Journal of the American Mosquito Control Association,g(3):275-284, 1993 Copyright @ 1993 by the American Mosquito Control Association, Inc. CHIRONOMID (DIPTERA: CHIRONOMIDAE) LARVAL OCCURRENCE AND TRANSPORT IN A MUNICIPAL WATER SYSTEM'

E. C. BAY

Washington State University, Research and Extension Center Puyallup, WA 9837l-4998

ABSTRACT. Chironomid larval populations contaminating a municipal water systemwere monitored over a 24-month period betweenJune 1987and June I 989. Iarval drift within distribution lines occurred throughout the year, and larval suspension was uniform at all depths within open reservoirs. Draining and cleaning reservoirs to rid them of chironomid populations was of queslionable value becauselarval populations recovered fully within weeks, and larval drift occurred throughout the system.

INTRODUCTION 3-basin, 757,000-kl lead distribution reservoir, Portland Avenue, and North End reservoirs.All Chironomid midge larvae in potable water were of complete concreteconstruction, approx- constitute an aestheticnuisance, but not a public imately 7 m deep, with sloping 30o sides sur- health problem. Larvae have never been found rounded by I m-high safety barriers (Fig. lA). to be diseasevectors, or otherwise injurious to In addition, a system of wires, several meters man (Gerardi and Grimm 1982, Ali l99l). These apart, was installed over each reservoir to dis- larvae, however, can cause serious public rela- courage bird activity. Bottom sediments rarely tions concernsfor municipal water suppliers.Such exceeded I cm deep. These features precluded an incident occurred in the city of Tacoma, WA. the useofconventional dredgesand bottom sam- in the fall of 1986,when a prisonerin the city's pling techniques. new jail found a red "worm" in his drinking Standinglarval populations were monitored in water. 2 ways. One was by setsof tethered, submerged, All but 2 of Tacoma's 5 city reservoirsare now sampling plates designedto monitor larval col- enclosed,but in 1986, the systemwas entirely onization by tube formation, and the other was open. Complaints of larvae in Tacoma's tap wa- by dragnet. Plateswere l5 cm2 placed uniformly ter have historically averaged4-6 per year. Prior in sets of 4 around each reservoir perimeter ap- to this study, the standard responseofthe water proximately 3 m beyond the water's edge.Plates department to larval complaints was to drain and were cut from a cement-like sheetof Flexboard@ scour the suspectedsource reservoir. The jail in- (Johns Manville Corp.) and marked with a 28- cident prompted the city of Tacoma to commis- mm grid to facilitatelarval and tube counts.Plates sion this researchto determine the magnitude of were inspectedand replaced weekly. the problem, identifu and evaluate source res- Weekly bottom tows were made with a cus- ervoirs, and recommend possible preventive tom-designed, 500-pm pore mesh, dragnet pre- measuresagainst larval encounter. cededby a weighted 500-g squareframe of spiral nylon brushes (Fig. 2). The l5-cm2 net opening MATERIALS AND METHODS collected larval-containing sediments dislodged by the advancing contact brush. Samples were Three primary midge larval monitoring pro- regarded as spatially and seasonally relative, cedureswere employed over a 24-month period rather than absolute. Four tows were made at betweenJune 1987 and June 1989. These in- eachreservoir in the vicinity of the tetheredplates cluded bottom sampling, larval drift and strata previously mentioned. Tow sampleswere made sampling. by casting the net apparatusfor a distance of l0 Four reservoirs were initially selectedfor this m, then retrieving it as a drag for 3 m. At 3 m study, but sampling in one, Alaska Street Res- the net wasjerked offofthe bottom and recov- ervoir, was discontinuedin July 1988,when it ered. Samples were concentrated into 120-ml was dismantled for reconstructionas an enclosed volumes of reservoir water for transport. tank reservoir. Others included McMillin. a In the laboratory,0.5 g of NaSO, was added to each sample for 5-10 min to distress larvae ' This researchwas supported in part by a grant from into evacuating their feeding tubes. Larvae were the city of Tacoma, WA, Public Utilities Water Di- then simultaneously stirred and bubbled into vision. suspension,from which one 20-ml aliquot was 275

