Iron is not responsible for Didymosphenia ... - Kevin Bladonfews.forestry.oregonstate.edu/publications/Bothwell_CJFAS_2012.pdf · Carole-Anne Gillis, Kevin D. Bladon, and Uldis Silins

RAPID COMMUNICATION COMMUNICATION RAPIDE

Iron is not responsible for Didymosphenia geminatabloom formation in phosphorus-poor rivers

Max L Bothwell Cathy Kilroy Brad W Taylor Eric T Ellison Daniel A JamesCarole-Anne Gillis Kevin D Bladon and Uldis Silins

Abstract Blooms of the river benthic diatom Didymosphenia geminata are an enigma because they occur underphosphorus-poor conditions A recent proposal that ferricndashferrous iron redox shifts sequester the additional phosphorus neededto stimulate and sustain D geminata blooms does not agree with published experimental data showing that blooms onlyoccur when cells are phosphorus-limited The ldquoiron hypothesisrdquo also infers that blooms would be favoured in rivers withelevated iron and management should target iron Surveys of rivers around the world affected by D geminata show thatblooms most often occur in iron-poor rivers Phosphorus uptake experiments conducted under realistic environmentalconditions with living D geminata colonies showed no effect of iron enrichment on phosphorus uptake Iron does notsolve the mystery of D geminata growth causing nuisance blooms worldwide

Reacutesumeacute Les prolifeacuterations de la diatomeacutee benthique de riviegravere Didymosphenia geminata constituent une eacutenigmepuisqursquoelles se produisent dans des conditions drsquoappauvrissement en phosphore Il a reacutecemment eacuteteacute proposeacute que deschangements drsquooxydoreacuteduction du fer ferrique au fer ferreux retiennent le phosphore suppleacutementaire neacutecessaire pourstimuler et soutenir les prolifeacuterations de D geminata Ce postulat ne concorde toutefois pas avec les donneacuteesexpeacuterimentales publieacutees qui deacutemontrent que les prolifeacuterations ne se produisent que quand le phosphore est restreint Ildeacutecellsle eacutegalement de cette laquo hypothegravese du fer raquo que les riviegraveres preacutesentant de fortes teneurs en fer favoriseraient cesprolifeacuterations et que les mesures drsquoameacutenagement devraient cibler le fer Les eacutetudes de riviegraveres affecteacutees par D geminataaux quatre coins du monde montrent que les prolifeacuterations se produisent le plus souvent dans des riviegraveres pauvres en ferDes expeacuteriences de mobilisation du phosphore meneacutees dans des conditions ambiantes reacutealistes avec des colonies deD geminata vivantes nrsquoont montreacute aucun effet drsquoun enrichissement en fer sur la mobilisation du phosphore Le fer nereacutesout pas lrsquoeacutenigme des formes de croissance de D geminata qui entraicircnent des prolifeacuterations neacutefastes partout sur laplanegravete

[Traduit par la Reacutedaction]

Sundareshwar et al (2011) have proposed that complexesbetween iron (Fe) and phosphorus (P) that occur on the poly-saccharide stalks of the benthic diatom Didymosphenia gemi-nata play a central role in the development and maintenance ofblooms in oligotrophic streams In this view ferricndashferrous

redox shifts within mats result in P accumulation and recy-cling providing D geminata with an advantage not availableto competing species The importance of ferricndashferrous Feredox shifts in the cycling of P in lakes has been known sincethe 1940s (Mortimer 1941) and is one of the most important

Received 6 July 2012 Accepted 11 September 2012 Published at wwwnrcresearchpresscomcjfas on 19 October 2012J2012-0309

ML Bothwell Environment Canada Pacific Biological Station Nanaimo BC CanadaC Kilroy National Institute of Water and Atmospheric Research Ltd Christchurch NZBW Taylor Department of Biological Sciences Dartmouth College Hanover NH USA Rocky Mountain Biological LaboratoryCrested Butte CO USAET Ellison Rocky Mountain Biological Laboratory Crested Butte CO USADA Jamesdagger Department of Natural Resource Management South Dakota State University Brookings SD USAC-A Gillis Institut National de la Recherche Scientifique Centre Eau Terre Environnement Quebec QC CanadaKD BladonDagger Department of Natural Resource Sciences Thompson Rivers University Kamloops BC CanadaU Silins Department of Renewal Resources University of Alberta Edmonton AB Canada

Corresponding author Max L Bothwell (e-mail maxbothwelldfo-mpogcca)

