in the middle.

The offshore area is so vast and there are so many different life histories and timings of a chinook lifecycle that what can be good for one group of fish could adversely affect another. Ecologically it is called the portfolio affect. The more diverse the array of life histories and timings the more likely it is that while one group may be harmed another may benefit and offset each other. that difference could be simply 2 months between when a fish migrates into the ocean.

With the satellite and remote sensing equipment we have today we can get a tonne of data from the offshore rearing grounds. But it is of no use if we are unable to apply or even quantify how it will affect salmon.

Likewise juvenile mortality being as important as it is to recruitment, the timing of there arrival to the ocean and coinciding food availability could be a strong indicator. but when you look at for example the eagle river, wap, lower and mid Shuswap and all the life histories, these being jacks, 3yr, 4yr, 5yr returners, each with tendencies of "immediate", "ocean type", and "stream type". We have fish hatching from a single spawn year migrating to the ocean over the course of 14 months and returning over the cours of up to 4 years from a tiny portion of the North Thompson section of the Fraser. So even if you have all the data on ocean conditions when they make it to the ocean. Good luck quantifying the affect when you add in all the other tributaries that are producing similarly diverse fish.

And this all doesn't even take into account river conditions, predation, fishery, hatcheries etc

IMO There is neither the resources nor the knowledge to accurately predict the next years return. But the assessment itself of the escapement, that being the return to rivers is quite accurate and gives info on the survival of these different life histories which we can use retrospectively to observe how conditions or events affect salmon survival
 
Last edited:
Hello Rover. I think you are bang on with your last post. imo dfo does a pretty good job forecasting returns with the tools available.
There seems to be a gap in knowledge with the outmigration and transition to sea phase for salmonids. I personally observed coast wide depopulation of macro invertebrates which explains the low s/h productivity. There is no recognition of this yet. Do you know of any long term plankton sampling that could be used to compare with sockeye and pink abundance trends? I've tried googling it and keep getting lost in many other interesting reports that don't have consistent or recent data. After a couple hours of searching and reading my back is sore and frustration kicks in. I am curious to know if there was healthy abundance of plankton flowing down the Thompson and Fraser during the out migration years of this years returns? If you or anyone else knows of this info please direct me to it. Thank you.
 
Back
Top