SegrsNrsnR.1993 CmnoNovrn Lenver TRANSPoRT 277

Fig.2. Dragnet for collecting benthos from concrete surfaces.The weighted brush dislodgessediments rn advance ofthe net opening. receptaclewas a 100-liter vinyl barrel, and the vented larvae from concentrating against and secondaryreceptacle was a 2 1-liter covered plas- working through the mesh.Collected larvae were tic food storagecontainer. The primary overflow almost always alive, or at least intact. Overflow guard was a 500-pm pore mesh, l4-cm wide, to the secondarycontainer was filtered through submergedscreen through which water entered a submerged,finer meshbag made from a section from beneath.This configuration evenly distrib- of woman's nylon stocking. Water dischargedto uted filtrate, reducedoverflow velocity, and pre- waste through a floor drain. Upon weekly col-

rl =: ii !_+------

ar =:-p i: f------.- [ / - l=t L ,fj\ I /!.\ I i|'l I \ Bi I \.-.-'

Fig. 3. Drift collection device. Most adrift larvae are retained in the primary receptacle@anel) by an inverted, submerged,overflow screen(A). Pass-throughlarvae are captured by a submergedfilter bag (B). 278 JounNnL or rnr Ar"renrc;rNMoseurro CoNrnol AssocrATroN Vor, 9, No. 3 lection. both the overflow screen and the sec- 6,000 larvaelm'(Fig. 4). Densitiesvaried both ondary mesh bag were rinsed into the primary within reservoir and year. Speciescomposition receptacle,and the contents were then emptied varied among reservoirs, years and seasons. through a tilt-standpipe drain onto a submersed Three speciesof larvae dominated all collec- concentrating sieve. tions (Figs.4 and 5). The most conspicuousspe- Larval drift was monitored at 4 Tacoma sites. cies was Conchapelopia americanus Fittkau, a which included the chlorinator plants at Mc- relatively large, 4-6 mm, tanypodine larva. The Millin and Portland Avenue reservoirs,a tap off most numerous larvae in all reservoirswere usu- of the distribution line near Hood Street Res- ally small, <3 mm long, Psectrocladius species. ervoir, l6 km downstream from McMillin Res- Paralauterbornie Ila subcincla (Townes), the 3rd ervoir, and l0 km beyond that, the Marine View species,was common during the cooler months Drive Pump Station. Larval drift for these lo- of late fall through early spring and completely cations was monitored for a minimum of 3 replacedPsectrocladius as the dominant larva in months, from August to October 1987, at Hood McMillin Reservoirbetween late November I 988 Street,and a maximum of 2l months, from July and early June 1989. Other less frequently en- 1987through April 1989,at McMillin Reservoir. countered larvae included Tanytarsus sp., In addition, during the summer of 1989, single Phaenopsectra profusa (Townes), Micropsectra l-wkcollectionsweremade foreach of 5 different n.sp. and Cricotopus sp. municipal water systemsbetween Olympia and Standing larval populations comprising most- Bellingham, WA. ly Psectrocladiusspp. were maximum for each Strata sampling: l,arval uptake for different reservoir during August, with I 987 peaksranging water depths was monitored, in 1988,over 3 and from I,700 to 5,600 lawae/m2 for Alaska Street 7 months for McMillin and Portland Avenue and Portland Avenue reservoirs, respectively. reservoirs,respectively. Monitoring was accom- During 1988, the sameAugust pattern prevailed plished by a system of pumps and manifolds but with peaks of 800 and 6,800 larvae/m2 for constructed with 2-cm pvc pipe attached to ser- North End and Portland Avenue reservoirs. vice platforms within each reservoir (Fig. lB). Alaska Street Reservoir was discontinued at the Each manifold had 4 separateintakes, eachwith time becauseofreconstruction. In McMillin Res- delivery controlled by a dedicated centrifugal ervoir, peak August larval density comprised pump. Pumps were attached to collection units mostly Psectrocladiusspp., and ranged between similar to the secondaryreceptacles of the drift 1,500 and 2,500 larvae/m2 in both 1987 and monitoring apparatus previously described, I 988. These were located upon the service platforms Spring reservoir cleaningappeared to have lit- above the reservoir. Intakes were oriented hor- tle influenceon seasonallarval population peaks. izontally, one above the other, at depths of 15, Portland Avenue Reservoir, cleaned March l, 45,135 and 405 cm from the reservoir floor. 1988,recorded 1,000 larvae/mz within l8 wk in Intake openings were guyed 3 m outward from early July, compared to l0 wk postcleaningfor platform scaffolding to prevent the scaffolding McMillin Reservoir, which was refilled May 24, from serving as a confounding larval source. and indicated 1,000larvaelm2 by August 3.I.ar Twenty-four hour pump collections were made vae in North End Reservoir, which was cleaned at varying day intervals commencing with res- on May 2, 1988,never measuredmore than 800 ervoir postcleaning refill. Pump delivery was lamae/m2, and except for a very brief period in 0.875 kl/h. All larval collections were transport- October, seldom exceeded 100 larvae/m2 ed to the laboratory for counting and identifr- throughoutApril 1989. cation. Peak larval populations, in 1988, exceeded 2,O00lawae/m2 for Portland Avenue Reservoir, and slightly more or lessthan 2,000 for McMillin RESULTS Reservoir into mid-September.During 1987, Standing lamal populallons.' Sampling plates both Alaska Street and North End reservoirs. for assessingstanding larval populations proved cleanedand refilled May 27 and June 8, respec- to be of limited value. Only occasionally were tively, attained 1,000 larr.ae/m2by mid-August. they well colonizedwith easilydiscernable larvae Both McMillin and Portland Avenue reservoirs and tubes. More oflen plateswere obfuscatedby exceededthese populations even earlier. periphyton and a combination of inhabited and Drift: l-,awal drift proved to be a constant vacated early instar larval tubes too numerous phenomenon at all locations (Figs. 6 and 7), but and fragmented to count. Sampling plate data was characterizedby different numbers and spe- are, therefore, not included here. cies ratios. Mean drift rates for various locations Bottom tows were very reliable, and standing ranged from a high of 4.73larvaelkl for Marine larval populations were recorded in excess of View Drive pump station in 1987,to a low of SsPTtrvisrr1993 CnrnoNouro Lexvel TneNsPonr 279