Present address Department of Earth and Environmental Sciences Lehigh University Bethlehem PA USAdaggerPresent address US Fish and Wildlife Service Great Plains Fish and Wildlife Conservation Office Pierre SD USADaggerPresent address Department of Renewable Resources University of Alberta Edmonton AB Canada

Can J Fish Aquat Sci 69 1723ndash1727 (2012) Published by NRC Research Pressdoi101139f2012-112

biogeochemical processes influencing productivity of manyfreshwater lakes and some streams However an associationbetween Fe and D geminata dominance of river benthiccommunities is not supported by water chemistry data frommultiple regions experiencing D geminata blooms and surveydata show that elevated P concentrations that can occur ininterstitial waters of D geminata mats do not influence the rateof division of D geminata cells on the periphery of coloniesFurther sorption of Fe and other metallic cations onto algalexopolysaccharide is well known and often occurs in high Penvironments (Lawrence et al 1998) Here we outline evi-dence that FendashP complexes that occur on D geminata stalksare not responsible for the development or the maintenance ofblooms in oligotrophic rivers

Published research from New ZealandExperiments in New Zealand show that D geminata maintains

its dominance of river benthic algal communities while remainingin a highly P-limited state (Bothwell and Kilroy 2011 Kilroy andBothwell 2012) P-limited cell division rates of D geminatawere not influenced by attachment to thick mats of polysac-charide stalk material (Bothwell and Kilroy 2011) Experi-ments document that the elongation of D geminata stalksassociated with blooms preferentially occurs under P-limitedconditions (Kilroy and Bothwell 2011) and blooms did notform when supplemental P was available (Kilroy and Bothwell2012) Surveys around the world find that D geminata bloomsare usually associated with very low P conditions (Bothwell etal 2009) While this could be interpreted as an indication thatD geminata has access to P sources not available to other taxathe application of the frequency of dividing cells technique toD geminata in situ across a spectrum of dissolved reactivephosphorus (DRP) concentrations in rivers on South Island NewZealand indicated that this was not the case (Kilroy and Bothwell2012) Kilroy and Bothwell (2012) found that cell division ratesof D geminata correlated positively with DRP in the overlyingwater and that bloom formation was inversely related to DRPDidymosphenia geminata blooms end when cell division rates arenot limited by P and blooms do not begin unless these rates areP-limited

Fe concentrations in rivers with and withoutD geminata blooms

In natural oxygenated waters Fe exists primarily as a finelydivided precipitate of hydrated ferric hydroxide with particlesizes ranging from filterable (045 m) to colloidal (0001ndash05 m) (Wetzel 2001) For this reason the fraction of totaliron (Fe-T) operationally defined by filtration as dissolved (Fe-D)varies widely depending on the size distribution of ferric hydrox-ide particles and the porosity characteristics of filters In the caseof rivers affected by D geminata on Vancouver Island BritishColumbia the percentage of Fe-T classified as Fe-D spans therange from 35 to 100 with a median of 44 (n 216Supplemental Table S11) Because of this wide variabilitycaution is needed when comparing Fe data Unfortunately insupporting their contention that higher Fe levels in riversmight foster D geminata blooms Sundareshwar et al (2011)compared Fe-T in affected rivers with Fe-D in non-D geminata

rivers (see supplemental table S1 in Sundareshwar et al 2011)Our comprehensive surveys show no consistent pattern be-tween Fe-T or Fe-D in river water and the blooming behaviourof this nuisance alga Moreover it is unclear what if anytemporal changes in Fe or other ions (eg sulphate) haveoccurred at the disparate sites where D geminata is blooming

Many streams in the Black Hills South Dakota have nat-urally high levels of Fe and in those with highest Fe ferrichydroxide precipitate significantly reduces attached algal bio-mass (Holcomb 2002) Fe-T levels in Rapid Creek (mean031 mgmiddotLndash1 n 3 (Sundareshwar et al 2011) versus mean014 mgmiddotLndash1 median 0079 mgmiddotLndash1 n 64 (our 2008ndash2009surveys) Table 1a Supplemental Table S21) the stream af-fected by D geminata that was the focus of the Sundareshwaret al (2011) study are at the lower end of the range for bothmean and median Fe-T values for Black Hills streams (Sup-plemental Table S21) Rapid Creek remains the only stream inthe region affected by extensive D geminata blooms

