TJ'A-670/2- 75-064
June 1975
P B 242177
ANALYSIS OF TOLUJriON FROM MAR I NT: F-VGIXHS AW)
~tilts on 'ii q: i-inv l ronmilvi*
Sunrury Report
By
Bo.it in^ Indiisfy Associations
Chiai',o, Illinois 60611
Cr.int No. R-SO 1799
Pro^nin lilmt-nt No. lhftO.iS
Project. Officer
I.co T. McCarthy, Jr.
Industrial Waste Tre.itnent Research Liharniory
ison, Now Jersey USS!"
,\\ri.\A'iKiT:.nxi'.\i. Ri>"i:.w.ii r.i'xn.R
on;icr. of RisrARai .wo M-VM/Hrs-vr
U.S. l^YI&WKNTAI. rROTTCI ION AQ'.NO'
ci:;«:iNx\ri, >.'• iio 4r,:(.s
-------
rizviw NOTici:
The National l-jivi ronmental Research Center, Cincinnati, has reviewed
this report and approved it for publication in order that t!;c data
niriv lie made widely available and serve as a basis for future itudies
by others, should they be desired. Approval does not signify tint
the contents necessarily reflect the vic-us and policies of the U.S.
linvironnental Protection Agency nor docs mention of :rade names or
comercial products constitute endorsement or recommendation for
their use.
-------
FOREWORD
v.an and his environment must he protectee! from the adverse
effects of pesticides, radiation, noise and other forms of pol-
lution and the unwi:;e monagoment of solid waste. Efforts to
protect the environment require a focus :hat recognizes the in-
terplay between the compon- ts of our physical environment—air,
water and land. The Nat'-nal Environmental Research Centers
provide the multidisciplinary focus throu-jn programs engaged in
° studies on the effects of environmental contaminants on
man and the biosphere, and
° a search for ways to prevent contamination and to re-
cycle valuable resources.
The northern lakes, southern lakes and laboratory studies
were conducted by independent consultants who were subcontractors
to the Boating Industry Associations. The project design was ap-
proved by the U. S. Environmental Protection Agency and monitored
during the course of the study by EPA personnel. As a result of
the studies and the contractors' analysis of the data, no signifi-
cant environmental impacts from outboard motor boating activity
were found under the study conditions. However, since it was in-
possible in this study to cover fully all the diverse factors which
may be impacted by emissions from outboard motors, care should be
taken in extrapolating the results to situations other than those
covered by this specific study. These data are being presented in
order that they may be widely available and serve as a basis for
possible future studies by others should they be required in con-
sidering the implementation of regulations relating to the
recreational utilization of the Nation's waters.
A. W. Breidenbach, Ph.D.
Di rector
National Environmental Research
Center, Cincinnati
-------
AI-S'IR-VT
This is a I'riTi't of n research project which Njm'.iii in April
I'.Tl . The ohiective c\ : 1 " i'r.vi ronriental Science and 1 iv.\ ir.eerim:, Inc., (la inesvi 1 le, l:lorid:i.
To achieve the project objective four :t>;u!s '.-ere sp'-:ected to out-
hoard ermine emissions at a rate calculated to be t;,rec tires creator
than l;,it fron saturat ion ' rating levels. Sopo rar;.;in.il e' mi^os i;i
the la-es biota wore noted I-'.if the differences ^ore -neb tivit it is
not certain, whether rhev were frop natural or stress effects. As a
result it was no* "OssiMc to detop-ine cvw'iusively tV Precise
point at which outboard emissions effect t'-e r.ruat ic er.\ i rojirent .
I'ased on the results, it is ">1 aucKlo to conclude, ho-.-evor t'-at b"-
ca::S'"' of ti:e hiu" stress levels o—Moved in this stt:dv, "o.i:board
nofor epissims do rot s ! f i car: 1 v affect .laiat ic ecosvsteps.
His report was subp.i ttcd in :"i; 1 fi) l^evf of 'Iran .'.'ir'-c:' I'.-S'Mb\-
I be heating Indu.-try A^soc i::t iorsurder t "e partial sroiw-orsh i p of the
! r.v i ror." en;al Protection V:e:cv. ' ieid. a-d Inboratorv vor1 v.as con-
dieted as of 'Kxjep.ber 10 ~ 5.