PORTLANDAVENUE 6qn 1987 | 1988 1989 5000 I I 4{p0 t = 3ooo 6 93 OM i3 :3 TE la ,aa H rom A,na,nhuflhin:8 V\n\u,Vm,uh,nh ln s8@ IE Eem J e {00 200 0 I 80 17:2 215

o MCMILLIN o 6mo E a o 1987 1988 1989 5000 I 5 J 1 4{t00 o ,n t ul o ; q nn G U I o l smo E a o u I .,' ano I G t 3a G 3 o H rmo |alat,, g \|alaa,,afl,lcflnn d ul U 000 G s t !em 4@ 200 0 80 172

NORT}IEND 6{n0 1988 1989 som I {000 I ! r @NcHAPELoPIAAMERIoANUS ' suBclNcrA I a'@PARALAUTE.BO'NTEL; i a FlnaLaweaeoRNlELl-A I-ilI gomI ad ; I =w psecrloc-t-rotussp. ur ,",,i Eri s 800 ! ooo 4(x) mo 0 2Gr zil i 80 172 2g i 94 Fig. 4. Population densities by Julian days of benthic chironomid larvae in 3 Tacoma City reservoirs between June 1987and June 1989. 280 JounNel or rns ArvrrnrceNMoseuno Corrnor Assocn'rroN Vor. 9, No. 3

ALASKASTREET 6m0 1987 1988 ul I 5qt0 o I 5 I G C) 4(xxl ul at CONCHAPELOPIAAMERICANUS tt a, tI U| = lr ol t@ PARALAUTERBORNIELI.A SUBCINCTA 3(xrc o PSECTROCLADIUS c, 5 im SP. U' o o I 2(xx, - I E ul = -Aa n o. 1000 f- ul 800 G 5 600 tl00 200 0 175 2s i s0 188 Fig. 5. Population densitiesby Julian days of benthic chironomid larvae in Alaska StreetReservoir, Tacoma, WA, in 1987and 1988.

