Annual and seasonal changes in fish presence in the lower ...
Annual and seasonal changes in fish presence in the lower Penobscot River 2010-‐2013 Garrett Staines and Gayle Zydlewski University of Maine, School of Marine Sciences
Open Rivers Ini@a@ve
Acknowledgements:
Patrick Erbland data collec@on Megan Altenri2er fieldwork Rachel Kocik data processing Haixue Shen data processing
Objec@ve Create metric for seasonal and annual comparisons
Fisheries Acous@cs Non-‐invasive with large spa@al and temporal coverage
Some examples: • Roanoke River; American shad and striped bass (Hightower) • Kenai River; Chinook and sockeye salmon (Burwen) • Chandalar River; chum salmon (Daum)
And the Penobscot
et al.
Split-‐beam Echosounders
“Fish tracks” for quan@fica@on transducer sound beam
Study Area Veazie Dam Site
Hampden
Verona Island
Brewer
Study Site
Sample Site Cross Sec@on Coverage East Cianbro Side
West Mean Low-‐@de
180 m wide x 10 m deep 3 meter @dal flux
Waterfront Side
Data Collected JAN
2010
2011
2012
FEB
MAR
APR
MAY
JUN
JUL
AUG
Waterfront Cianbro Waterfront Cianbro Waterfront Cianbro Waterfront
2013
Cianbro
*some gaps from equipment malfunc@on
SEP
OCT
NOV
DEC
Current data subset
• One side, limited sampling range, certain @dal and diel • Eight 15 minute files per month • Visual counts with subsamples; compare with tracking • Try several fish tracking parameters; choose best by year
Full Data Set JAN
2010
2011
2012
FEB
MAR
APR
MAY
JUN
JUL
Waterfront
AUG
SEP
OCT
Sub-‐set NOV
JAN
DEC
FEB
Waterfront
Waterfront
JUN
JUL
AUG
SEP
OCT
Waterfront Cianbro
Waterfront
2012
Cianbro
2013
MAY
Cianbro
2011
Cianbro
APR
Waterfront
2010
Cianbro
MAR
Cianbro
Waterfront
Waterfront
2013
Cianbro
~80 m
Cianbro
Cianbro, night, high slack @de only 15 m
NOV
DEC
Current data subset
• Cianbro side with first 15 m range, night high slack @de • Eight 15 min files per month; May-‐Oct; 2010-‐2013 • Visual counts with subsamples; compare with tracking • Try several fish tracking parameters; choose best by year
Visual Handcount Fish Tracks
Automated “Fish Tracks”
Current data subset
• Cianbro side with first 15 m range, night high slack @de • Eight 15 min files per month; May-‐Oct; 2010-‐2013 • Visual counts with subsamples; compare with tracking • Try several fish tracking parameters; choose best by year
Handcounts Compared to “Fish Tracks” 800
160
y = 2.3093x r² = 0.0490
1
400
120 Fish Tracks
Fish Tracks
600
y = 0.6391x r² = 0.3696
200
2
80
40
0 0
200
400
600
0
800
0
40
Hand Counts
y = 1.2901x r² = 0.4163
160
80 Hand Counts
250
40
Fish Tracks
Fish Tracks
3
80
160
y = 0.9984x r² = 0.2733
200 120
120
150
4
100 50
0 0
40
80
120
Hand Counts
160
0 0
50
100 150 Hand Counts
200
250
Handcounts vs. “Fish Tracks” for Each Year 250
2010
y = 0.9984x r² = 0.2733
150 100
y = 0.6648x r² = 0.3502
400 300 200
50
100 0
0 0
50
250
100 150 Hand Counts
2012
200
0
250
250
100
50
50
0 100
150
Hand Counts
300
400
500
600
200
250
2013
y = 0.754x r² = 0.613
150
100
50
200
200 Fish Tracks
150
0
100
Hand Counts
y = 0.8591x r² = 0.4681
200 Fish Tracks
2011
500
Fish Tracks
200 Fish Tracks
600
0 0
50
100 150 Hand Counts
200
250
2010
Density Gauge
40.0
10.0
20.0
5.0
0.0
0.0
60.0
2011
Density Gauge
40.0
Fish Density
15.0
15.0 10.0
20.0
5.0
0.0
0.0
60.0
2012
Density Gauge
40.0
15.0 10.0
20.0
5.0
0.0
0.0
60.0
2013
Density Gauge
40.0
15.0
10.0
20.0
5.0
0.0
0.0
May
Jun
Jul
Aug
Date
Sep
Oct
Nov
Gauge Height (m)
60.0
30.0
2010
20.0 10.0 0.0
*
30.0
2011
20.0
Fish Density
10.0 0.0 30.0
2012
20.0 10.0 0.0 30.0
2013
20.0
*
10.0
0.0
May
Jun
Jul
Aug
Month
Sep
Oct
Summary: Biological Implica@ons • Peak in May (except 2010; start of gear deployment) • Decrease in density through summer • Density spike at beginning of Fall (except 2013; gear malfunc@ons) • Daily/weekly increases in density follow increases in gauge height • Used as baseline for dam removal
Thoughts/Ques@ons: Biological Implica@ons
Future • Con@nued sampling • Addi@onal subsets of ebbing and flooding @des • Addi@onal range added to analysis • Minimum subset requirements for acceptable precision *doubled sub-‐sample in 2010
Open Rivers Ini@a@ve
Acknowledgements:
Patrick Erbland data collec@on Megan Altenri2er fieldwork Rachel Kocik data processing Haixue Shen data processing
Objec@ve Create metric for seasonal and annual comparisons
Fisheries Acous@cs Non-‐invasive with large spa@al and temporal coverage
Some examples: • Roanoke River; American shad and striped bass (Hightower) • Kenai River; Chinook and sockeye salmon (Burwen) • Chandalar River; chum salmon (Daum)
And the Penobscot
et al.
