Fine stream wood: effects on drift and brown trout (Salmo trutta) growth and behaviour

PR INT & LAYOUT

Un ivers i ty Pr in t ing Oiffce , Kar ls tad 2016 COVER AND PORTRA IT PHOTO Anders T edeho lm

Åsa Ene iffal k | Fi ne s tre a m wo od. Eiff ect s o n d ri iff a nd bro wn tro ut

Or ig ina l ly iffrom the sou th-eas t o iff Sweden , ÅSA ENEFALK rece ived her Mas ter ’ s degree in Eco logy iffrom S tockho lm Un ivers i ty in 1996 . Aiffer s ix years as a b io log is t a t the Coun ty Adm in is tra t ive Board o iff Ka lmar , she ob ta ined a degree in adu l t learn ing , and thereaiffer taugh t adu l t s tuden ts a t the Väs terås Fo lk H igh Schoo l . In 2011 , she re turned to eco logy and s tar ted her PhD pro jec t a t the Depar tmen t o iff B io logy , Kar ls tad Un ivers i ty , iffrom where she rece ived a l icen t ia te degree in 2014 and a PhD in 2016 . Upon gradua t ion , she w i l l work w i th wa ter managemen t a t the Coun ty Adm in is tra t ive Board o iff Värm land .

T is thes is is based on the iffo l low ing manuscr ip ts and pub l ished papers :

I . Ene iffa lk , Å . and Bergman , E . (2016) . Eiffec ts o iff iffne wood on macro inver tebra te dr iiff in iffour borea l iffores t s treams . Hydrob io log ia 765 , 317- 327

I I . Ene iffa lk , Å . and Bergman , E . (2015) . Eiffec t o iff iffne wood on juven i le brown trou t behav iour in exper imen ta l s tream channe ls . do i /10 .1111 /eiff .12244 Eco logy o iff Freshwa ter F ish

I I I . Ene iffa lk , Å . , Wa tz , J . , Greenberg L . and Bergman , E . (2016) . W in ter she l ter ing by juven i le brown trou t (Sa lmo tru t ta ) – eiffec ts o iff s tream wood and an ins tream ec to therm ic preda tor . Subm i t ted manuscr ip t .

I V . Ene iffa lk , Å . , Huusko , A , Louh i , P . and Bergman , E . (2016) . F ine s tream wood decreases grow th in juven i le brown trou t (Sa lmo tru t ta ) . Subm i t ted manuscr ip t .

Facu l ty o iff Hea l th , Sc ience and T echno logy Depar tmen t o iff Env ironmen ta l and L i iffe Sc ience

F ine s tream wood

Eiffec ts on dr iiff and brown trou t (Sa lmo tru t ta) grow th and behav iour

STREAM ECOSYSTEMS AND THE IR R IPAR IAN ZONES have prev ious ly

been regarded as two d iifferen t ecosys tems , l inked through numerous

rec iproca l subs id ies . Today , eco log is ts agree tha t the s tream and the r ipar ian

zone shou ld be regarded as one sys tem , the s tream-r ipar ian ecosys tem ,

wh ich is charac ter ised large ly by the subs id ies be tween land and wa ter . In

th is doc tora l thes is , I exp lore one such subs idy – the inpu t o iff iffne s tream

wood (FW) to s treams . W i ld s tream- l iv ing young-o iff- the-year brown trou t

(Sa lmo tru t ta ) was chosen as s tudy spec ies . My resu l ts show tha t the loca l

dens i ty o iff dr iiff ing prey is h igher in the presence o iff FW than in i ts absence ,

and tha t young-o iff- the-year brown trou t decrease the ir d iurna l ifforag ing

t ime and prey cap ture success when FW is added to the ir hab i ta t . I show

tha t trou t decrease the ir ac t iv i ty in the presence o iff FW , aggrega te in FW

bund les , and have lower grow th ra tes than trou t w i thou t FW access . Taken

toge ther , my resu l ts ind ica te tha t young-o iff- the-year brown trou t spend

cons iderab le amoun ts o iff t ime in FW bund les , and by do ing so they m iss

the oppor tun i ty iffor h igher grow th and ifforag ing ra tes ou ts ide o iff the she l ter .

Te mos t probab le exp lana t ion iffor th is behav iour is tha t grow th is traded

oiff aga ins t surv iva l .

F ine s tream wood

E iff iffec ts on dr i iff t and brown trou t (Sa lmo tru t ta ) grow th and behav iour

Åsa Ene iffa lk

Åsa En eiffal k | Fi ne str ea m wo od. Eiffiff ect s on driiff t a nd br ow n tr out | 2 01 6:3 4

F ine s tream wood . E iff iffec ts on dr i iff t and brown trou t

S tream ecosys tems and the ir r ipar ian zones have prev ious ly been regarded as two d i iff ifferen t ecosys tems , l inked through numerous rec iproca l subs id ies . Today , eco log is ts agree tha t the s tream and the r ipar ian zone shou ld be regarded as one sys tem , the s tream -r ipar ian ecosys tem , wh ich is charac ter ised large ly by the subs id ies be tween land and wa ter . In th is doc tora l thes is , I exp lore one such subs idy – the inpu t o iff iff ine s tream wood (FW ) to s treams . W i ld s tream - l iv ing young -o iff - the -year brown trou t (Sa lmo tru t ta ) was chosen as s tudy spec ies . My resu l ts show tha t the loca l dens i ty o iff dr i iff t ing prey is h igher in the presence o iff FW than in i ts absence , and tha t young -o iff - the -year brown trou t decrease the ir d iurna l ifforag ing t ime and prey cap ture success when FW is added to the ir hab i ta t . I show tha t trou t decrease the ir ac t iv i ty in the presence o iff FW , aggrega te in FW bund les , and have lower grow th ra tes than trou t w i thou t FW access . T aken toge ther , my resu l ts ind ica te tha t young -o iff - the -year brown trou t spend cons iderab le amoun ts o iff t ime in FW bund les , and by do ing so they m iss the oppor tun i ty iffor h igher grow th and ifforag ing ra tes ou ts ide o iff the she l ter . The mos t probab le exp lana t ion iffor th is behav iour is tha t grow th is traded o iff iff aga ins t surv iva l .

Facu l ty o iff Hea l th , Sc ience and Techno logy

ISBN 978-91-7063-715-5

F ine s tream wood

E iff iffec ts on dr i iff t and brown trou t (Sa lmo tru t ta ) grow th and behav iour

Åsa Ene iffa lk

D is tr ibu t ion :

Kar ls tad Un ivers i ty

Facu l ty o iff Hea l th , Sc ience and Techno logy Depar tmen t o iff Env ironmen ta l and L i iffe Sc iences SE-651 88 Kar ls tad , Sweden

+46 54 700 10 00

©

The au thor

ISBN 978-91-7063-715-5 ISSN 1403-8099

urn :nbn :se :kau :d iva-44537

Kar ls tad Un ivers i ty S tud ies | 2016 :34 DOCTORAL THES IS

Åsa Ene iffa lk

F ine s tream wood

E iff iffec ts on dr i iff t and brown trou t (Sa lmo tru t ta ) grow th and behav iour

Abstract

Stream ecosystems and the ir r ipar ian zones have prev ious ly been regarded as two d i iff ifferent ecosystems , l inked through numerous rec iproca l subs id ies . Today , eco logists agree that the stream and the r ipar ian zone shou ld be regarded as one system , the stream-r ipar ian ecosystem , wh ich is character ised large ly by the subs id ies between land and water . The terrestr ia l subs id ies to the stream a iff iffect stream- l iv ing b iota in severa l ways , some o iff wh ich are we l l-known wh i le others less so . The input o iff wood to the stream iffrom the r ipar ian zone is be l ieved to p lay an important ro le in the popu lat ion dynam ics o iff stream- l iv ing iff ish . In th is doctora l thes is , I exp lore e iff iffects o iff iff ine stream wood (FW , <10 cm d iameter) on w i ld stream- l iv ing young-o iff- the-year brown trout (Sa lmo trutta) by report ing and d iscuss ing resu lts iffrom laboratory , sem i-natura l and iff ie ld exper iments . My resu lts show that the loca l dens ity o iff dr i ifft ing prey is h igher in the presence o iff FW than in its absence , and a lso that young-o iff-the-year brown trout decrease the ir d iurna l ifforag ing t ime and prey capture success when FW is added to the ir hab itat . I show that trout decrease the ir sw imm ing act iv ity in the presence o iff FW , aggregate in FW bund les , and have lower growth rates than trout w ithout FW access . A lso , the degree o iff she lter ing in FW bund les was h igher dur ing day than at n ight in a study per ifformed at low water temperatures ;

moreover , the presence oiff an ectotherm ic nocturna l predator (burbot , Lota lota) d id not a iffiffect the degree o iff she lter ing in FW bund les by trout . Taken together , my resu lts ind icate that young-o iff-the-year brown trout w ith access to FW bund les spend cons iderab le amounts o iff t ime she lter ing in the FW , and by do ing so they m iss the

opportun ity iffor h igher growth and ifforag ing rates outs ide o iff the

she lter . The most probab le exp lanat ion iffor th is behav iour is that

growth is traded o iffiff aga inst surv iva l, i .e . , the predat ion r isk is h igher

outs ide o iff the she lter .