The absence of association between D geminata colonies andFe-T was also evident within Rapid Creek Although D geminatawas widespread within the stream colonies were not visible at allsites or on all occasions The average Fe-T concentration at sitesand on occasions with colonies was not different from that whencolonies were absent (Supplemental Table S21) LikewiseD geminata standing crop in Rapid Creek quantified using abiovolume index was unrelated to Fe-T concentration of thewater (Supplemental Fig S11) Collectively these results in-dicate no positive relationship between the concentration ofFe-T in river water and D geminata colonies within or amongstreams in the Black Hills of South Dakota

Didymosphenia geminata blooms commencing in 2006 inQuebec spread to numerous rivers along the Gaspe PeninsulaThose rivers eventually supporting blooms had significantlylower levels of Fe-D than those rivers remaining D geminatabloom free (Table 1b Supplemental Table S31) In Quebecrivers alkalinity and pH are related to the distribution ofD geminata blooms (Gillis et al 2010)

Seven rivers on Vancouver Island supporting large D gemi-nata blooms had median Fe-D and Fe-T concentrations of0037 and 008 mgmiddotLndash1 respectively (Table 1c) Mean Feconcentrations for these rivers ranged between 0028 and0058 mgmiddotLndash1 for Fe-D and between 008 and 017 mgmiddotLndash1 forFe-T (Supplemental Table S11) Fe-T data were available intwo rivers both prior to and during D geminata bloom yearsLower Fe-T during blooms in the Puntledge River mightindicate adsorption onto stalk material However such a de-cline was not observed in the Oyster River (Table 1c)

Two remote adjacent Canadian Rocky Mountain streamswere monitored for 8 years as part of a long-term watershedstudy Didymosphenia geminata was not detected in eitherstream until 2010 when blooms commenced in one (NorthYork) Both streams had the same level of Fe-T for the entireperiod of study (Table 1d) Concentrations of Fe-T (median001 mgmiddotLndash1) in these high elevation streams were among thelowest in our surveys and did not differ significantly betweenthe streams affected and unaffected by D geminata (Ta-ble 1d)

On South Island New Zealand Fe-D data were availablefrom several rivers affected by D geminata and their spring-

1Supplementary data are available with the article through the journal Web site at httpnrcresearchpresscomdoisuppl101139f2012-112

1724 Can J Fish Aquat Sci Vol 69 2012

Published by NRC Research Press

Table 1 Summary of iron (Fe) and dissolved organic carbon (DOC) concentrations measured in rivers in Canada USA and New Zealand in relation to the occurrence ofD geminata blooms

No ofriversor sites

Total Fe (mgmiddotLndash1) Dissolved Fe (mgmiddotLndash1) DOC (mgmiddotLndash1)

Samplen Med Max Statistical test

(a) Black Hills South DakotaD geminata affected 1 64 0079 157 MannndashWhitney U test NA NANot affected 6 57 015 801 p 0028

(b) QuebecD geminata affected 12 NA NA 576 0035 160 Two-tailed t test 511 24 500 Two-tailed t testNot affected 17 NA NA 510 0150 413 t 651

p 0001501 48 940 t 617

p 0001

(c) Vancouver IslandD geminata affected 7 567 008 239 271 0037 0370 NA NA NA

Before D geminata Oyster 104 01 196 MannndashWhitney U test NA NA NAAfter D geminata 24 01 043 p 0460 NA NA NA

Before D geminata Puntledge 19 009 096 MannndashWhitney U test NA NA NAAfter D geminata 32 0041 0335 p 0005 NA NA NA

(d) Canadian RockiesD geminata affected North York 118 001 098 Nonparametric sign test NA NA NA 137 105 61 Nonparametric sign testNot affected Star 118 001 036 p 0461 NA NA NA 158 13 79 p 0001

(e) South Island New ZealandD geminata affected 10 NA NA NA 21 0016 0071 MannndashWhitney

U testp 0463

Note g440 values available see SupplementalFig S2Not affected 25 NA NA NA 34 0011 067

(f) Colorado Rocky MountainsD geminata affected 6 24 0079 0315 Two-tailed t test 24 0036 0196 Two-tailed t test 24 62 109 Two-tailed t testNot affected 5 20 0017 0079 t 335

p 000920 0004 0028 t 312

p 001220 26 59 t 213

Note Comparisons are between medians or single values from multiple rivers or sites NA refers to no data available or not applicable Statistical tests used for the comparisons were selectedaccording to the data distribution and type (paired or independent) Complete data sets are in Supplemental Tables S1ndashS51

elletal1725

Publishedby

esearchPress

fed tributaries in 2007 and in 2011 from 19 South Island riversites selected randomly from the New Zealand National WaterQuality Network (Supplemental Table S41) Fe-D spanned awide range and there was no significant difference betweenFe-D concentrations in rivers affected or unaffected byD geminata (Table 1e) None of the sites with blooms hadFe-D 0071 mgmiddotLndash1