-------
CO.VIT.Vi'S
Pace
AHsrmct i
l.ist of !:icures ii
I.ist. of T.-tMes i i i
Soct j QMS
I Conclusions 1
II IVcorncntlnt ions 6
Ml 1 nt rodiict ion n
!Y Laboratory Stmlios 19
V l'ioki Stmlics -JC
Nort!'cr» Hole! Studies
Soutiu-rv. i ield Studies
"i"i ;ViV!0!?cos 61
v
-------
FIQIRI-S
No. Pago
1 Northern Field Study - Test Facility 1Z
2 Simulated Motcrboat Iriissio:i Loading 14
-> S inula tec Motorlioat linission Leading l.r>
-J Stress Ixvels 17
5 Schematic Ilia gran of tlic Sarrple
Collection and Analysis System 22
6 Compos i t ion by Hydrocarbon Families of
Condensate and of Fxhnust Gas Stre:in,
Suncrinrcsed Over That of !^n>' Fuel.
Disappearance of Outboard Fngine F.xhaust
Products fron Aerated and Ouiescent
vi
-------
No. I'n ee
I Test linj'.inc and Spec i fic.it ions 2(1
II Surnary of Condensate Analyses 2!1
!II Condensable Material in l:;ic)i Stare As
IV;11cr and Or;_;:inic ''aterinl
(Chrysler 3.6 I'!' Inline) ~o
r\' V-cdian Tol enir.ee I. in its for CioMfish
!ixiK5sctl to Arorvitic I'ydrncarbons 3S
V Median Tolerar.ce Units for Cold fish
l.xposed to (\i:hr,ai\i Inline l.xliaust
Water 3S
VI W-Analyt ical <>;»t :i for Saline Test
roiuis (vaii urn .V.1
VI! Kni'.ines I'sed l>:irinu lest Period -Ii
\'I I : Stress Levels 42
!\ Hydrocarbons* f:'oi 1 ir.'j Point
Ranv;c >]
Saturated I'vdrocar'~ons in Sediments S2
V) 1
-------
sr.cn on i
COXCI.I ISION'S
my>!t\TORY siim-
1. (Vis-phase carbon nonoxide concentrations were closely related to
the fuel/air ratio supplied by the carburetor, and ranged from less
thin 2 percent at 1,00'i rpm for a <> hp I'vinmde engine to greater than
0.? percent at 5,POO inn for a 35 hp Chrysler engine. fHn;inic fish tox-
icity studies showed tint carbon nonoxide, even at near-saturation levels,
did not produce fish mortality.
2. lVis-phase carbon dioxide concentrations ranged from a low of approx-
irateiy 3.5 pcrc-ent to rore than 9.5 percent, and were also principally
a function of the fuel/air ratio.
3. No trend with speed and load was observed for either carbon monoxide
or carbon dioxide emissions.
•3. Total gas-phase hydrocarbon emission concent nit ions range from a
low of -l,50r> ppri neasured as Cf-.l'].: to a high of ]n,opn pnn as
These concentrations, attributabie to overscavenging, which in tum is
related to engine trapping cf'iciencv, generally decreased with increas-
ing speed and. load.
5. fngir.e trapping efficiency ranged between S'l and SO jiercent and in
general was observed to increase with engine speed.
(>. !'.iss er.issier. rates of both carbon rionoxide and unburned hydrocar-
Ivir.s increased with increasing speed and load in the test engines. l:or
exarple, tb.e total hydrocarbon emissions ranged fror less than 0.02 kg/
hour for an ivinnide <> hp engine at 1 ,"00 rp>n ff.2 bhp'j to approximately
3 kg/hour for a 105 hp Chrysler engine at -1,0nn rp-i (.V blip).
". Tie corpcsition of the gas-phase exhaust hydrocarbons resenMed the
ccr-posit ion of the fuel with the principal exceptions That the olefin con-
centration va? greater and tb.e paraffin concentration slightly less than
in tb.e test fuel. A rode rate variation in co-rpos i t ion was evident, from
engine tc engine.
S. ihe con.leasable naterial fror1 outboard engine exhaust was found to
contain paraffinic. olefinic and aronatic hwlrocar^ons as well as srnall
arc: m is of pbenols and carbcnyl corpo-.uicls.