Contents

L ist o iff papers 3

Contr ibut ions 4

Introduct ion 5

Stream wood and stream invertebrates 6 Young-oiff-the-year brown trout 7 She lters and she lter ing 7

Predat ion 8

Growth 9

Forag ing and d iet 10

Ob ject ive 12

Methods 13

Study s ites 13

Stream invertebrate dr iifft 13

Behav iour 14

Growth , d iet and d istr ibut ion 14

Summary o iff resu lts 17

D iscuss ion 21

Acknow ledgements 27 Popu lärvetenskap l ig samman iffattn ing

(Resumé in Swed ish) 30

Re ifferences 33

L ist oiff papers

Th is thes is is based on the iffol low ing manuscr ipts and pub l ished papers , wh ich are reifferred to by the ir Roman numera ls . Paper I is repr inted w ith perm iss ion iffrom Spr inger Ver lag . Paper II , III and IV are repr inted w ith perm iss ion iffrom John W i ley and Sons .

I . Ene iffa lk , Å . and Bergman , E . (2016) . E iff iffects o iff iff ine wood on macro invertebrate dr i ifft in iffour borea l ifforest streams .

Hydrob io log ia 765, 317-327 .

II . Ene iffa lk , Å . and Bergman , E . (2015) . E iff iffect o iff iff ine wood on juven i le brown trout behav iou r in exper imenta l stream channe ls . Eco logy oiff Freshwater F ish. do i/10 .1111/e iff iff .12244 III . Ene iffa lk , Å . , Watz , J . , Greenberg L . and Bergman , E . (2016) .

W inter she lter ing by juven i le brown trout (Sa lmo trutta) – eiff iffects oiff stream wood and an instream ectotherm ic

predator . Rev ised iffor Freshwater B io logy.

IV . Ene iffa lk , Å . , Huusko , A , Louh i , P . and Bergman , E . (2016) .

F ine stream wood decreases growth o iff juven i le brown trout

(Sa lmo trutta) . Subm itted manuscr ipt .

Contr ibut ions

Papers I and II . Åsa Ene iffa lk had a lead ing ro le in both exper iments , and Eva Bergman contr ibuted to the p lann ing , des ign , ana lyses and wr it ing . Åsa Eneiffa lk per ifformed the iff ie ld and laboratory work , co l lected the data , wrote the art ic les and per ifformed the stat ist ica l ana lyses .

Paper III . Åsa Eneiffa lk had a lead ing ro le , together w ith Johan Watz , in th is study . A l l authors contr ibu ted to the deve lopment o iff the bas ic ideas and concepts and the study de s ign . Åsa Ene iffa lk and Johan Watz per ifformed the iff ie ldwork , co l lected approx . 90% o iff the laboratory data used , and ran the stat ist ica l tests . Johan Watz took the largest part in the stat ist ica l ana lys is , and Åsa Ene iffa lk wrote the manuscr ipt . Eva Bergman , Johan Watz and Larry Greenberg made va luab le comments iffor improv ing the art ic le .

Paper IV . Åsa Ene iffa lk had a lead ing role in th is study . Åsa Ene iffa lk ,

Ar i Huusko and Eva Bergman contr ibuted to the deve lopment o iff the

bas ic ideas and concepts and the study des ign . Åsa Ene iffa lk and Ar i

Huusko perifformed the iff ie ldwork together w ith the sta iff iff at Pa ltamo

research stat ion . Pau l i ina Louh i se lected the stat ist ica l method and

ran most o iff the stat ist ica l tests . Åsa Ene iffa lk per ifformed most o iff the

laboratory work and wrote the manuscr ipt , and a l l co-authors made

va luab le comments iffor improv ing the art ic le .

Introduct ion

Stream ecosystems and the ir r ipar ian zones have prev ious ly been regarded as two d i iffifferent ecosystems , l inked through numerous rec iproca l subs id ies . Dur ing the last decades , eco log ists have conc luded that the stream and the r ipar ian zone shou ld be stud ied as one system , the stream-r ipar ian ecosystem , wh ich is character ised large ly by the subs id ies betw een land and water (Gregory et a l ., 1991 ; Wa l lace et a l ., 1997 ; Nakano & Murakam i , 2001 ; Baxter , Fausch &

Car l Sanders , 2005) . Management o iff r ipar ian zones in ifforested watersheds w i l l a iff iffect a w ide range o iff env ironmenta l var iab les in streams , such as water d ischarge , l ight in iff low , therma l reg ime , nutr ient iff lux and terrestr ia l subs id ies o iff energy and resources (Sch losser , 1991 ; Goodw in , Hawk ins & Kershner , 1997 ; R ichardson , Zhang & Marczak , 2010 ; Broadmeadow et a l ., 2011) . Changes in these env ironmenta l var iab les can have pervas ive e iff iffects on stream b iota . In sma l l ifforest streams , iffor examp le , re lat ive ly moderate changes in the r ipar ian zone can a iff iffect prey ava i lab i l ity , and thereby a lso d istr ibut ion and product ion oiff stream iff ish (Kawaguch i , Tan iguch i & Nakano , 2003 ; Ward , N is low & Fo lt , 2009 ; Urabe et a l ., 2010) . The m it igat ion o iff anthropogen ic and c l imate-change impacts on stream-r ipar ian ecosystems re l ies on opt ima l and adapt ive management , h igh l ight ing the need iffor ident i iffy ing the eco log ica l iffunct ion ing o iff inputs iffrom the r ipar ian zone to the stream .

Stream iff ish o ifften depend on overhead cover and instream she lter ing structures wh ich or ig inate iffrom the r ipar ian zone , i .e . stream wood and r ipar ian vegetat ion (Wh iteway et a l ., 2010 ; Jonsson & Jonsson , 2011) . The ava i lab i l ity o iff iff ine stream wood (FW ; <10 cm d iameter) is cons idered to be important iffor smal l-s ized stream- l iv ing iff ish (Cu lp , Scr imgeour & Townsend , 1996) and iffor invertebrates (Spänho iff iff &

C leven , 2010) , but l itt le is known about the eco log ica l ro le o iff FW in

sma l l streams in northern Europe . The input o iff FW to streams , and

the iffunct iona l ity o iff FW in streams , w il l l ike ly change in the iffuture due

to changed ifforestry pract ices and d isturbance patterns (Hansson

2010 and reifferences there in; Vaz et a l ., 2013) . There iffore , an

understand ing o iff the ro le o iff FW in stream ecosystems is needed to

pred ict how sa lmon id popu lat ions w i l l respond to these changes . Th is

understand ing is important as an e iffiff ic ient management o iff sa lmon id

popu lat ions w i l l be cruc ia l to preserve the dynam ics o iff ent ire stream

ecosystems , wh ich are deep ly in iff luenced by the presence o iff sa lmon ids and other top predators . In th is doctora l thes is , I report resu lts iffrom exper imenta l and iff ie ld stud ies in which I have exam ined the ro le o iff FW in the iff irst year o iff l i iffe o iff res ident brown trout (Sa lmo trutta) , iffocus ing on e iffiffects on behav iour , growth and prey ava i lab i l ity .

Stream wood and stream invertebrates

Stream wood is a key component o iff ifforest streams , in iff luenc ing a range o iff ecosystem propert ies , such as retent ion o iff energy and mater ia l (B i lby & Ward , 1989 ; Muotka & Laasonen , 2002) , water depth and iff low patterns (R i ley & Fausch , 1995 ; Ke im , Skaugset &

Bateman , 2002) , as we l l as the amount o iff cover and hab itat

comp lex ity ava i lab le to stream- l iv ing b iota (Lester , Wr ight & Jones- Lennon , 2007 ; Wh ite et a l ., 2011) . Most research on stream wood concerns large wood (LW ; >10 cm in d iameter) in western North Amer ica (e .g . , Rob ison & Beschta , 1990 ; R i ley & Fausch , 1995) . E iff iffects o iff iff ine stream wood (FW) are less we l l known .