The only data set in which D geminata bloom formationwas associated with higher levels of Fe was from a suite of 11Rocky Mountain streams near Gunnison Colorado Sixstreams with D geminata blooms in 2010 had significantlyhigher levels of both Fe-D and Fe-T compared with five othernearby streams without D geminata blooms during the sameperiod (Table 1f Supplemental Table S51) Experiments wereundertaken to determine if higher Fe concentrations mightstimulate P uptake by D geminata colonies

Fe enrichment experiments in the ColoradoRocky Mountains

Sundareshwar et al (2011) used poisoned nonliving seg-ments of D geminata colonies to demonstrate that additions ofFe increase sequestration of P onto D geminata stalks Theyalso used concentrations of Fe and P in their trials that werethree to six orders of magnitude greater than dissolved levelsin typical rivers dominated by D geminata Our P-uptakeexperiments with Fe enrichment were conducted with livematerial on substrates from Rocky Mountain streams incu-bated in containers for 2 h under in situ stream conditions with20 g PmiddotLndash1 sodium phosphate added and with or without Feadditions of 03 mg FemiddotLndash1

Our realistic level of Fe augmentation failed to increase theuptake of P and rates of P uptake were similar betweensubstrates with versus without visible colonies of D geminata(Fig 1a) Biomass-specific P uptake was five times higher onsubstrates without D geminata colonies than on substrateswith colonies and was also not affected by the addition of Fe(Fig 1b) There was also no interaction between Fe andD geminata colonies on P uptake (Fig 1b) Hence using livematerial under environmentally realistic conditions the uptakeof P by D geminata was not stimulated by additional Fe Moreimportantly P removal from the water column (adsorption oruptake) by substrates with D geminata colonies was signifi-cantly lower on a per ash-free dry mass basis than P removalby substrates without D geminata colonies indicating theimportance of cellular uptake rather than noncellular adsorp-tion onto D geminata polysaccharide stalks as has been pro-posed In addition these experimental results also suggest thatthe low P concentration typical of streams in which D gemi-nata blooms (Kilroy and Bothwell 2012) is not a result ofgreater uptake

Reactivity of FeFendashorganic complexes are important in increasing Fe levels

in surface waters but are problematic in predicting Fe reactiv-ity because dissolved organic carbon (DOC) in natural watersis an unknown mixture of compounds with binding constantsfor Fe that are either unknown or unpredictable among naturalwaters (Schecher and McAvoy 2003) The best that can bedone to evaluate the potential for DOC to hinder FendashP com-plexes is to compare DOC concentrations in rivers

Across a broad range rivers dominated by D geminata tendtoward lower DOC (Bothwell et al 2009) It is unclear if thenegative relationship between D geminata blooms and DOCis due to reduced light penetration in highly coloured rivers orthe association with lower pH in more humate-stained watersIn the view of Sundareshwar et al (2011) low DOC mightalso favour D geminata blooms if DOC chelation of Fereduces reactivity and minimizes sequestration of P Howeverwe found little evidence to support the latter argument

In the Colorado Rocky Mountain streams where D gemi-nata blooms were associated with higher concentrations ofFe-T and Fe-D DOC was also twice as high potentiallyoff-setting the higher Fe (Table 1f) Median DOC concentra-tions in two high elevation streams in the Canadian Rocky

Fig 1 Effects of D geminata and iron on water column phos-phorus (a) Removal rates (1 standard error of the mean SE) ofphosphorus (20 g PmiddotLndash1 as sodium phosphate) by substratescovered with visible colonies of D geminata (solid squares) versussubstrates lacking visible colonies of D geminata (open circles) atambient iron concentrations and with iron added (03 mg FemiddotLndash1as ferrous sulfate) did not differ (two-way analysis of variance(ANOVA) F[331] 165 p 019) (b) Phosphorus removal rates(1 SE) normalized to ash-free dry mass (AFDM) on thesubstrates were significantly higher in the absence of D geminatacolonies (two-way ANOVA F[131] 3863 p 00001) but alsonot affected by the addition of iron (two-way ANOVA F[131] 0031 p 086) Uncovered containers with substrates fromWashington Gulch and Coal Creek were incubated at ambientstream temperature and light