Split-‐beam Echosounders
“Fish tracks” for quan@fica@on transducer sound beam
Study Area Veazie Dam Site
Hampden
Verona Island
Brewer
Study Site
Sample Site Cross Sec@on Coverage East Cianbro Side
West Mean Low-‐@de
180 m wide x 10 m deep 3 meter @dal flux
Waterfront Side
Data Collected JAN
2010
2011
2012
FEB
MAR
APR
MAY
JUN
JUL
AUG
Waterfront Cianbro Waterfront Cianbro Waterfront Cianbro Waterfront
2013
Cianbro
*some gaps from equipment malfunc@on
SEP
OCT
NOV
DEC
Current data subset
• One side, limited sampling range, certain @dal and diel • Eight 15 minute files per month • Visual counts with subsamples; compare with tracking • Try several fish tracking parameters; choose best by year
Full Data Set JAN
2010
2011
2012
FEB
MAR
APR
MAY
JUN
JUL
Waterfront
AUG
SEP
OCT
Sub-‐set NOV
JAN
DEC
FEB
Waterfront
Waterfront
JUN
JUL
AUG
SEP
OCT
Waterfront Cianbro
Waterfront
2012
Cianbro
2013
MAY
Cianbro
2011
Cianbro
APR
Waterfront
2010
Cianbro
MAR
Cianbro
Waterfront
Waterfront
2013
Cianbro
~80 m
Cianbro
Cianbro, night, high slack @de only 15 m
NOV
DEC
Current data subset
• Cianbro side with first 15 m range, night high slack @de • Eight 15 min files per month; May-‐Oct; 2010-‐2013 • Visual counts with subsamples; compare with tracking • Try several fish tracking parameters; choose best by year
Visual Handcount Fish Tracks
Automated “Fish Tracks”
Current data subset
• Cianbro side with first 15 m range, night high slack @de • Eight 15 min files per month; May-‐Oct; 2010-‐2013 • Visual counts with subsamples; compare with tracking • Try several fish tracking parameters; choose best by year
Handcounts Compared to “Fish Tracks” 800
160
y = 2.3093x r² = 0.0490
1
400
120 Fish Tracks
Fish Tracks
600
y = 0.6391x r² = 0.3696
200
2
80
40
0 0
200
400
600
0
800
0
40
Hand Counts
y = 1.2901x r² = 0.4163
160
80 Hand Counts
250
40
Fish Tracks
Fish Tracks
3
80
160
y = 0.9984x r² = 0.2733
200 120
120
150
4
100 50
0 0
40
80
120
Hand Counts
160
0 0
50
100 150 Hand Counts
200
250
Handcounts vs. “Fish Tracks” for Each Year 250
2010
y = 0.9984x r² = 0.2733
150 100
y = 0.6648x r² = 0.3502
400 300 200
50
100 0
0 0
50
250
100 150 Hand Counts
2012
200
0
250
250
100
50
50
0 100
150
Hand Counts
300
400
500
600
200
250
2013
y = 0.754x r² = 0.613
150
100
50
200
200 Fish Tracks
150
0
100
Hand Counts
y = 0.8591x r² = 0.4681
200 Fish Tracks
2011
500
Fish Tracks
200 Fish Tracks
600
0 0
50
100 150 Hand Counts
200
250
2010
Density Gauge
40.0
10.0
20.0
5.0
0.0
0.0
60.0
2011
Density Gauge
40.0
Fish Density
15.0
15.0 10.0
20.0
5.0
0.0
0.0
60.0
2012
Density Gauge
40.0
15.0 10.0
20.0
5.0
0.0
0.0
60.0
2013
Density Gauge
40.0
15.0
10.0
20.0
5.0
0.0
0.0
May
Jun
Jul
Aug
Date
Sep
Oct
Nov
Gauge Height (m)
60.0
30.0
2010
20.0 10.0 0.0
*
30.0
2011
20.0
Fish Density
10.0 0.0 30.0
2012
20.0 10.0 0.0 30.0
2013
20.0
*
10.0
0.0
May
Jun
Jul
Aug
Month
Sep
Oct
Summary: Biological Implica@ons • Peak in May (except 2010; start of gear deployment) • Decrease in density through summer • Density spike at beginning of Fall (except 2013; gear malfunc@ons) • Daily/weekly increases in density follow increases in gauge height • Used as baseline for dam removal
Thoughts/Ques@ons: Biological Implica@ons
Future • Con@nued sampling • Addi@onal subsets of ebbing and flooding @des • Addi@onal range added to analysis • Minimum subset requirements for acceptable precision *doubled sub-‐sample in 2010