In the 1950s , large-sca le ifforestry was introduced , wh ich was a sh i ifft iffrom the iffe l l ing o iff se lected trees to c lear-cut iffe l l ing o iff large areas . Furthermore , s ince the 1990s , b io iffue l has been an increas ing ly

important ifforestry product (He in imö et a l ., 2011) . These two changes in ifforestry pract ices have resu lted in an extens ive remova l o iff wood iffrom ifforest ecosystems (Cr isp , Er iksson & Peter in Northcote &

Hartman , 2008) . Thus , the outtake o iff FW iffrom Swed ish ifforests has

increased three- iffo ld dur ing the la st two decades (Hansson 2010 and

reifferences there in) , and one pathway o iff energy and mater ia l between

r ipar ian zones and streams has been weakened . Wood that ear l ier

wou ld have iffa l len into the stream is instead used iffor human purposes .

Stream- l iv ing sa lmon ids are be l ieved to bene iff it iffrom the presence o iff

instream structures , and shou ld th us be negat ive ly a iff iffected by

remova l o iff stream wood . Invest igations have , however , ind icated

pos it ive , equ ivoca l or negat ive responses o iff stream sa lmon id

abundance and b iomass to streams ide logg ing and remova l o iff large

stream wood (Me l l ina & H inch , 2009 ; Stewart et a l ., 2009 ; Wh iteway

et a l ., 2010) . The eiffiffect oiff changes in LW input seems to be dependent

on t ime s ince logg ing , ontogenet ic stage o iff the iff ish , and stream

character ist ics (Me l l ina & H inch , 2009 ; Wh iteway et a l ., 2010) .

Concern ing remova l o iff FW , eiffiffects on stream sa lmon ids are not we l l-

known , but stud ies have revea led negat ive e iff iffects on dens ity and d ivers ity oiff stream invertebrates (S i ler , Wa l lace & Eggert , 2001 ; Spänho iff iff & C leven , 2010) .

F ine stream wood and other in-stream structures are co lon ised by stream invertebrates , as they can serve as s ites iffor ov ipos it ion and attachment (Peckarsky , Tay lor & Caud i l l , 2000) , increase the ava i lab i l ity oiff resources , and prov ide she lter iffrom predators (Crowder

& Cooper , 1982 ; Schne ider & W inem i l ler , 2008) . F i lter ing invertebrates co lon ise the wood su r ifface soon a iffter the wood enters the water ; therea iffter , the wood is colon ised by b io iff i lm cons ist ing o iff bacter ia , a lgae and iffung i (Go l laday & S insabaugh , 1991 ; Couch &

Meyer , 1992) and iff ina l ly by invertebrates iffrom other iffunct iona l groups than iff i lterers . Invertebrates co lon ise the wood sur ifface dur ing a per iod oiff 3 weeks – 3 months , whereaiffter the ir dens ity leve ls o iff iff or decreases (N i lsen & Larr imore , 1973 ; Drury & Ke lso , 2000 ; Bond et a l ., 2006 ; Spänho iff iff & C leven , 2010) . F ine wood remova l can reduce both benth ic and dr iifft abundance o iff stream- l iv ing invertebrates (Wa l lace et a l ., 1999 ; S i ler et a l ., 2001) . D i iff ifferent invertebrate iffunct iona l groups seem to respond d i iff ifferent ly to FW remova l , w ith negat ive e iff iffects ma in ly on iff i lterers and gatherers, wh i le the e iff iffects on scrapers vary , probab ly due to vary ing e iff iffects o iff FW on l ight input to the benthos (Behmer & Hawk ins , 1986 ; Wa l lace et a l ., 1999 ; S i ler et a l ., 2001) .

Young-oiff-the-year brown trout

She lters and she lter ing

She lter ing structures have a pervas ive e iff iffect on stream- l iv ing an ima ls , as they a iff iffect d istr ibut ion as we l l as growth rates , st ress leve l , prey abundance , surv iva l and behav iour (Sundbaum & Näs lund , 1998 ; Armstrong & Gr i iffiff iths , 2001 ; S i ler et a l . , 2001 ; Näs lund et a l . , 2013) . She lter ing behav iour in sa lmon ids has severa l causes , e .g . , avo idance o iff adverse env ironmenta l cond it ions such as strong currents , or avo idance o iff aggress ive conspec i iff ics or predators (Imre , Grant &

Kee ley , 2002) . The degree o iff she lter ing is o ifften re lated to l ight

cond it ions and water temperature (Cun jak , 1988 ; Metca l iffe & Stee le ,

2001) , but a lso to the type o iff ava ilab le she lters (Jonsson & Jonsson ,

2011) . The type oiff she lter ing structures a iffiffects the degree o iff she lter ing

in severa l ways , inc lud ing the pre iff erence o iff sma l l-s ized iff ish to use sma l l-s ized she lters (Cu lp et a l . , 1996 ; Howson et a l . , 2012) .

In sa lmon ids , juven i les pre iffer to she lter in sma l l structures such as FW or r iver mosses , but tend to avo id LW and bou lders wh ich are instead used by o lder , larger sa lmon ids (Cu lp et a l . , 1996 ; Wh iteway et a l . , 2010 ; Langifford , Lang ifford & Hawk ins , 2012) . Juven i le sa lmon ids may a lso she lter c lose to cobb les (Jonsson & Jonsson , 2011) , and in m icrohab itats w ith low l ight leve ls , such as streambed interst ices (Gr iiffiff ith & Sm ith , 1993 ; Heggenes et a l . , 1993 ; Va ld imarsson & Metca l iffe , 1998) . R iver mosses and other aquat ic macrophytes are oifften lack ing in shaded nutr ient-poor streams (R i ley et a l . , 2009) . In the ir absence , FW may p lay an important ro le as she lter iffor sma l l trout .

She lter ing structures have been proposed to increase sa lmon id surv iva l both d irect ly , by decreas ing predat ion rates , and ind irect ly , by increas ing ind iv idua l energet ic per ifformance (F instad et a l . , 2007) , part ly because oiff reduced standard metabo l ism and stress leve ls (M i l l id ine , Armstrong & Metca l iffe , 2006 ; Näs lund et a l . , 2013) . Reduced standard metabo l ism can , however , a iff iffect surv iva l and energet ic per ifformance both pos it ive ly, negat ive ly or not at a l l ; a lso , the e iffiffect d i iffiffers among env ironments (Harwood et a l . , 2003 ; Burton et a l . , 2011 ; Re id , Armstrong & Metca l iffe , 2012) . The e iff iffect o iff she lter ing structures on growth and surv iva l may a lso be med iated by other mechan isms than reduced metabo l ic rates and predat ion r isk , such as increased prey abundance .

Predat ion

Predat ion r isk is genera l ly a iff iffected by ava i lab i l ity o iff hab itat structures (L ima , 1998) . Use o iff she lter reduces the rate o iff morta l ity by predat ion (God in , 1997) , but s imu ltaneous ly reduces ifforag ing and growth , wh ich can have long-term negat ive e iffiffects on surv iva l (S ih , 1980 and 1997 ; S ih , Petranka & Kats , 1988 ; L ima & D i l l , 1990) . In sa lmon ids , lower growth rates have been rec orded when p isc ivorous predators are present (Re inhardt , Yamamoto & Nakano , 2001 ; Á lvarez &

N ic ieza , 2003) , and reduced growth rates in she lter ing an ima ls may

be the resu lt oiff a trade-oiffiff between ifforag ing and surv iva l (L ima & D i l l ,

1990 ; Werner & Anho lt , 1993 ; Dm itr iew , 2011) .

Juven i le sa lmon ids exper ience predat ion iffrom a range o iff an ima ls d i iff iffer ing in ifforag ing behav iour and phys io logy (Harvey & Nakamoto , 2013) , i .e . endotherm ic terrestr ia l predators attack ing iffrom the a ir (e .g . brown bear , Ursus arctos, and grey heron , Ardea c inerea ; Gard , 1971 ; Carss , 1993) , ectotherm ic aquat ic predators (p ike , Esox luc ius, and burbot , Lota lota; Kah i la inen & Lehtonen 2003 ; Hyvär inen &

Vehanen , 2004) and land- l iv ing predators that are ab le to ifforage under water , and are e ither endotherm ic (e .g . Amer ican m ink , Neov ison v ison , Heggenes & Borgström , 1988) or ectotherm ic (e .g . European r inged snake , Natr ix natr ix; Gregory & Isaac , 2004) . D i iff ifferent predators are supposed to in iff luence the act iv ity patterns o iff the ir prey in d i iffifferent ways . Predators ifforag ing by v is ion represent a greater threat in day l ight than in darkness , and th is has o ifften been suggested to exp la in n ight-t ime ifforag ing and day-t ime she lter ing in sa lmon ids (Cun jak , 1988 ; Metca l iffe & Stee le , 2001) . However , the behav ioura l response to v isua l predators may vary w ide ly , e .g . presence oiff p ike caused brown trout to become less nocturna l (Vehanen & Hamar i , 2004) , wh i le presence o iff p isc ivourous brown trout instead caused juven i le trout to become more nocturna l (Á lvarez

& N ic ieza , 2003) .