Ambient Iron added0

(micro

Ambient Iron added0

(microg

Mountains were the lowest of all streams surveyed (11 and13 mgmiddotLndash1 Table 1d) and although the mean DOC washigher in the stream free of D geminata- compared with thestream affected by D geminata the differences were small(05 mgmiddotLndash1 Table 1d) In New Zealand there was norelationship between the coloured DOC content of river waterand the presence or absence of D geminata blooms (Supple-mental Fig S21) Only the Quebec data set showed lowerconcentrations of DOC in rivers affected by D geminata butthey were also associated with Fe-D concentrations four timeslower than in rivers lacking blooms (Table 1b)

ConclusionValues of Fe-T in natural surface waters typically range

between 005 and 020 mgmiddotLndash1 (Wetzel 2001) We have foundthat D geminata often blooms in rivers at the lower end of thisspectrum which contradicts the inference of Sundareshwaret al (2011) that higher Fe concentrations favour D geminatablooms More importantly quantitative measurements indicatethat D geminata blooms do not rely on P sourced from withinthe mat and uptake rates of P by living D geminata coloniesare not increased by realistic levels of Fe enrichment RatherD geminata blooms occur and are sustained because of aninadequate supply of P This conclusion is consistent withobservations that D geminata blooms disappear in riverreaches immediately downstream of point-source nutrient out-falls in Canadian and US Rocky Mountain rivers (ML Both-well and BW Taylor personal observation) and observationsin South Dakota that nutrients added to a stream curtailedD geminata bloom coverage up to 06 km downstream (Jamesand Chipps 2012)

The global emergence and spread of D geminata is widelyattributed to new introductions While the root causes ofblooms in Northern Hemisphere environments where D gemi-nata is thought to be endemic are unknown the associationbetween D geminata bloom formation and very low P sug-gests increasing intensity and frequency of P limitation inpristine waters perhaps associated with global nitrogen depo-sition or other regional mechanisms of P depletion While thisremains speculation the available data do not support the ideathat Fe is responsible for promoting and sustaining D gemi-nata blooms in P-poor rivers

AcknowledgementsSteven R Chipps (United States Geological Survey South

Dakota State University) provided some of the Fe data fromstreams in the Black Hills of South Dakota Susie Wood(Cawthron Institute) coordinated the Fe analyses for NewZealand rivers in 2011 John Deniseger (BC Ministry of En-vironment) assisted with the compilation of Fe data from theBritish Columbia Environmental Monitoring System databaseThis research was supported by Dartmouth College funds to BWTand a National Science Foundation Research Experience for Under-graduates to ETE through Rocky Mountain Biological Laboratory

P Chambers Y deLafontaine S Chipps and G Scrimgeour pre-viewed the manuscript

ReferencesBothwell ML and Kilroy C 2011 Phosphorus limitation of the

freshwater benthic diatom Didymosphenia geminata determinedby the frequency of dividing cells Freshw Biol 56(3) 565ndash578doi101111j1365-2427201002524x

Bothwell ML Lynch DR Wright H and Deniseger J 2009 Onthe boots of fishermen the history of didymo blooms on Vancou-ver Island Fisheries (Bethesda Md) 34(8) 382ndash388 doi1015771548-8446-348382

Gillis CA Gabor R Cullis J Ran L and Hassan M 2010 Therole of water chemistry and geomorphic control in the presence ofDidymosphenia geminata in Queacutebec [online] Poster presentationAmerican Geophysical Union Fall meeting San Francisco CalifAvailable from httpadsabsharvardeduabs2010AGUFMH41G1162G

Holcomb BM 2002 Nutrient inputs iron availability and algalbiomass in Black Hills watersheds implications for reservoir andstream productivity [online] MSc thesis Wildlife and FisheriesSciences South Dakota State University Brookings South Da-kota Available from httppubstoragesdstateeduwfsthesisHolcomb-Benjamin-M-M-S-2002pdf

James DA and Chipps SR 2012 An evaluation of the efficacy ofwhole-stream phosphorus enrichment to reduce coverage of Did-ymosphenia geminata in an oligotrophic stream [online] Oralpresentation American Fisheries Society annual meeting Minne-apolis Minn Available from httpsafsconfexcomafs2012webprogramPaper9471html