0.60 for the samelocation in 1988(Table l). The through September 1987 (Fig. 7). Drift for the highest recorded drifts, mostly Psectrocladius same location and period during 1988 did not spp., up to I l.5 larvae/kl, occurredat the Marine exceed O.2 lamae, and the highest earlier inci- View Drive pump station during late August dence, on July 6, was less than 4 larvae/kl. In comparison, larval drift for McMillin Reservoir, almost exclusively C. americanus, was highest HOODSTREET 1987 during April 1988, when it ranged from 4.2 to 5.8 larvae/kl, and in October 1987 never ex- ceeded 3.3 larvae. Peak larval drifts for other 6 locations were Portland Avenue Reservoir, 6.3 5 larvae (mostly Psectrocladir.rs spp.) in early July 1988, and Hodd Street Reservoir, 2.7-3.Olanae J 4 in September 1987. v 3 Isolated l-wk drift collections for other west- tr ul 2 ern Washington water systems ranged from 0 for o. tl the cities of Puyallup, Bellingham, and Everett, IIJ 1.0 to 0.05 and 0. l0 larvae/kl for the cities ofSeattle and Olympia, respectively. Both Bellingham and tr 0.8 Everett have filtered water systems. Samples for all sites except Salmon and Maple springs, Puy- J 0.6 allup, contained minute quantities of insect con- 0.4 taminants, including exuvial and adult remains ofvarious aquatic insects. Puyallup's spring sys- 0.2 tem is enclosed storage. The single sample from McCallister Springs, Olympia, also contained a 0.0 meager collection ofcopepods, ostracods and hy- 215 M 292 dracarina. pumps Fig. 6. Chironomid larval drift interception at 3 Strata distribution: One or more failed tap locations as recordedby Julian days,Tacoma, WA, during the collection period on several occasions. 1987-89. However, 8 successful collections (24-h uptakes) SpprsMspn1993 CnrnoNorr,rroLlnver Tr,cNsPonr 28r

PORTLANDAVENUE I 7 6 5 )4 Y3 :I CONCHAPELOPIAAMERICANUS tr2 'ZI PARALAUTERBORNIELLASUBCINCTA c \m psectaooLADtussP. ru1.0 e 0.8 5 o.o 0.1 o.2 0.0

8 7 6 5 r4 Ya 62r 1.0 ,,.ll|,, u e 0.8 J 0.6 0'4 0.2

80

MARINEvtEw 11 10 9 I 7 6 5 I 4 G 3 ut G 2 ul 1.0 ,au" G 0.8 J 0.6 0.4 0.2 0.0 80 172 M it01

Fig. 7. Chironomid larval drift interception as recorded by Julian days, Hood Street Reservoir, Tacoma, wA, 1987. 282 JourNer- or ruE AuenrclN Mosquno CoNrnol AssocnrroN Vou 9, No. 3

Table l. Mean chironomid larval drift expressed as larvaelkl and determined by flow interception at 3 collection sites, Tacoma, WA, 1987 and 1988. Portland McMillin Reservoir Marine View Drive Avenue l 987 I 988 I 987 I 988 I 988 n l9 35 t4 3l l8 x + SD larvae/kl 1.50+ 1.02 1.08+ l.4l 4-73+ 4.46 0.60+ 0.83 2.20+ 2.28 were made for each depth stratum at Portland lawa/kl the chancesof encounteringa larva in a Avenue Reservoir, and l0 for each stratum at given 200-ml cup of water would be I in 5,000, McMillin Reservoir,except for the 135 cm depth, and then only ifone was observant. which had I I collections. Larvae occurred at all Magnitude of problem: Although there is sub- strataon all occasionswith analysisofvariance stantial literature on the biology, systematics, showingno differenceat the P: 0.5 level (Table ecology and control of midge flies of the family 2). As with bottom samplesand drift, larvae did Chironomidae in most aquatic habitats (Ali vary in speciesand magnitude for time and lo- l99l), few accountsoccur for potablewater dis- cation as shown in combined strata collections tribution systems(Smalls and Greaves 1968, (Fig. 8). The predominant specieswere, as with bottom and drift collections, Psectrocladiusand C. americanus. PORTLANDAVENUE

DISCUSSION

The most significantfindings ofthis study were ! productivity s2 the degreeofchironomid in the res- ul ervoirs, the amount and frequencyoflarval drift, I and the even distribution of pelagic larvae TlJ throughout reservoir profiles.With thesefindings ioa it might seem surprising that complaints about I 0.. larvae are not more frequent than occur. That they are not is partly due to the minute size, translucence,and pale coloring of most speci- mens, and more importantly, the way in which we use our water. The greatestlikelihood of ob- serving a contaminant larva is in a drinking cup, a glass or an enamel basin. The quantities of water so viewed are minusculecompared to those used for bathing, laundry, cooking, gardening, pools and industry. Assuming a drift rate of I