9. The corposii ion of the total cor.bined condensate was very sinilar to
that of the fuel. Arenatic conprunds constitute 2n-25 pr-rcent of the
tnt.i! coniU'nscfi iivdrocarbon ar.ount . Toluene is slir.btlv lover on a Per-
centage basis in the condensate than in the fuel, and binuclear aroratics
are slightly higher.
10. Tb.e total arount of conder,s:\Me rater i.-il which can reasonably be ex-
pected to l-e condensed in a beating situation varic\' fron al-out 1.5 - 7
iKTcent of the fuel used.
-------
11. Three engines were investigated for cranlxase drainage: all exhib-
j ted the general trend of decreasing drainage with increasing engine
speed and load.
12. At low speed and load test conditions the spread of the drainage re-
sults were relatively large. !:or example, at 1 ,F.nn the average drain-
age for an IS hp hvinrude was approximately 60 grams per hour whereas
drainage froiv a 35 hp Chrysler was in excess cf 275 grans per hour. At
tlie 1,500 rpm test condition drainage, expressed as a percent cf the fuel
used, ranged from 5 percent for a SO bp "crcury to in excess of 8 percent
for a 55 lip Chrysler engine.
15. 'Hie oil conposi ticn of the cranlcasc drainage was about 20-50 per-
cent. Since the ratio of oil to gasoline in the fuel was 1:50, crankcase
drainage represents a 10-1S fold increase in oil content over the mixture
fed to the engine.
1-1. Maintenance in the form of a conventional "tune-up" had little influ-
ence on either the gaseous or condensable emission characteristics of two
field engines tested.
15. The aromatic constituents of the first stage condensate have an evap-
oration half-life of about 11 days in a lake or other water body, assuming
conservatively, a quiescent body of water at j"i°C, the condensate being
unifornly distributed initially to a depth of one meter.
16. 'Hiere is a small, non-vDlatilo hydrocarbon fraction which is not re-
noved I1}' evaporation from water exposed to submerged two-cycle engine ex-
haust emissions.
17. Pb-hour T!.\f values for goldfish mortality were"dcternir.ed in dynamic
bioassay tests as P-.1.0 pom as toluene for outboard engine condensate.
¦WlllilX IA'Ih snipy
IS. No significant differences were seen in periphyton diaton ri chiles s
and species distribution between ponds during two ye,ars of study.
1P. Although variable, organic production was not s i gp i f.icantl v dif-
icrent nor was chlorophyll a_ production different between panels. I'ow-
ever. when placed >n the ratio of the autotrophic index there was a sig-
nificant difference between the control and nop.-.leaded treatr.cnt sections
during the 1P72 sampling period. 'Il'ese
-------
Phytoplankton species associations varied annually in a manner indicative
of natural lake systems.
21. Phytopiankton productivit.y treasured by ^ carbon fixation indicated
lower piiotosynrhctic carbon production in both the r.on-leaded and leaded
stress sections when compared to their respective control sections in
most collections during 1971, 1972, and 1973. Of these differences, only
the-lower carbon production in the non-leaded stress section conpared to
its control during 1972 was significant. These differences cannot be
directly attributed to natural popr.lat.ion variations as species associa-
tions and rictincss were similar throughout the pond system during this
study.
22. Chlorophyll a measurements of t.hc phytoplankton recorded durinp
1975 showed no significant difference between the non-leaded test pond
and its control pond. During 1073 a significant difference was recorded
in chlorophyll a measurements between the leaded stress [Kind and its
adjacent control pond.
2j>. Phytoplankton productivity index values showed no photosynthetic
inhibition in terms of a ^ carbon production to chlorophyll a. Al-
though only a few data points were anlayzed this index is fclF to be
legitimate and useful in studies of primary productivitv.
24. Zooplankton population dynanics, comparative species richness, abun-
dance, and occurrance were indicative of normal temporal periodicity en-
countered in small temperate lakes. No statistically significant effects
on the zooplankton community can be attributed to the outboard motor
emissions in the northern ponds.
25. Hie benthic macro invertebrate coixainity demonstrated normal vari-
ations in population corinositon and dynamics. The 1972 shift in dominant
organisms composing the benthic fauna! assemblage was copimcnsurate with
the change in trophic structure of the ponds and could not lie correlated
with stressing by outboard motor emissions.