Kilroy C and Bothwell M 2011 Environmental control of stalklength in the bloom-forming freshwater benthic diatom Didymo-sphenia geminata J Phycol 47(5) 981ndash989 doi101111j1529-8817201101029x

Kilroy C and Bothwell ML 2012 Didymosphenia geminatagrowth rates and bloom formation in relation to ambient dissolvedphosphorus concentration Freshw Biol 57(4) 641ndash653 doi101111j1365-2427201102727x

Lawrence JR Swerhone GDW and Kwong YTJ 1998 Naturalattenuation of aqueous metal contamination by an algal mat CanJ Microbiol 44(9) 825ndash832 doi101139w98-083

Mortimer CH 1941 The exchange of dissolved substances betweenmud and water in lakes J Ecol 29(2) 280ndash329 doi1023072256395

Schecher WD and McAvoy DC 2003 MINEQL A chemicalequilibrium program for personal computers Version 46 Envi-ronmental Research Software Hallwell Maine

Sundareshwar PV Upadhayay S Abessa M Honomichl SBerdanier B Spaulding SA Sandvik C and Trennepohl A2011 Didymosphenia geminata Algal blooms in oligotrophicstreams and rivers Geophys Res Lett 38(10) L10405 doi1010292010GL046599

Wetzel RG 2001 Limnology lake and river ecosystems AcademicPress San Diego Calif

Bothwell et al 1727

Supplemental materials for Bothwell et al CJFAS

Supplemental Figure Captions

Figure S1 D geminata biovolume index versus the total iron concentration at sites

within Rapid Creek during 2008-2009 Biovolume index was computed using visually

estimated percent coverage and thickness of the D geminata mat (mm) from one hundred

randomly selected rocks from each sampling riffle Thickness was assigned a score from

0 to 5 based on the following 0 1 (lt 1 mm thick) 2 (1-5 mm) 3 (6-15 mm) 4 (16-30

mm) 5 (gt 30 mm) The coverage of D geminata was multiplied by the thickness to

yield an index ranging from 0 to 500 The biovolume index was unrelated to Fe-T

concentration of the water (linear regression F1 46=209 p=0155 r2=0

Figure S2 Mean optical density of water at 440 nm (g440) a measure of colored

dissolved organic matter (CDOM) plotted against dissolved iron (Fe-D) concentration

measured at 19 South Island New Zealand river sites Filled circles are sites with D

geminata blooms present and open triangles with blooms absent The positive

relationship between CDOM and concentrations of Fe-D was significant (r=0822

plt0001 n=19) but there was no difference in CDOM between streams with and without

D geminata blooms

Figure S1

Figure S2

Supplemental Tables

Table S1 Summary of data on dissolved and total iron concentrations in Vancouver

Island rivers All seven rivers were affected by D geminata blooms from the early 1990s

Fractions of dissolved Fe were calculated from the ratios (as ) of dissolved to total iron

for pairs of samples collected at the same time

Dissolved Fe (mgbullL-1

) Total Fe (mgbullL-1

) Fraction dissolved Fe ()

River Yrs Mean sem n Med Mean sem n Med Min Max n Med

Cowichan 1985-1993 0046 0005 39 0040 0092 0013 53 006 44 100 5 50

Englishman 1996-2010 0558 0522 21 0034 0136 0020 181 0086 60 81 14 35

Nanaimo 1989-2010 0058 0006 74 0044 0115 0012 75 0082 15 100 73 56

Oyster 1980-1993 0056 0008 69 0040 0172 0022 128 01 77 100 69 33

Puntledge 1980-2008 0084 0020 51 0047

Somass 1993-2004 0031 0004 35 0022 0080 0029 47 0031 35 92 27 45

Stamp 1993-1995 0028 0005 37 0020 0113 0037 32 0025 50 92 30 27

Data from the British Columbia Environmental Monitoring System (EMS)

httpwwwenvgovbccaepdwamrems_internetindexhtml

Filter pore size 045microm used to separate dissolved and total Fe

Table S2 Total iron and D geminata bloom status in streams of the Black Hills South

Dakota

River D geminata

blooms

Year Total Fe (mgbullL-1

Mean sem n Median

Castle Creek no 2001 018 002 3 019

Castle Creek no 2008 057 023 10 017

North Fork Castle no 2001 495 155 3 384

South Fork Rapid Creek no 2001 067 019 3 085

North Fork Rapid Creek no 2001 062 015 3 051

Spearfish Creek no 2008-2009 0058 0018 25 004

Whitewood Creek no 2008 044 0098 10 042

Rapid Creek

yes 2008-2009 013 0028 47 008

Rapid Creeksect

no 2008 017 0089 17 007

Total iron analyses conducted by Olson Analytical Services Laboratory at South Dakota State University

