predicting herring distributions during winter · 2016. 11. 24. · jed macdonald 1, kai logemann ,...
TRANSCRIPT
![Page 1: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/1.jpg)
1 MARICE, University of Iceland, Reykjavík, Iceland2 Department of Mathematical Sciences, NTNU, Norway3 Hafrannsóknastofnun, Reykjavík, Iceland4 Institute of Marine Research, Bergen, Norway5 Department of Biology, University of York, York, UK
Predicting herring distributions
during winter
Jed Macdonald1, Kai Logemann1, Elias Krainski2, Þorsteinn Sigurðsson3,
Geir Huse4, Colin Beale5, Solfrid Hjøllo4, Guðrún Marteinsdóttir1
Rannsóknasjóðs síldarútvegsins
![Page 2: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/2.jpg)
Collective
learning and
sociality
![Page 3: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/3.jpg)
Atlantic herring (Clupea harengus)
![Page 4: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/4.jpg)
Atlantic herring (Clupea harengus)
Conservatism (Jakobsson 1969; Corten 1993, 2002)
![Page 5: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/5.jpg)
Conservatism defined…
&
Younger fish learn traditional migration routes from older ones and these are remembered.
Disruption of learning opportunities can trigger a
distribution shift.
TraditionsHabits
![Page 6: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/6.jpg)
Why and when but…
![Page 7: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/7.jpg)
Can we predict where
herring will spend the winter?
Photo: George McCallum
![Page 8: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/8.jpg)
Icelandic summer spawners (ISS)
Age 3+ adults targeted
from October to January
each year.
Major fishery in Icelandic
waters.
Purse-seine, pelagic
trawl.
Fishing season follows
the summer NSS /
mackerel fishery.
![Page 9: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/9.jpg)
Shifting winter distribution (Óskarsson et al. 2009)
Offshore off east coast1991/92 – 1997/98
Offshore off east
and west coasts
1998/99 – 2006/07
Inshore in small fjords on the north
coast of Snæfellnes Peninsula
2007/08 – 2013/14
![Page 10: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/10.jpg)
A model for wintering
herring
![Page 11: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/11.jpg)
=
+
+
Winter distribution
![Page 12: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/12.jpg)
Fishery and survey records – 23 years
Catch record (occurrence)
+ Survey record (absence)
![Page 13: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/13.jpg)
Habits and traditions
Describes the
occurrence pattern of
the stock the previous
season (t-1).
presencet-1
occupied cell
Describes the number
of occurrence records
in t-1.
countst-1
3.5
0
![Page 14: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/14.jpg)
Environment, predators and prey
Dynamic Static
Zooplankton
biomass in prev.
August
Bottom depth
Bottom slope
Distance to shore
Predators Prey
Sources
- CODE
- GEBCO
- Fishery logbooks
- Hjøllo et al. 2012
Salinity
Current velocity
Temperature
gradient
Fishing intensity in
prev. week
Temperature
Stratification
![Page 15: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/15.jpg)
Results – spatial predictions
Models accurately
predicted the
observations for
each year of the
time series...
Both in years when
the wintering area
stayed the same.
And, when the
distribution shifted
to a new location.
![Page 16: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/16.jpg)
Results – covariate influence
Warmer
Fresher
Recent fishing activity and static environment variables
had little influence.
stratified
velocity
zooplank.
![Page 17: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/17.jpg)
Results – demographic effects?
Strong
correlation
between the
population size
and the
probability of
herring returning
to the same
wintering area.
Suggests that prediction of the winter distribution one-
year ahead is possible…
![Page 18: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/18.jpg)
Results – prediction one-year ahead
Prediction to last four seasons of the time series.
Season r cv AUC cv Brier score
2010_11 0.704 0.961 0.142
2011_12 0.703 0.976 0.137
2012_13 0.631 0.976 0.129
2013_14 0.432 0.588 0.194
Performance was good in three out of four seasons.
Unusual mass mortality events forcing the SSN
estimate down, concurrent with strong overlap in
fished area b/n 2011_12 and 2012_13.
![Page 19: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/19.jpg)
Our model gave highly accurate spatial predictions
both within the time series, and to independent
observations one-year ahead.
The occurrence pattern of herring in one winter was
the best predictor of where they were the next.
Attachment to wintering sites increased with adult
population size.
Environmental factors (i.e. temperature, salinity,
stratification, summer prey avaliability) were also
influential.
Discussion
![Page 20: Predicting herring distributions during winter · 2016. 11. 24. · Jed Macdonald 1, Kai Logemann , Elias Krainski2, Þorsteinn Sigurðsson3, Geir Huse 4, ... 2013_14 0.432 0.588](https://reader036.vdocuments.net/reader036/viewer/2022081623/61471203f4263007b1359534/html5/thumbnails/20.jpg)
Conclusions
Collective learning often influences decision making in
animal societies.
Yet, capturing such phenomena in large-scale spatial
models remains challenging.
The models presented here attempt to incorporate
these processes in an intuitive, flexible framework.
The approach and predictions can (we hope!) benefit
ecologists, fishers and fisheries managers alike.