Size Based Indicators - a Helcom perspective rjan stman (SLU) with - PowerPoint PPT Presentation

Size Based Indicators - a Helcom perspective Örjan Östman (SLU) with much help of Kristiina Hommik (EMI) & Szymon Smoliński (MIR)

Size based indicators of coastal fish • Easy to measure for all species • Precision & Accuracy • Sample sizes needed • Is it possible to identify ’regional’ assessment criteria? • Spatial consistency? • Temporal consistency? • Influence of season and gear? • (Relation to other reference points) • Size based indicators could increase the spatial resolution of assessments for many species

Min size Sizer based indicators Median length investigated Mean Length (mL) Mean and median 0.075 length Length at 90% quantile (L90) N/(sum(N)) Length at 90% Large Fish Index (% above threshold 0.050 length) (LFI) quantile (L90) Large Fish Index Size-spectra 0.025 A minimum size threshold Size spectrum (SS) (slope of to remove influence of log(N)~Length recruitment 0.000 10 20 30 40 50 LengthClass

Data • Data on cod, flounder, perch, pikeperch, and whitefish from around 50 fishery independent surveys from Poland, Estonia, Finland, and Sweden • Commercial data of pikeperch from Estonia and Finland • Different gears, gillnets dominating but fykenets and trawl data as well • Data over the whole ice-free season, summer dominates

Pikeperch • We have data from: Precision generally good at a sample size of  Estonia (commercial + survey) 300 individuals. Size-spectra poor accuracy,  Sweden (survey) up to 500 individuals needed  Finland (commercial + survey) • Threshold values  MinSize = 25 cm  LFI = 40 cm  Q90

Estonia ( ) and Finland SD30 ( ), SD29( ) Spring Estonia Finland

Estonia ( ) and Finland SD30 ( ), SD29( ) Spring

Estonia ( ) and Finland SD30 ( ), SD29( ) Autumn

Estonia ( ) and Finland SD30 ( ), SD29( ) Spring L90 LFI Estonia Finland

Åland ( ), Sweden combined ( ), Kvädöfjärden ( ) and Galtfjärden ( ) * mL median

Åland ( ), Sweden combined ( ), Kvädöfjärden ( ) and Galtfjärden ( ) * LFI L90

Muskö no-take-zone and reference area, years 2009-2016 mL median

Muskö no-take-zone and reference area, years 2009-2016 LFI L90

Conclusions Pikeperch • Mean, median, L90, LFI good precision for 200-300 individuals • Differ between gears, commercial data related to fishing regulations • Regional differences, largest in Estonia – smallest around Ålands Sea • Decreased over time in many areas • A couple cm smaller in autumn than in spring • Not larger in a no-take area

Perch precision 27 L90 mL IndicatorValue • We have data from: IndicatorValue 20 25  Estonia (survey) 19 23  Sweden (survey)  Finland & Åland (survey) 21 18 mL, L90 & LFI good  Poland (survey) precision at 150 1000 2000 1000 2000 100 150 200 250 300 400 500 100 150 200 250 300 400 500 25 50 75 25 50 75 perch, for size as.factor(Threshold) as.factor(Threshold) spectra > 500 • Threshold values needed (not a LFI SS IndicatorValue -0.1 major  MinSize = 15 cm IndicatorValue 0.2 problem as sample  LFI = 25 cm size for perch -0.2 is large).  Q90 0.1 -0.3 0.0 1000 2000 4000 100 150 200 250 300 400 500 25 50 75 1000 2000 4000 100 150 200 300 400 500 50 as.factor(Threshold) as.factor(Threshold)

Perch 23 28 -0.2 0.3 22 26 -0.3 21 Period 0.2 Period Period Period mL L90 LFI SS 2000-2010 2000-2010 2000-2010 2000-2010 2011-2016 2011-2016 2011-2016 2011-2016 24 20 -0.4 0.1 19 22 18 Brunskär Finbo Kvädöfjärden Puck Vilsandi -0.5 BrunskärFinbo Kvädöfjärden Puck Vilsandi Brunskär Finbo Kvädöfjärden Puck Vilsandi WaterBody BrunskärFinbo Kvädöfjärden Puck Vilsandi WaterBody WaterBody WaterBody Different indicators show similar patterns SS stands out a bit