Data from Holcomb 2002

sect sampling sites and dates when no D geminata coverage was seen

Comparison of Fe-T during bloom presence vs absence in Rapid Creek Mann Whitney U-test p =0670

Table S3 Summary of dissolved iron and dissolved organic carbon (DOC) data from

Quebec Rivers

Fe-D (mgbullL-1

) DOC (mgbullL-1

River D gem blooms

Years mean sem n med Years mean sem n med

Bonaventure 1 1984-1986 005 012 28 001 1984-1986 2004-2009

11 06 122 10

Cascapeacutedia 1 1984-1986 005 007 35 003 1984-1986 21 13 61 18

Nouvelle 1 1984-1986 003 005 33 003 1984-1986 09 07 79 07

Matapeacutedia 1 1984-1992 007 008 98 004 1984-1990 2004-2009

32 08 198 32

Causapscal 1 1984-1988 1990-1992

008 007 24 006 1984-1988 2004-2009

42 32 42 31

Madawaska 1 1984-1990 006 004 69 005 1984-1988 1993-2009

44 05 267 44

Saint-Jean Gaspeacute

1 1984-1985 003 005 33 001 1984-1986 09 06 56 08

York 1 1984-1985 009 015 24 003 1984-1986 1995-1997

19 11 60 16

Madeleine 1 1984-1986 004 004 14 002 1984-1986 16 10 19 12

Sainte-Anne 1 1984-1986 005 006 30 004 1984-1986 15 08 50 13

Matane 1 1984-1986 011 018 33 004 1984-1986 1995-1997

34 16 75 30

Mitis 1 1984-1986 021 033 32 010 1984-1986 1995-1997

43 53 80 36

Rimouski 0 1984-1986 014 014 32 010 1984-1986 49 12 56 48

Trois-Pistoles 0 1984-1986 012 009 31 010 1984-1986 55 21 53 54

du Loup 0 1984-1986 040 012 36 039 1984-1986 1994-2010

141 8 241 118

Ouelle 0 1984-1986 038 056 32 025 1984-1986 2005-2010

111 44 106 105

Des escoumins

0 1984-1986 047 049 9 026 1984-1986 2004-2009

49 16 57 47

Du sault 0 1984-1986 049 011 9 051 1984-1986 53 16 9 50

Bestiamites 0 1984-1986 038 058 31 017 1984-1986 49 1 52 46

Outardes 0 1984-1990 012 003 93 012 1984-1990 1993-1996

47 06 147 46

Manicouagan 0 1984-1990 010 003 91 009 1984-1990 1993-1996

42 06 141 42

Godbout 0 1984-1986 024 014 9 020 1984-1986 58 13 9 54

Pentecocircte 0 1984-1986 109 117 9 070 1984-1986 67 07 9 68

aux Rochers 0 1984-1986 020 005 27 020 1984-1986 64 13 52 60

Sainte-Marguerite

0 1984-1986 029 01 9 028 1984-1986 65 15 9 68

Moisie 0 1984-1990 028 047 74 019 1984-1990 1993-1996

48 15 130 46

Manitou 0 1984-1986 011 009 8 007 1984-1986 5 07 8 48

Magpie 0 1984-1986 004 002 7 004 1984-1986 36 03 7 34

Saint-Jean 0 1984-1985 073 017 2 073 1984-1986 54 03 2 54

Footnote for Supplemental Table S3

Data on dissolved Fe (Fe-D) and DOC for Quebec rivers from the MDDEP 2012 Banque de donneacutees sur

la qualiteacute du milieu aquatique (BQMA) Queacutebec ministegravere du Deacuteveloppement durable de lrsquoEnvironnement

et des Parcs Direction du suivi de lrsquoeacutetat de lrsquoenvironnement Filter pore size 045microm used for Fe-D and