J Table 2. Mean number of chironomid larvae/ kl by strata for McMillin Reservoir (in service F2 as of March 5, 1988)and Portland Avenue t10 n"r"*.t. tt" r"*i. ioa lo" Depth (cm) l5 45 135 405 McMillin n l0 l0 ll l0 x 0.38A' 0.544 0.394' 0.24,A

Portland n 8888Fig. 8. Chironomid larvaelkl for all strataby Julian i 1.364 1.07A 1.03A 0.844 days in eachof 2 reservoirs. Shown are 24-h uptakes I Horizontal values followed by the sme letter are not sig- following reservoir filling in May 1988,and continuing nifrcant at the P: 0.5 level. through December 1988. 283 Spgrerrrsrn1993 CnrroNor'nn Lenvlr TnlNsPonr

Mitcham and Shelley 1980, Gerardi and Grimm reservoirs as does Tacoma, and with approxi- 1982,Seppanen 1982). The most definitivework mately the same frequency.Visual inspection of Seat- on Chironomidae in water storage reservoirs bottom samplestaken in June 1987 from concerns the distribution and seasonalchanges tle's South Spokane Street Reservoir indicated to of the chironomid fauna of 2 English reservoirs sedimentsand larval populations comparable between 1950 and 1953 (Mundie 1957)'These those in Tacoma's reservoirs. Seattle lies ap- reservoirs differed from those in Tacoma, how- proximately 50 km north of Tacoma, and takes ever, in that they had mud bottoms and sup- its water from the Cedar River. Tacoma's open ported a complex population of flora and fauna' water source is the Green River. Encounter of midge larvae in domestic water Curiously, the city of Portland, OR, 240 km is less rare than might be presumed from the south of Tacoma, with 2 major open water res- prob- scientific literature, including water trade jour- ervoirs, reportedly does not have a midge nals (Rickards 1943, Flentje 1945). Most in- lem. This was confirmed by a visit on April 21, stancesgo unpublicized, or occasionally,as hap- 1989, during which several sample drags taken pened with the Tacoma jail incident mentioned from both Mt. Tabor and Washington Park res- earlier, are restricted to local news media. An- ervoirs yielded a total of only I I larvae. These other occurrence made local headlines in An- reservoirs were at the time overdue for cleaning napolis,WA, in October 1989,when customers on a 6-month schedule,but were nearly free of of an independent water company experienced sediment. repeatedlarvae. While working in California, I Reasons for the productivity differential be- consultedin severalsuch incidents, including one tween the city of Portland's reservoirs,and those from an enclosedLaCienega reservoir. This res- of Tacoma and Seattle are not known, but are prac- ervoir had become contaminated during a sand suspectedto be at least partly due to logging filter backwash with larvae-infested water. Es- tices. Portland's watershedhas historically been tablished larvae emergedto mate and reproduce. unlogged, and Seattle's Cedar River watershed Most chironomid mating is associatedwith logged only moderately. Tacoma's Green River aerial swarming, for which enclosed reservoirs watershed,by contrast, has been severelyclear- would not seemconducive. I have, however, ob- cut. Clearcutting contributes substantiallyto wa- servedterrestrial mating by males walking up to ter runoff and sediment distribution. females for a variety of specieson several oc- The lack of sediments in Portland reservoirs casions. Parthenogenicchironomid infestations also suggeststhat there may be lessairborne silt- in water systemshave been reported by Krueger ation than for Seattleand Tacoma, but this was (1941) and Williams (1974). Stenogamousspe- not measurable.Reasons for the atypical char- cieshave also been maintained in laboratory cul- acter of Tacoma's previously uncovered, ter- ture (Biever I 965). Reservoir enclosure,as noted minal, North End Reservoir, when compared above, is not a guaranteedsolution to midge lar- with other Tacoma Reservoirs remain equally va contamination. Other reports of midge larvae obscure. in enclosedreservoirs and distribution systems Preventiverecommendallons.' Although other includeWilhelmi (1925)and Smallsand Greaves factors may be contributing, it would