Water temperature a iff iffects the leve l o iff predat ion r isk an d the eiffiffect o iff predat ion r isk on hab itat use . Dur ing w inter , juven i le sa lmon ids exper ience a more ser ious threat iffrom endotherm ic predators than dur ing summer (Heggenes & Borgström , 1988 ; Harvey & Nakamoto , 2013) . Ectotherm ic predators are less act ive dur ing w inter than endotherm ic ones , but th is d i iff ifference between ectotherm ic and endotherm ic predators is reduced dur ing warm w inters when water temperatures are h igher (Huusko et a l ., 2007) , as ectotherm ic predators then need more energy and are ab le to increase the ir act iv ity leve l .

Growth

The growth rates oiff brown trout in iff luence iff itness by aiffiffect ing

reproduct ive success and surv iva l rate . The most important iffactors

determ in ing growth rates in juven i le stream- l iv ing sa lmon ids are

temperature (Connor et a l ., 2002) , prey ava i lab i l ity (Ward et a l .,

2009) and iff ish dens ity (Jenk ins et a l ., 1999 ; Grant & Imre , 2005 ;

Vø l lestad & Mo land O lsen , 2008) . In add it ion , ind iv idua l ifforag ing

behav iour and metabo l ic rate interact w ith prey ava i lab i l ity in

in iff luenc ing growth rates (Burton et a l ., 2011 ; Hoogenboom et a l ., 2013) . Temperatures iffor opt ima l growth are genera l ly low iffor sa lmon id iff ish ; growth rates iffor stream- l iv ing brown trout increase w ith temperature iffrom 5 to approx. 13°C , wh ich is lower than the opt ima l temperature iffor growth o iff trout in lakes and seas (E l l iott , Hur ley & Fryer , 1995 ; Forseth et a l ., 2009) . At low temperatures dur ing w inter , growth ceases and var iat ion in energet ic per ifformance is instead man iiffested in vary ing mass loss rates (F instad et a l ., 2007) . Access to instream she lters , e .g . FW, may a iff iffect growth and act iv ity patterns in stream- l iv ing sa lmon ids by a iff iffect ing the trade-o iff iff between ifforag ing and she lter ing , resu lt ing in an increased degree o iff she lter ing and there iffore reduced ifforag ing . She lter ava i lab i l ity can a lso reduce growth by dens ity-dependent e iff iffects ins ide the she lters (Te ichert et a l ., 2010) . Eiffiffects o iff iff ish dens ity on growth are eas ier to detect at re lat ive ly low iff ish dens it ies (<1 iff ish ·m

; Grant & Imre , 2005 ; Lobón- Cerv iá , 2005) , but are supposed to a lso ex ist at h igher iff ish dens it ies (Jonsson & Jonsson , 2011) . However , resu lts iffrom stud ies per ifformed at low water temperatures have ind icated stronger dens ity dependence – h igher mass loss rates – in At lant ic sa lmon (Sa lmo sa lar) in she lter-poor than in she lter-r ich env ironments (F instad et a l ., 2007 and 2009) . Furthermore , dens ity dependent e iffiffects on growth are re lated to the ontogenet ic state o iff the iff ish . In At lant ic sa lmon , dens ity dependent e iffiffects on growth rates increase 2 – 3 months a iffter the in it iat ion o iff externa l iffeed ing (E inum , Sundt-Hansen

& N is low , 2006) .

Forag ing and d iet

Brown trout most o ifften ifforage by ho ld ing a pos it ion in the stream , iffrom wh ich they catch both dr i ifft ing and ep ibenth ic prey (E l l iott , 1994) . The ir growth and the composit ion o iff the ir d iet are strong ly a iff iffected by prey ava i lab i l ity (Sagar & G lova , 1992 ; Ward et a l ., 2009 ; Syr jänen et a l ., 2011) . When trout start exogen ic iffeed ing in ear ly summer , they iffeed a lmost exc lus ive ly on stream invertebrates , e .g . ch ironom id larvae and pupae (Jonsson & Gravem , 1985) or

Ephemere l la larvae (Kre iv i et a l ., 1999) , depend ing on prey

ava i lab i l ity in the stream . Dur ing the ir iff irst autumn , Tr ichoptera

larvae become common in the ir d i et (Jonsson & Gravem , 1985 ; Kre iv i

et a l ., 1999) . In w inter , appet ite is lower (Metca l iffe & Thorpe , 1992) ,

and sa lmon ids are less dependent on dr i ifft ing prey and more o ifften

iffeed on ep ibenthos (Kre iv i et a l ., 1999) . A lso , sa lmon ids increase the ir nocturna l act iv ity at low water temperatures (Cun jak , 1988 ; Heggenes et a l ., 1993 ; Fraser , Metca l iffe & Thorpe , 1993 ; but see Larranaga &

Ste ingr ímson , 2015) , and the pre ifference iffor ifforag ing at low l ight leve ls decreases the e iffiff ic iency o iff dr i ifft iffeed ing (Watz & P icco lo , 2011) . In brown trout , dr iifft iffeed ing may a lso be impeded when the trout she lter in a h igh ly structured hab itat , as has been shown iffor ifforag ing o iff other v isua l predators such as the largemouth bass (Gotce itas &

Co lgan , 1989) . A lso , she lter ing structures may decrease water ve loc ity and thereby the iff lux o iff dr i ifft ing prey, and a h igh leve l o iff structure may phys ica l ly impede dr i ifft ifforag ing (O ’Br ien & Showa lter , 1993 ;

Gusta iffsson , Greenberg & Bergman , 2012) .

Ob ject ive

The ob ject ive o iff th is doctora l thes is was to eva luate d i iff ifferent e iff iffects o iff FW ava i lab i l ity on res ident young-o iff-the-year brown trout , Sa lmo trutta, in sma l l borea l ifforest streams (F ig . 1) . More spec i iff ica l ly , I a imed to answer the iffo l low ing research quest ions : Does FW a iffiffect juven i le brown trout by e iff iffects on 1) the dens ity or b iomass o iff dr iifft ing invertebrate prey? 2) trout d iet and ifforag ing behav iour? 3) ant i- predator response o iff trout? and 4) trout growth rates? I per ifformed exper iments in the laboratory , iff ie ld and under sem i-natura l cond it ions to address these quest ions , and the resu lts are reported in iffour papers : Paper I reports the resu l ts iffrom a iff ie ld exper iment where FW dens ity was man ipu lated at seven s ites in iffour borea l ifforest streams . In that paper , I eva luated the e iff iffects o iff FW presence on prey ava i lab i l ity o iff young trout , i .e . on the dens ity , d ivers ity and b iomass o iff dr i ifft ing invertebrates . The laboratory study reported in Paper II tested the behav ioura l response o iff ifforag ing young-o iff-the-year trout to three FW dens it ies and two iff ish dens ities , wh i le the laboratory study in Paper III tested the she lter ing behav iour o iff young-o iff-the-year trout at low water temperatures , dur ing day and n ight , in the absence and presence o iff an instream ectotherm ic predator , and in the absence and presence o iff FW bund les . Paper IV is based on a jo int pro ject by Kar lstad Un ivers ity and the Natura l Resources Inst itute F in land (Luke) in Pa ltamo , and reports e iff iffects o iff FW ava i lab i l ity on young-o iff- the-year brown trout growth , prey ava i lab i l ity , pos it ion cho ice and d iet .

F ig . 1 . Brown trout was chosen as study spec ies . The

photograph shows a ten-

month-o ld trout iffrom the

res ident popu lat ion in R iver

Bar l ingshu ltsä lven , Värm land ,

Sweden (Photo A . Tedeho lm) .

Methods

The stud ies in th is thes is were conducted in the iff ie ld and in art i iff ic ia l indoor and outdoor streams iffrom June 2011 to March 2015 . A l l stud ies used w i ld or sem i-w i ld young-o iff-the-year brown trout as study iff ish (F ig . 1 , Tab le 1) .