Perch Relative similar between sites Most sites within L90 22-27 cm mL 18-22 cm

Perch 0.4 Most sites within 0.3 0.2 LFI 5-30% 0.1 Size spectra -0.1 -- -0.5 0 Pärnu Vilsandi Brunskär Helsinki Tvärminne Puck Vistula Asköfjärden Forsmark Holmön Kinnbäcksfjärden Kvädöfjärden Lagnö Långvindsfjärden Norrbyn Råneå Torhamn Vinö Finbo Kumlinge Lumparn 2000-2010 - Medel av LFI > 25 2000-2010 - Medel av SS -0.1 2011-2016 - Medel av LFI > 25 2011-2016 - Medel av SS -0.2 Estonia Finland Poland Sweden Åland -0.3 -0.4 -0.5 -0.6

Conclusions Perch • Mean, median, L90, LFI good precision for 150 individuals • No major difference between gears and seasons, commercial data not included • Relative small regional differences, largest in Estonia – smallest in Puck? • Should be possible to set regional assessment criteria

Whitefish precision 46 36 L90 mL • We have data from: 44 IndicatorValue IndicatorValue 35  Estonia (survey) 42 34  Sweden (survey) 40 33  Finland & Åland (survey) 38 mL, L90 & LFI good  Poland (survey) 32 precision at 200 36 100 150 200 300 400 500 600 700 800 100 150 200 300 400 500 600 700 800 50 ind., for size 50 as.factor(Threshold) as.factor(Threshold) spectra > 400 • Threshold values LFI needed (which is a SS 0.3 0.0 high sample size for  MinSize = 25 cm IndicatorValue IndicatorValue whitefish).  LFI = 40 cm 0.2 -0.1  Q90 0.1 -0.2 100 150 200 300 400 500 600 700 800 50 100 150 200 300 400 500 600 700 800 50 as.factor(Threshold) as.factor(Threshold)

Whitefish Southern sites 52 0.6 40 -0.06 48 Period mL Period L90 Period Period 0.4 2000-2010 LFI -0.09 36 SS 2000-2010 2000-2010 2000-2010 2011-2016 44 2011-2016 2011-2016 2011-2016 -0.12 32 0.2 40 Northern sites -0.15 Brunskär Kinnbäcksfjärden Kumlinge Kvädöfjärden NorrbynVilsandi WaterBody Kinnbäcksfjärden Kumlinge Kvädöfjärden Norrbyn Vilsandi Brunskär Kinnbäcksfjärden Kumlinge Kvädöfjärden NorrbynVilsandi Brunskär Kinnbäcksfjärden Kumlinge Kvädöfjärden NorrbynVilsandi WaterBody WaterBody WaterBody Different indicators show similar patterns

Whitefish

Conclusions Whitefish • Mean, median, L90, LFI good precision for 200 individuals • No major difference between gears and seasons, commercial data not included • Smaller in Gulf of Bothnia than Baltic Proper • Should be possible to set assessment criteria per basin

Cod 38 30.0 IndicatorValue IndicatorValue 29.5 36 • We have data from:  Estonia (survey) 29.0 34  Sweden (survey) 28.5  Lithuania? 32 Good 28.0 precision • Threshold values 1000 1400 100 150 200 250 300 400 500 600 700 800 50 at around 30 1000 1400 100 150 200 250 300 400 500 600 700 800 50  MinSize = 25 cm as.factor(Threshold) 300 as.factor(Threshold) 0.20 -0.1 individuals  LFI = 40 cm  Q90 IndicatorValue 0.15 IndicatorValue -0.2 0.10 0.05 -0.3 0.00 1000 1400 1000 1400 100 150 200 250 300 400 500 600 700 800 100 150 200 250 300 400 500 600 700 800 50 50 as.factor(Threshold) as.factor(Threshold)

Cod 45.0 0.20 34 42.5 32 0.15 Period Period Period mL LFI L90 2000-2010 2000-2010 40.0 2000-2010 2011-2016 2011-2016 2011-2016 30 0.10 37.5 28 0.05 35.0 26 Hoburgsbank Litauen Torhamn Vilsandi Hoburgsbank Litauen Torhamn Vilsandi Hoburgsbank Litauen Torhamn Vilsandi WaterBody WaterBody WaterBody Similar pattern between indicators

Cod Evident differences between sites, smaller at the Swedish west coast (likely a gear effect)

Conclusions cod • Mean, median, L90, LFI good precision for 300 individuals • Difference between gears and seasons, commercial data not included • Smaller in Kategatt & Skagerrak than Baltic Sea, but likely a gear effect • Larger in the Sound (SD 23) • Not an evident decrease during 2010’s

Flounder – On Kristiina’s computer  • Very good precision • Difference between areas, smallest along Swedish and Polish coast, largest in central Baltic (SD 25) • Relatively similar from the Strait to Skagerrak (SD21 -23) • I think…

Next • What do spatial/temporal changes in size-based indicators reflect?