26. A single fish taste test in 1971 showed an alteration in the taste
of fish taken from the stressed ponds at a treatment level ot: 52 gallons
of fuel burned per million gallons. Subsequent fish taste studies during
1972, at treatment levels of 1 .-1, 1.5, 2.8, 4.D, 4.2, 11.2, 76.9 and 110.5
gallons of fuel burned per million gallons showed no tast-e alteration in
the fisli population.
27. Xo major variation in the general water uual.it>" of the test pond
were observed as a result of stressing..
2S. Field and laboratory studies during J 971 and 1072 c-n aromatic hydro-
carbons (gasoline fraction) usinj; the cyclohcxar.o extraction - UV spectro-
photometry procedure indicated little difference between stressed and
control sections. The maximum concentration observed wis 50 ug/1 (as
toluene) in the stress ponds. Poth field and laboratory results indi-
cate that the majority of these aromatic hydrocarbons remain in the water
-------
Cuiimin for a relatively short tine, less than a day under conditions
nomally encountered in natural water syster-.s, before they are renoved
by natural physical (evaporation), chemical fadsorntion) and/or biolopi-
cal (hioxidat ion) processes.
29. No significant chance in the concentration of saturated hvdrocarlxns
with boiling points in tlie ran^e of 175° to Jnn°C ('corresponding to in
nolecular -.-.eight Cjp to n-raraffins) in the water colirm wis observed
as a result of three years of outboard engine stressing.
30. No stati>tically si ficr.nt (PS percent confidence level) buildup
of saturated hydrocarbons was observed in the test ponds sedirents after
three years of engine operation. The data collected in this investigation
cannot rule out the buildup of those natorials in the s'.'d inents. The re-
search data collected to date indicate that any increase in sati'i-atod hy-
drocarbons-present in the sodircnts involves saturates with carbon ninbers
C17 and above.
31. In the leaded fuel stress rond an increase in lead in the water cohcn
frcr an avera.ee background v.ilue of -1.3 to >.7 rarts :vr billion Ot^) was
deserved anil is ili recti;-' attributed to stressing by out boa id cm: ines using
leaded fuels. In hard water lakes such as t!io nortbcj-n stndy pew'., the
concentration of lead in the \ntor cohrtn is limited to a;-m-ox inarch- 1craturo, 1 i «*.! 11 , and n;rrient lew's, and can-
not he conclusively correlated with treatment eff'cts rf outboard r.otor
opcrat ion.
34. Phytorlankton bicassays conducted in si;:: en a linitcd basis in the
southern test lakes shew out'"oani rotor rxhanst water at a ratio of 1:2-1
(ratio of exhaust, wares to lake water) to inhibit rhoT-osviithosis. This
treatment level is 3?P 'fines the exhaust ;.-as -atei cc cent rat ion antici-
pated in waters receiving norral outboard -otor usage. It is unrealistic
to think-that boat usage would reac:> this level on a:'.-' recreational lake.
33. Growth nues and bic-rass of pci"hytor. culture ;n th.e I iryici ic iciv--
or. artificial substrates '.-ere net affected ;-y outboard rotor'onera*. ion in
th.e southern test lakes.
4
-------
56. In the southern test lakes, dist rihut ion , si^ecios comosi t ion and
diversity of henthic racroinvertchratos were not affected I1)- rotor o;>or-
:ition except in a sruill area irviediatelv heneath the pertinent ly .vomited
notors where scour ins*, of tl.e hotter: scilir.er.ts resulted ! nn rotor o:x'r;i-
t i cn.
37. '-Wor operation in the southern lnl.es increased t hr* concent nit ion
of dissolved a ror.11 i c hvditvnrhoiis, r.ixed and circulatcii r".:e lake inter,
and therefore distributed t no hydrocarbon emissions as well as otlier
emissions t hrouy.hout the lakes. The concert nit ion of arr~ttic hydro-
carbons increased fror knekcroitnd levels of less t!un 0.01 r^/1 to levels
of 1.0 r;;/l durini; mtcr ex"rat ion. V-hen the rotors were not operated
for tv.o days, hydrocarbon levels declined to less tlian '*.1 r-.j;/l.