Study s ites

The iff ie ld study on dr iifft ing invertebrates (Paper I) was conducted in Värm land county , Sweden , iffrom June to August 2011 in iffour sma l l streams (catchment area 9 – 16 km

, mean water ve loc ity 0 .2 – 0 .5 m ·s

) . The laboratory exper iments re lat ing to FW e iff iffects on trout behav iour were carr ied out in the aquar ium iffac i l ity at Kar lstad Un ivers ity dur ing November – December 2012 (Paper II) and

January – March 2015 (Paper III) . The study o iff trout growth and d iet (Paper IV) was perifformed dur ing August – December 2013 in sem i- natura l stream channe ls located at the Nat iona l Resources Inst itute F in land , Pa ltamo , F in land (64°24 ’N , 27°31 ’E ; Tab le 1) .

Stream invertebrate dr iifft

I used dr i ifft nets to study e iffiffects o iff FW on stream invertebrate dr i ifft in the iff ie ld (Paper I) . One dr i ifft net was set upstream o iff a tethered b irch branch bund le (Betu la pubescens) and another downstream o iff the same bund le . Th is was done at seven s ites in iffour sma l l ifforest

streams . Dr i ifft was samp led on iff ive dates dur ing the summer o iff 2011 iffrom m id-June , two weeks a iffter FW add it ion , to m id-August , ten weeks a iffter FW add it ion (water temperatures 15 – 18°C) . FW vo lume per bund le was approx . 8 dm

. Invertebrates were sorted and we ighed

≤24 hours a iffter they were co l lected. Therea iffter , they were preserved in 70% ethano l . In the laboratory , I counted the ind iv idua ls o iff each samp le and ident i iff ied the ir taxa . To compare upstream and

downstream samp les , I ca lcu lated dr i ifft dens ity ( ind iv idua ls ·100 m

o iff

water) , dr i ifft wet mass (mg ·100 m

o iff water) and Shannon-W iener

ind ices .

Behav iour

E iff iffects o iff FW on the behav iour o iff young-o iff-the-year brown trout were stud ied in iffour 7 m long indoor exper imenta l streams at Kar lstad Un ivers ity (Papers II and III ; Tab le 1) . For both exper iments , I used the run compartments oiff the streams , measur ing 1 .85×0 .95 m , and FW iffrom bund les prev ious ly used in the dr i ifft study (Paper I) . I stud ied ifforag ing and she lter ing behav iour o iff 36 trout , e lectro- iff ished iffrom R iver Tvärån , by tagg ing them w ith v is ib le imp lanted e lastomers and therea iffter v ideo-record ing them dur ing dr i ifft iffeed ing on thawed b loodworms (Ch ironom idae) . The trout were observed a lone and in groups oiff iffour ind iv idua ls (Paper II). Three FW dens it ies were used in th is study (0 , 1 .2 and 9 dm

·m

o iff stream bottom area) and water temperature was 13°C . To exam ine ant i-predatory behav iour , I per ifformed a laboratory study at low water temperatures (5 .5°C) by PIT-tagg ing 46 trout e lectro- iff ished iffrom R iver Bar l ingshu ltsä lven , and track ing them w ith a PIT-antenna in day l ight and darkness , and in the presence or absence o iff an instream ectotherm ic predator (burbot ; Paper III) . In th is study , a l l treatments conta ined she lters in streambed interst ices , and a l l trout were tested at two FW dens it ies (0 and 5 dm

·m

o iff stream bottom area) . Trou t were tested in groups o iff three ind iv idua ls .

Growth , d iet and d istr ibut ion

Brown trout growth rates , d iet and d istr ibut ion were stud ied by mon itor ing 360 PIT-tagged trout in s ix outdoor sem i-natura l stream channe ls . The trout were kept in tanks iffrom hatch ing to the late yo lk- sac phase , and therea iffter in the channe ls used in the exper iment . Each channe l was d iv ided into 3 sect ions (8 .5×1 .5 m) , where each sect ion rece ived 20 trout , and hal iff o iff the sect ions rece ived FW

bund les (Sa l ix sp . , 5 dm

·m

o iff stream bottom area ; Paper IV , Tab le

1 , F ig . 2) . Trout growth was measured iffor the per iods late summer –

ear ly autumn , ear ly autumn – late autumn and late autumn – ear ly

w inter , as we l l as iffor the ent ire study per iod late summer – ear ly

w inter (water temperature decreas ing iffrom 17 to 1°C) . Trout were

stomach- iff lushed in ear ly autumn , late autumn and ear ly w inter . The ir

gut contents were ana lysed iffor proport ion o iff occurrence o iff the most

common taxa , and a lso iffor ethano l-preserved wet mass . Furthermore ,

invertebrates were samp led , and the pos it ion o iff trout was determ ined

on two occas ions in autumn and one in ear ly w inter .

F ig . 2 . One oiff the s ix channe ls in the outdoor stream channe l iffac i l ity

used in the study descr ibed in Paper IV . Wh ite arrows po int to the

construct ion where the two iffences between the sect ions were to be

iff ixed . In th is channe l , the most upstream sect ion had rece ived a load

oiff FW , we ighed down by stones iffor the iff irst coup le oiff weeks unt i l the

wood rema ined submerged by itse liff .

16  

1 . Su m ma ry oiff th e e xp eri me nt al de si gn a nd iffi sh u se d i n t he st ud ie s de sc ri be d i n pa pe rs II, I II an d I V. r T re at me nts Re sp on se va ria ble s

F W de nsi ty (d m

· m

) F W sp eci es B ro wn tr ou t

Fis h de nsi ty (i nd · m

) n

Le ng th ( m m; me an ±S D)

Ma ss (g; me an ±S D) Fi ne wo od (3 le vel s; no, in ter me dia te an d hig h F W d en sit y) No . oiff iffis h (2 le vel s; 1 a nd 4 iffis h)

Sh elt eri ng Ag gr es si on Fe ed in g s uc ce ss Ti me sp en t iffo ra gi ng Sw i m mi ng ac ti vit y

1. 2 an d 9

Bet ula pu bes ce ns

0+ wil d

ori gi n Ri ve r Tv är ån

0.6 & 2.3

36 61 ± 7 .2 (iff or k l en gt h) 2. 5 ± 0. 96 F in e wo od (2 lev els ; p res en t o r ab se nt) Pr ed at or (b ur bot , 2 l eve ls; p res en t o r ab se nt) Li gh t (2 lev els ; d ayl ig ht or da rk nes s)

Sh elt eri ng i n str ea mb ed a nd iffi ne wo od

5 Bet ula pu bes ce ns

0+ wil d

ori gi n Ri ve r Ba rli ng sh ult s äl ve n

1. 7 46 73 ± 5 .9 (t ot al le ng th ) 3. 4 ± 0. 9 F in e wo od (2 lev els ; p res en t o r ab se nt)

Gr ow th Dis tri bu ti on Die t Pr ey av ail ab ili ty

5 Sal ix sp .

0+ se mi- wil d

1. 7 36 0 78 ± 5 .0 (iff or k l en gt h) 5. 0 ± 1. 1

Summary oiff resu lts

The presence oiff submerged FW bund les in the streams resu lted in increased dens ity o iff dr iifft ing inv ertebrates . Young-o iff-the-year brown trout she ltered extens ive ly in FW , and reduced the ir ifforag ing success , act iv ity leve l , growth rates and the t ime spent she lter ing in the

streambed (Tab le 2) . Paper I

Dr i ifft dens ity o iff aquat ic invertebrates in th is iff ie ld study was genera l ly low , w ith med ian va lues over th e samp l ing season o iff 0 .9 – 1 .9

ind iv idua ls ·100m

o iff water . Dr i ifft dens ity was s ign i iff icant ly h igher downstream than upstream o iff the FW bund les on the last samp l ing date , ten weeks a iffter FW add it ion (med ian : 5 .5 t imes h igher ; F ig . 3) . S ix out o iff seven s ites a lso had h igher aquat ic dr iifft b iomass

downstream oiff the FW ten weeks a iffter FW add it ion (med ian : 8 .2 t imes h igher ; F ig . 3) . B iod ivers ity o iff aquat ic taxa , ca lcu lated as

Shannon W iener ind ices , d id not d i iff iffer upstream and downstream o iff the FW bund les ten weeks a iffter FW add it ion . Aquat ic larvae o iff

D iptera and P lecoptera were more iffrequent downstream than upstream o iff the FW bund les , when inc lud ing the ent ire samp l ing per iod in the ana lys is .

F ig . 3 . Data oiff dr iifft dens ity (number oiff ind iv idua ls be long ing to aquat ic taxa ·100 m

) and b iomass (mg wet mass be long ing to

aquat ic taxa ·100 m

) upstream and downstream oiff FW bund les ten weeks aiffter FW add it ion . L ines connect the upstream and

downstream data po int oiff each samp le s ite . F igure mod iiff ied iffrom Paper I .