DOC Fe-D data are primarily from the pre-D geminata bloom period (1984 and 1986) DOC was

monitored consistently from 1984 to the present at most stations Pre- and post-incursion DOC data are

combined for the Matapeacutedia Bonaventure Causapscal Madawaska du Loup des Escoumins and Ouelle

rivers

Table S4 Summary of dissolved iron dissolved organic carbon (DOC) and water colour

data from South Island New Zealand rivers

D gem blooms

) DOC g440

River Year mean sem n med mean sem n med med

Mararoa 2007 yes 0015 0007 5 0010 083 037 5 080

Oreti 2007 yes 0013 0008 3 0010 107 062 3 100

Waitaki (ds) 2007 yes 0055 0023 6 0038 040 018 5 025

Buller (us) 2011 yes 0001 1 036

Clutha 2011 yes 0001 1 006

Hurunui 2011 yes 0003 1 026

Opuha 2011 yes 0032 1 069

Shotover 2011 yes 0016 1 027

Waiau 2011 yes 0071 1 075

Waitaki (us) 2011 yes 0015 1 035

Ashton Flats Spring 2007 no 0010 1

Ferry Rd Spring 2007 no 0037 0021 3 0040 047 027 3 025

Flaxy Creek 2007 no 0020 0014 2 0020 053 037 2 053

Hen amp Chicken Spr 2007 no 0200 1

Oreti Roadend Spr 2007 no 0010 0006 3 0010 025 018 2 025

Otiake Spring 2007 no 0046 0026 3 0047 025 014 3 025

Wakakahi Spring 2007 no 0200 1

Wash Creek 2007 no 0010 0006 3 0010 047 027 3 025

Wilson Road Spring 2007 no 0030 1

Brightwater Creek 2011 no 0001 1

Buller (ds) 2011 no 0077 1 245

Flaxy Creek 2011 no 0001 1

Grey 2011 no 0050 1 260

Mataura (ds) 2011 no 0140 1 147

Mataura (us) 2011 no 0035 1 036

Monowai 2011 no 0007 1 044

Motueka (Gorge) 2011 no 0006 1 028

Opihi (ds) 2011 no 0004 1 038

Opihi (us) 2011 no 0005 1 040

Sutton 2011 no 0290 1 281

Taieri 2011 no 0670 1 378

Upper Oreti 2011 no 0001 1

Waimakariri 2011 no 0009 1 016

Wairau 2011 no 0011 1 012

Wash Creek 2011 no 0001 1

ds us downstream and upstream sites where sampling locations on the same river were far apart

Analyses in 2007 and 2011 had detection limits (dissolved iron) of 002 and 0002 mgbullL-1

respectively

median values for g440 (gilvin calculated from absorbance at 440 nm) from a long-term record

Table S5 Total and dissolved iron concentrations and dissolved organic carbon (DOC)

concentrations in Colorado Rocky Mountain streams

River years

D geminata blooms

(mgbullL-1

Coppersect 2010-2011 Absent 00113 00025 13

East Fork Crystalsect 2010-2011 Absent 00098 00013 59

Poverty Gulchsect 2010-2011 Absent 00299 00123 12

Rustlerrsquos Gulchsect 2010-2011 Absent 00168 00038 31

Upper Eastsect 2010-2011 Absent 00787 00284 26

median 00168

Cement 2010-2011 Present 00569 0011 109

Coal 2010-2011 Present 02195 01054 36

Lower East 2010-2011 Present 00895 00332 77

Oh-Be-Joyful 2010-2011 Present 00569 00391 2

Washington Gulch 2010-2011 Present 03154 0196 62

West Brush 2010-2011 Present 0069 00273 62

median 00793

Present indicates visible colonies of D geminata covering gt 1 m2 of the stream bottom

sect D geminata cells were detected in these streams but not visible colonies of D geminata covering gt 1 m

of the stream bottom

All Fe samples and blanks were acidified with 400 microL of Optima-grade nitric acid Concentrations of field

blanks were 00014 mgbullL-1

total Fe and 00006 mgbullL-1

dissolved Fe

Gelman AE glass fiber filters nominal pore size 1microm were used for Fe-D and DOC samples

Dissolved and total iron concentrations were higher in streams with visible colonies of D geminata

(dissolved iron two-tailed t test t9=312 p lt0012 total iron two-tailed t test t9=335 plt0009)

DOC concentrations were not significantly different between streams with versus without visible

colonies of D geminata (two-tailed t test t9=213 p lt007) and the trend for higher DOC in streams with

visible colonies was driven by one stream (Cement Creek)

Published research from New Zealand

Fe concentrations in rivers with and without D geminata blooms

Fe enrichment experiments in the Colorado Rocky Mountains

Reactivity of Fe

Conclusion

Acknowledgements

References