appear, (1968).Enclosure does, however, protect against with the generaldecline in larval complaints, that airborne dust and particles, which contribute to enclosing 3 out of 5 Tacoma city reservoirs has larval nutrition. largely solved that city's midge larva problem. In 1988, a few sparseCDC light trap collec- Yearly population variation, however, cannot be tions of emerged midges were taken within Ta- wholly discounted. Although McMillin Reser- coma's newly enclosed Hood Street Reservoir. voir. still uncovered,has not been sampled since It was never determined whether a sustaining 1990. resident service workers report that adult population developed within this or other res- midge activity has been noticeably less than in ervoirs that were later covered.No midge larvae previous years.The absenceof spider webs, midge were reported by Tacoma water usersin 1991, remains, and spotting of structuressupports this and only a singlecomplaint was receivedin April observation. of 1992. In April 1993, however, several com- The purpose of stratification studies was to plaints were receivedfrom a singlelocalized area. determine whether larval intake might be influ- Of particular interest was the differencein chi- enced by modifying the elevation of reservoir ronomid midge productivity and encounter for outfalls. Uniform pelagic larval distribution in- open reservoir systemsin the major northwest dicates that it cannot be influenced. metropolitan areasof Seattle,Tacoma and Port- Reservoir cleaningduring active midge breed- land, OR. According to city of Seattlewater de- ing in spring and summer is of temporary value. partment personnelwho we visited in June I 987, Ali et al. ( 1976),working with concrete-linedflood Seattleexperiences similar problems with its open control channels in southern California, found 284 JounNer-or rns Arr.rpnlcexMosqurro CoNrnol AssocnrroN Vol. 9, No. 3 that cleaning provided satisfactory midge control Flentje, M. E. 1945. Elimination ofmidge larvaewith for only 2 wk. DDT. J. Am. Water Works Assoc.37:1053. Faucet filter screens and aerators were inves- Gerardi, M. H. and J. K. Grimm. 1982. Aquatic tigated and found to be ineffective as larval bar- invaders. Water Eng. Management lO:.22-23. Krueger, V. F. 1941. Eine parthenogenetische riers. Early instar chironomid larvae passed Chi- ronomide als Wasserleitungsschadling.Naturwis- readily through these when applied in water sus- senschaften36: 5 56-5 58. pension from a syringe. Mitcham, R. P. and M. W. Shelley. 1980. The control Enclosing and/or filtering municipal reservoirs of animals in water mains using permethrin, a syn- appears now to be the most practical and effi- thetic pyrethroid. J. Inst. Water Eng. Sci. 34:474- cient, if imperfect, means of reducing complaints 483. about midge larvae from water users. Com- Mundie, J. H. 1957. The ecologyof Chironomidae plaints of occasional larvae will undoubtedly in storagereservoirs. Trans. R. Entomol. Soc.Lond. continue to occur. Most of these. however. can lO9:149-232. Rickards, C. 1943. Overcoming a bloodworm (chi- be quietly and successfully handled with a simple J. ronomid) problem. Public Works Mag. 74:11-38. factual explanation. Seppanen,H. 1982. Microbiology of drinking water: organisms in filters and the distribution network. REFERENCES CITED Vesitalous 23:24-29. Smalls,I. C. and G. F. Greaves. 1968. A surveyof Ali, A. I 99 I . Perspectiveson managementof pestif- animals in distribution systems. J. Water Treat. (Diptera), global erous Chironomidae an emerging Exam.l7:150-183. problem. Am. Mosq. Control Assoc.T:260-281. J. Wilhelmi, J. 1925. Zuckmuckenlarven in Hochbe- MullaandF.W. Ali.A..M. S. Pelsue.1976. Removal halstereiner Wasserverorgunsanlage.Kl. Mitt. Mitgl. of substratefor the control of chironomid midges in Ver WasserversorgungAbwasserbe s. l: I l9-120. flood control Environ. En- concrete-lined channels. Williams, D. A. 1974. An infestationby a partheno- tomol. 5:755-758. genic chironomid. J. Water Treat. Exam. 23:215- Biever,K. D. 1965. A rearingtechniquefor the col- 229. onization of chironomid midees.Ann. Entomol. Soc. Am.58:135-136.