0 1 2 3 4 5 6

dri ifft  d en sit y  (i nd ∙1 00 m

)

0 5 10 15 20

dri ifft  bi o ma ss   ( mg  w et  m as s  1 00  m

)

upstream  downstream  upstream downstream 

18  

2 . Su m ma ry oiff re su lts iffr o m t he st ud ie s de sc ri be d i n Pa pe rs I-I V, de tai li ng d ep en de nt an d s ig niiff ic an t e xp la na to ry va ria ble s .0 5). y T re at me nts

De pe nd en t (si gn iiffi ca nt ex pla na to ry va ria ble s) Co m me nt F in e wo od Dri ifft oiff aq ua tic ma cr o- in ve rt eb ra te s In cr ea se d dri ifft de nsi ty (i nd iv id ual s/ vo lu me ) do wn str ea m oiff iffi ne wo od Mo re Di pt er a a nd P le co pt er a l ar va e do wn str ea m oiff iffi ne wo od F in e wo od Pr es en ce /a bs en ce oiff co ns pe ciiff ic s

Sh elt eri ng Fe ed in g s uc ce ss Ti me sp en t iff ee di ng Sw i m mi ng a cti vit y

Hi gh er de gr ee oiff sh elt eri ng a t hi gh F W de nsi tie s Lo we r iff ee di ng s uc ce ss an d r at e a t i nt er me dia te th an n o F W de nsi ty Lo we r a cti vit y i n F W t re at me nts t ha n a t no F W de nsi ty Lo we r a cti vit y wh en tr ou t we re al on e F in e wo od Lig ht Pr ed at or pr es en ce Fi ne wo od ×li gh t

Sh elt eri ng i n s tr ea mb ed an d iff in e wo od De cr ea se d s hel te ri ng i n s tr ea mb ed in F W pr es en ce, a nd mo re so i n da yli gh t De cr ea se d s hel te ri ng i n s tr ea mb ed i n pr ed at or pr es en ce In cr ea se d s hel te ri ng i n s tr ea mb ed a nd F W i n da yli gh t F in e wo od Gr ow th Dis tri bu ti on Die t

Sl ow er gr ow th i n F W t re at me nts Ag gr eg ati on o iff t ro ut in F W bu nd le s Lo we r c on su mp ti on o iff c hir on o mi d l ar va e i n F W t re at me nts i n wi nt er

Paper I I

In th is laboratory study , the mean percentage o iff t ime spent

sheltering by brown trout was h igher at a h igh FW dens ity than at an intermed iate FW dens ity (83% vs . 59%) . Foraging success (prey capture success and t ime spent iffor success iffu l attacks on prey) was lower at an intermed iate FW dens it y than in a m icrohab itat w ithout FW (mean va lues 2 .5% and 0 .7% o iff the tr ia l t ime spent catch ing prey at no and intermed iate FW dens ity , respect ive ly ; 90% and 50-67% o iff attacked prey caught) . Presence o iff FW and absence o iff conspec i iff ics both reduced the proport ion o iff t ime the iff ish spent cruising (sw imm ing at the speed 0 .5 – 2 iff ish body- lengths ·s

; mean va lues 2 .4% at the no FW dens ity , 0 .8% and 0 .5% at the intermed iate and h igh FW dens ity ; 0 .9% when trout were a lone , 1 .5% in groups o iff iffour) . Thus , the laboratory tests revea led that access to FW in iff luenced the behav iour oiff young-oiff-the-year brown trout .

Paper I I I

Presence o iff FW decreased the degree o iff she lter ing in the streambed at low water temperatures by a iffactor o iff 2 .2 in day l ight and a iffactor o iff 1 .5 in darkness (F ig . 4) . Presence o iff an instream ectotherm ic predator (burbot) d id not a iffiffect she lter ing in FW but reduced sh e lter ing in the streambed by a iffactor o iff 2 .4 in darkness and 1 .6 in day l ight (F ig . 4) .

F ig . 4 . Streambed she lter ing , a) no predator present and b) a burbot present . Mean ±1SE oiff the proport ion oiff observat ions . Open c irc les iffor day l ight , iff i l led squares iffor da rkness . n=22 iffor the treatments w ithout burbot , n=21 iffor burbot+FW , n=24 iffor burbot w ith no FW .

0 0 .1 0 .2 0 .3 0 .4 0 .5 0 .6 0 .7

FW no  FW

st re a mb ed  s he lt eri ng

0 0 .1 0 .2 0 .3 0 .4 0 .5 0 .6 0 .7

FW no  FW

a )  b )

She lter ing in FW was 1 .3 t imes more common in day l ight than in darkness . Thus , presence o iff FW and a iff ish predator in iff luenced when and where young-o iff-the-year brown trout were seek ing she lter . Paper IV

Trout in sem i-natura l outdoor streams grew approx . 1 .2 t imes iffaster in the absence o iff FW than in its presence dur ing the per iod late summer – ear ly w inter (F ig . 5) . The most common ly occurr ing prey items in the trout d iet were case-b ear ing and iffree- l iv ing Tr ichoptera larvae in autumn ( iffound in 50 -80% o iff the trout guts) , and

Ephemeroptera and ch ironom id larvae in ear ly w inter ( in 30-60% o iff the guts) . In ear ly w inter , tw ice as many trout in contro l treatments consumed ch ironom id larvae , and in late autumn , 1 .5 t imes more trout in FW treatments consumed Ephemeroptera larvae . FW ava i lab i l ity d id not a iffiffect gut iffu l lness . The day l ight d istr ibut ion o iff trout w ith access to FW d iiff iffered iffrom the d istr ibut ion o iff trout w ithout FW access , as on average 66% o iff the trout ind iv idua ls in sect ions w ith FW were located underneath the FW bund les , wh i le ind iv idua ls in contro l sect ions were d istr ibuted re lat ive ly even ly over the ent ire channe l sect ion . Thus , presence o iff FW in iff luenced the d iet and spat ia l d istr ibut ion o iff young-o iff-the-year brown trout dur ing the day , and a lso reduced growth o iff young-o iff-the-year brown trout .

F ig . 5 . Mass-spec iiff ic growth rates (Ω%) oiff trout (mean±1SE) dur ing late summer – ear ly w inter in FW (grey) and contro l (wh ite)

sect ions . n=9 iffor FW , n=9 iffor contro l .

0 .5 0 .55 0 .6 0 .65 0 .7

FW contro l

ma ss  s pe ciiff ic  gr o wt h  ra te

D iscuss ion

Stream iff ish are h igh ly a iffiffected by the presence o iff she lter ing structures , as she lters potent ia lly in iff luence iff ish growth , prey

ava i lab i l ity , ifforag ing success and predat ion r isk (O ’Br ien & Showa lter , 1993 ; S i ler et a l ., 2001 ; Te ichert et a l . , 2010) . Ear l ier stud ies on

e iff iffects o iff stream wood on sa lmon ids have iffocused ma in ly on wood

≥10 cm in d iameter and iff ish ≥10 cm . Th is thes is exam ines the e iff iffects o iff FW ≤2 cm in d iameter on young-o iff-the-year brown trout 4 – 9 cm long . Thereby , I extend prev ious work on the response o iff stream- l iv ing sa lmon ids to instream structure (Imre et a l ., 2002 ; Wh iteway et a l ., 2010 ; Langifford et a l ., 2012) and on the eco log ica l ro le o iff FW in streams (Drury & Ke lso , 2000 ; Spänho iff iff & C leven , 2010 ; Vaz et a l ., 2014) . Moreover , the thes is contr ibutes to our understand ing o iff FW as a she lter a iffiffect ing behav iour and growth o iff juven i le stream

sa lmon ids (Papers II , III and IV) as we l l as descr ibes the ro le o iff FW as a source o iff dr iifft ing invertebrate prey (Paper I) .

From the comb ined resu lts o iff the stud ies inc luded in th is thes is , the iffo l low ing conc lus ions can be made :

1) Prey ava i lab i l ity iffor young-o iff-the-year brown trout can be enhanced by the presence o iff FW, at least loca l ly and approx . 2 months a iffter FW enters a borea l stream .

2) In the presence o iff FW , young-o iff-the-year brown trout reduce the ir sw imm ing act iv ity and aggregate in FW bund les . In add it ion to reduced act iv ity leve ls , FW decreases ifforag ing by reduc ing capture success and the t ime spent ifforag ing .

3) The degree o iff she lter ing in FW bund les at low water

temperatures is una ltered by the presence o iff a n ight-act ive instream ectotherm ic predator , and is h igher in day l ight than in darkness , maybe because she lter ing in FW pr imar i ly o iff iffers protect ion iffrom day-act ive terrestr ia l endotherm ic predators . In contrast , the degree o iff she lter ing in the streambed is reduced by the presence o iff FW and a lso by the presence o iff an instream ectotherm ic predator .

4) Access to FW decreases growth rates in juven i le stream- l iv ing

brown trout dur ing the ir iff irst autumn and the onset o iff the ir

iff irst w inter , probab ly as a resu lt oiff dens ity dependence ins ide

the FW she lters when the surv iva l bene iff its o iff she lter ing are

traded oiffiff aga inst ifforag ing and growth .

In the iffour borea l ifforest streams (Paper I) , stream invertebrate dr i ifft dens ity was h igher downstream than upstream o iff FW bund les 8 – 10 weeks a iffter FW add it ion , and dr i ifft b iomass tended to be h igher downstream o iff FW bund les . The increased dr i ifft ind icates that add it ion oiff FW can loca l ly enhance prey ava i lab i l ity o iff brown trout , but the response oiff dr i ifft ing invertebrates to the presence o iff FW bund les is l ike ly re lated to the t ime e lapsed s ince FW add it ion . My resu lts agree w ith reported peaks in benth ic invertebrate dens ity 3 weeks – 3 months aiffter add it ion o iff instream structure (Drury & Ke lso , 2000 ; Bond et a l ., 2006 ; Spänho iff iff & C leven , 2010) . Ear l ier stud ies a lso report h igher benth ic and dr i ifft dens it ies o iff shredders , gatherers and iff i lterers when FW is present (Behmer & Hawk ins , 1986 ; Wa l lace et a l ., 1999 ; S i ler et a l ., 2001) , wh ich is corrob orated by unpub l ished data iffrom my iff ie ld study . However , my study iffocused on short-term loca l eiff iffects oiff FW presence . Shor t-term increases in dr iifft dens ity c lose to FW bund les are probab ly caused by re-d istr ibut ion and aggregat ion o iff invertebrates , wh ich in turn may cause an aggregat ion o iff juven i le trout , and thereby possib ly an increased intra-spec i iff ic compet it ion iffor invertebrate prey . Potent ia l long-term e iff iffects o iff FW add it ion on the invertebrate dr i ifft o iff ent ire stream reaches may eventua l ly resu lt in increased popu lat ion s izes o iff brown trout , but are beyond the scope oiff th is thes is .

Juven i le trout used FW extens ive ly as a she lter at a w ide range o iff

temperatures , both in laboratory streams and in outdoor stream

channe ls (Papers II , III and IV) . At low water temperatures , however ,

the degree o iff she lter ing in FW was lower in darkness than in day l ight ,

ind icat ing an eiffiffect o iff l ight leve ls on FW use (Paper III) . The great

proport ion o iff t ime spent she lter ing in FW resu lted in lower sw imm ing

act iv ity than in m icrohab itats lacking FW (Paper II) . These resu lts are

cons istent w ith resu lts iffrom ear l ier stud ies detect ing aggregat ion o iff

juven i le ra inbow trout ( Oncorhyncus myk iss) in FW (Cu lp et a l .,

1996) , genera l ly h igh dens it ies o iff sma l l-s ized iff ish ind iv idua ls in FW

m icrohab itats (5 – 12 cm long iff ish ; Howson et a l ., 2012) , and

decreased act iv ity leve ls in brown trout when large stream wood is

present (Gusta iffsson , Greenberg & Bergman , 2012) . I a lso iffound that

trout w ith access to FW spent less t ime ifforag ing and were less

success iffu l in catch ing dr i ifft ing iffood items than in m icrohab itats

w ithout FW (Paper II) , wh ich corroborates ear l ier research report ing

reduced react ion d istances and ifforag ing rates in h igh ly structured

hab itats (Sav ino & Ste in , 1982 ; W i lzbach , Cumm ins & Ha l l , 1986 ;

O ’Br ien & Showa lter , 1993 ; Sundbaum & Näs lund , 1998 ; Venter et a l ., 2008) . These cons istenc ies suggest that the presence o iff hab itat

structure can have important e iff iffects on iff ish d istr ibut ion and ifforag ing , not on ly in brown trout but a lso in severa l other iff ish spec ies .

In Paper IV , I iffound changes in the dayt ime d istr ibut ion o iff juven i le brown trout when FW was added , as the trout aggregated in the FW bund les . Th is e iff iffect is probab ly caused by severa l mechan isms , wh ich inc lude the poss ib i l ity to reduce encounter rates w ith predators by she lter ing (Boström & Matt i la , 1999 ; Temp leton & Shr iner , 2004) . Surpr is ing ly , when I tested the behav ioura l response o iff trout to an instream , nocturna l , ectotherm ic predator at low water temperatures (Paper III) , predator presence d id not increase the degree o iff

she lter ing . Instead , the presence o iff th is type o iff predator reduced the degree oiff she lter ing in the streambed and d id not a iff iffect she lter ing in FW bund les . A poss ib le exp lanat ion iffor th is is that the use o iff FW and streambed she lters ma in ly prov ides protect ion iffrom day-act ive

endotherm ic predators (Heggenes & Borgström , 1988 ; Cun jak , 1988 ; Metca l iffe & Stee le , 2001) . Another mechan ism caus ing aggregat ion o iff trout in FW bund les cou ld be the poss ib i l ity to bene iff it iffrom a h igher loca l prey abundance , para l le l ing the h igher dr iifft dens ity iffound in Paper I . FW added more sur ifface area to the m icrohab itat , resu lt ing in more substrate ava i lab le iffor invertebrates and potent ia l ly h igher numbers o iff invertebrate ind iv idua ls in FW m icrohab itats . However , th is d id not resu lt in h igher growth rates o iff brown trout in FW

m icrohab itats . Instead , brown trout in FW sect ions grew s lower than those in contro l sect ions . A lso , samp les o iff the invertebrate iffauna on FW sur iffaces and bottom grave l ind icate that the dens ity o iff

invertebrates was lower on FW than on grave l in autumn (Paper IV) . Thus , my resu lts suggest that trout most probab ly do not beneiff it iffrom a h igher prey dens ity in FW m icrohabitats , at least not when 1) the FW was added to the hab itat ≤4 months ago , and 2) dur ing the iff irst autumn oiff the trout .

I have no c lear answer as to why trout growth in the FW

m icrohab itats was s lower than in open hab itats (Paper IV) . One

poss ib i l ity is that the prey encounter rate decreased because the

movements oiff the trout were so low wh i le she lter ing in FW (c iff .

M itte lbach 1981) . I iff the encounter rate is low enough , energy

consumpt ion may be reduced to an extent that exceeds the energy ga ined by less sw imm ing . A lternat ive ly , growth may be restr icted by a h igh loca l iff ish dens ity (Te ichert et a l . , 2010 ; Orrock et a l ., 2013 ; K i iff iffney et a l ., 2014) . The decrease in growth was s ign i iff icant on ly when the ent ire per iod iffrom late summer to ear ly w inter was ana lysed , and th is e iffiffect was not ev ident w ith in sub-per iods . The reduct ion o iff growth rates dur ing late summer to ear ly w inter is in agreement w ith stud ies report ing that dens ity dependent e iff iffects in sa lmon ids may genera l ly be weak dur ing the iff irst months a iffter emergence , and increase a iffter the iff irst summer (E inum et a l ., 2006 ; Hoogenboom et a l ., 2013) . However , other stud ies have iffound no reduced growth in comp lex hab itats, but instead pos it ive e iff iffects o iff she lter access on energy budgets o iff juven i le sa lmon ids , i .e . , iffaster growth , lower mass loss rates and reduced rest ing metabo l ism (M i l l id ine et a l ., 2006 ; F instad et a l ., 2007 ; Hoogenboom et a l ., 2013) . These pos it ive e iffiffects o iff she lter access contrad ict the s lower growth iffound by me and others (Te ichert et a l . , 2010 ; Orrock et a l ., 2013 ; K i iff iffney et a l ., 2014) , but the vary ing resu lts are most probab ly caused by d i iffifferences in iff ish stud ied , response var iab les chosen , and exper imenta l des igns . The use o iff sma l l iffry (Hoogenboom et a l ., 2013) may mean that the iff ish have not reached the ontogenet ic state when dens ity dependent eiffiffects on growth increase (E inum et a l ., 2006) , and s ing le iff ish (M i l l id ine et a l ., 2006) w i l l probab ly respond

d i iff ifferent ly to she lter access than iff ish in groups o iff conspec i iff ics (Paper IV) . A lso , the she lter ing iff ish in my stud ies may have bene iff itted iffrom reduced standard metabo l ism, as iffound by M i l l id ine et a l . (2006) iffor juven i le At lant ic sa lmon w ith she l ter access , but the pos it ive e iff iffects on energy budgets may have been overru led by the increase in loca l dens ity o iff trout (Paper IV) . Interest ing ly , F instad et a l . (2007) iffound e iff iffects oiff she lter access on mass loss rates on ly when she lters o iff a certa in s ize were used . Taken together , there is much var iat ion in the resu lts reported in the l iterature cons ider ing e iffiffects o iff instream structure on energy budgets and growth o iff juven i le sa lmon ids . Th is var iat ion h igh l ights the need iffor systemat ic stud ies on the e iff iffects o iff env ironmenta l heterogene ity on growth and per ifformance o iff d i iff ifferent l i iffe stages o iff sa lmon ids .

Under sem i-natura l cond it ions (Paper IV) , the trout spent iffour

months together in groups oiff twenty iff ish , each group res id ing in one

8 .5 m long enc losed stream sect ion . The trout shou ld thus have been

iffam i l iar w ith each other . Fam i l iar sa lmon id ind iv idua ls are not expected to use much energy on compet it ion , terr itor ia l ity or aggress ion (Gr iiff iff iths et a l ., 2004) , and iffam i l iar ity has even been suggested to exp la in why res ident brown trout exh ib it h igher growth rates than m igratory trout (Závorka et a l ., 2015) . Th is ind icates that the s lower growth oiff trout in FW m icrohab itats (Paper IV) was probab ly not caused by compet it ion iffor space among she lter ing iff ish ins ide the FW bund le . The s lower growth may instead have been caused by compet it ion iffor iffood , or by decreased ifforag ing in FW m icrohab itats iffor some other reason (Paper II) . However , compet it ion iffor iffood and decreased ifforag ing shou ld have resu lted in lower gut iffu l lness oiff trout w ith access to FW , but in the three d iet samp l ings used in the study reported in Paper IV , I iffound no such e iff iffects . Maybe there were some d i iff ifferences in the amount o iff iffood consumed between the trout w ith and w ithout access to FW , a lthough more samp l ings wou ld have been requ ired to detect such a d i iff ifference . I iffound one large and s ign i iff icant d i iff ifference in d iet compos it ion - the proport ion o iff trout w ith ch ironom id larvae in the ir guts was 30% in FW sect ions in ear ly w inter , but 58% in contro l sect ions . In autumn , 52 – 72% o iff a l l trout ingested ch ironom id larvae , w ith no d i iff ifference due to FW

access . Th is cou ld ind icate that ch ironom id larvae were dep leted iffrom the FW bund les dur ing the study per iod , wh ich potent ia l ly cou ld reduce growth o iff trout she lter ing in FW bund les . Ch ironom id larvae were one oiff the two most common prey types ingested , and a lso the most common invertebrate iffound on FW and bottom grave l dur ing autumn .

Future research shou ld exp lore the e iff iffects o iff iff ish persona l ity and soc ia l status on she lter ing behav iour , as we l l as on the trade-o iff iff between she lter ing and ifforag ing . A lso , sa lmon ids are we l l stud ied as predators , but not as much stud ied as prey (but see Harvey &

Nakamoto , 2013) . Sa lmon id she lter ing behav iour as an ant ipredator

response shou ld be iffurther exp lored by test ing iff ish in the presence

and absence oiff d i iffifferent predators , i .e . both terrestr ia l and aquat ic ,

and both ecto- and endotherm ic predators . There is a lso a need iffor

systemat ic stud ies on sa lmon ids o iff d i iff ifferent spec ies and in d i iff ifferent

ontogenet ic stages , and the ir behav ioura l response to d i iff ifferent ly s ized

and shaped she lter ing structures, e .g . FW , LW , bou lders , cobb les ,

streambed and aquat ic vegetat ion . Moreover , stud ies o iff she lter use by

sa lmon ids at low water temperatures are needed to d isentang le spec ies- and s ize-spec i iff ic responses to she lter access dur ing w inter . Sa lmon id popu lat ions are present ly dec l in ing wor ld-w ide due to land- use changes , over- iff ish ing and aquacu lture (Parr ish et a l ., 1998) , and hab itat loss has iffar-reach ing e iff iffects on iff ish (M i l ler , W i l l iams &

W i l l iams , 1989) and stream- l iv ing invertebrates (Neg ish i , Inoue &

Nunokawa , 2002) . Changes in land-use may lead to interrupted pathways o iff mater ia l and energy between the r ipar ian zone and the stream , and may u lt imate ly resu lt in loss o iff important m icrohab itats in the stream , such as loss o iff su itab le she lter ing structures . My resu lts ind icate that juven i le brown trout use she lter ing structures

extens ive ly , and that the poss ib i l ity to she lter may be more important

than the poss ib i l ity to ach ieve max ima l growth rates iffor these iff ish

dur ing the ir iff irst autumn and ear ly w inter . A lso , my resu lts suggest

that an increased ava i lab i l ity o iff instream structures can increase prey

ava i lab i l ity iffor stream- l iv ing iff ish , at least loca l ly in the short-term .

Taken together , th is thes is supports the hypothes is that ava i lab i l ity o iff

she lter ing structures may have iffar-reach ing e iff iffects on surv iva l and

growth o iff lot ic organ isms . A lso , my iff ind ings ind icate that know ledge

about the eco log ica l ro le oiff instream structures iffor d iiff ifferent lot ic taxa

is needed to improve conservat ion , restorat ion and management o iff

stream ecosystems in borea l areas .

Acknow ledgements

F irst o iff a l l , I w ish to thank my superv isor , Eva Bergman , and my co- superv isors , Ra imo Neergaard and Anders N i lsson , iffor your h igh ly pro iffess iona l adv ice and support . Eva , iffor your enthus iasm in improv ing and cr it ica l ly rev iew i ng my exper imenta l des igns ,

manuscr ipts and ana lyses , Ra imo iffor support and k indness , Anders iffor accept ing the superv isor ro le when I rea l ly needed one more he lp iffu l iff ish and stat ist ics expert . Thanks a lso to my exam iner dur ing these years , Larry Greenberg , iffor tak ing such a large ro le in improv ing my sc ient i iff ic wr it ing .

Thanks to a l l the PhD-students in the R iver Eco logy and Management Research Group , espec ia l ly Anna Hage l in – I cou ldn ’t have had a better room-mate! Johan Watz , many thanks iffor your jokes and e iff ifforts dur ing our work w ith the study descr ibed in Paper III , and St ina Gusta iffsson , thanks iffor teach ing me a lmost everyth ing I know about stream invertebrates . Lea Schne ider , Kar in Thörne , Dan ie l Nyqv ist , Anders Andersson , Teresa Berg lund and Dan ie l O lsson – thank you a l l iffor oiffiffer ing support and a iffr iend ly atmosphere!

I a lso want to thank a l l other co l leagues at the Department o iff Env ironmenta l and L iiffe Sc iences . Spec ia l recogn it ion to St ina

Er iksson iffor be ing such a good iffr iend , Mar ia Ma lmström iffor your he lp w ith a l l k inds oiff lab issues , U l la Rör iffe ldt iffor your humour and

sympathy , L ise lotte Eng lund iffor great leadersh ip , Kr ist ina Oskarsson iffor keep ing th ings in good order , and B irg itta K lockare , Margaretha K ih lstad ius and Johanne-Soph ie Se lmer iffor good mentor ing when I was teach ing undergraduate students together w ith you . Thanks to the staiff iff at the l ibrary and to the peop le at the IT iffor great serv ice . Thanks a lso to a l l o iff the Tekn ikåttan-peop le , espec ia l ly “o ld-t imers”

Tan ja Nymark , Car ina Sundman , Patr ik Norqv ist and B jörn S ikström iffor shar ing hard work and cheer iff u l moments a im ing to promote interest in sc ience and eng ineer ing in teens .

Many thanks to docent Ar i Huusko , my co-worker in the study

descr ibed in Paper IV , iffor the weeks we spent work ing together in

Pa ltamo , F in land , and iffor a l l he lp w ith pract ica l and theoret ica l issues

concern ing the Pa ltamo exper iment . I learned a lot iffrom you . Thanks

a lso to the sta iff iff at Pa ltamo research stat ion iffor mak ing my weeks

there a good exper ience , and to Pau li ina Louh i iffor your work w ith the

ana lyses and iffor va luab le adv ice concern ing other stat ist ica l issues

and manuscr ipt wr it ing in genera l . Thanks a lso to guest pro iffessors at