5S. The -evol of dissolved ort'.ar.ic carbon in the lirnetic ;ore of the
southern lake treated with drained tvpe engines was sir.ni ficar.i 1\- greater
than both the control lake and the lake treated i.ith dniir.less tvjx- engines.
The drainless t>r>e or.eine (lis a recirculating device which t-lirinatcs
c:' :r.i:case drainage. The drained t^e engine does not ha\e :!iis devi ""
and therefore enits a greater arc: sit of unhurried fiie 1 .
351. !-"ish tastes condivrted hv :he I'r.i vers i t v of I'loritla I ocvl Sciences
!V:\'i:"t:-ient have Iconst rated r.o evidence ef taint ir.v. hy eutl^eard -otor
«tissions even at treatment levels fur in excess of tnose in the ".S.
iVM ic i'ealth Service study - here tairt ir:-. - as oh^er/ed.
-I;1. As a result of Kiclvrm::,1 s:v~lii'.r i-eforo t ixnif-er* Saive a-ounts of
lead fS to '.'O T of dried •" lant tiss'so"! vere detected in t ho i-ooted
vei'Otat ion of the rmss !,cd ccrr-unity. Therefore, tiw effects of l?::d
erissiens on the southern lnles v. ore rot sv.die.i.
¦J 1 . There '.ins no overt evider.ee '.hat treat-vnt si-mi ficant 1 v affected
the levels of the fol lev. ir.;: c;-e*"i cal arareters in the southern lakes:
iion. "irnesi'.ri. chloride, siilfato, .'novice, total solids. su-rended
.•¦olids. dissolved solids, total '¦.•mines? , conduct ivity, tur'-iditv.
hiecherical ox wen demand, ar.d cherical ONv-en de~and.
-J2. ! xcc"i for carho:'., there ¦¦"as r.o direct evidence of difference :n t'-e
r.utiicnt re-.; iron of the sou*cm lake S'-'Ste~s associated v i th treafent.
-!3. ;'\:t hoard rotor oreration rroha'-lv increased r.rass lv«5 ^iTdivt h i ;y
in the southern test lake,-. ;hc dm ire-' er.i;: ne test lake exhibited a
s i ::r i f i cant! y !ni;icr (:r ^ro>. i-.at r 1 v ir.e rercer.t' i;rtss he ! rro :i;ct : vi t v
a:-:.! hio-"!SS r'\ir. t he cert it-i i ;ke. ,:1-c rn-s ' d productivity of t'e
dn in loss cnuire lake, nl ?'•«?•».: •'*. rot st.it i.-t ica. .y s i i fic~n* . averaccv!
u-'"!T\ iratcl v 1 :crcent 'i'.;:er than t h.r c.-rti'ol lake, i ass in the
control lake anti in the era ir. less or.- ir.e lake •.•;as siri lar.
4J. 'ihe effects of out!rv!7\: -»-t«r e-issier- ^r the .vasal ic ev-esvste"! in
the sent:'.err. lake-- coi:hi not ;~e di.-corned frr-~ T'-e t,ir1,uier.t eti'e.'is of
::ixi:iL: ;u;d stirring indeed in t V-e lakes hv rrtKr:
-------
SECTION' II
RECOMMENDATIONS
1. The effects of outboard marine engine exhausts on phyto-
plankton productivity is worthy of further inves»:ipation.
The data have shown reduced '"carbon fixation bv the phyto-
piankton under the treatment condicions of non-leaded and
leaded fuels, the former showing statistical sipnificar.ee.
This phenomenon could, nore closelv obser\'ed under alf.al
assay and continuous flow laboratory bioassay conditions
where prepared media and natural waters are exposed to
various levels of outboard enpinc treatment and fuels.
Various ^rowrh rate and productivity measurements can be
taken to relate carbon production on a per cell basis or
per unit chlorophyll a and expressed as a productivity index.
Such expressions ind." -ate phytosynthctic efficiency or
inhibition, and under controlled assay conditions can be
related directly to the stress variable being applied in
the experiment.
2. Chemical analysis of. ..the culture nacerial should also
be undertaken to determine hydrocarbon and heavy metal
sorption or uptake by alr.nl cells. These analyses could
identify the exhaust fraction involved in photosynthetic
inhibition if this is indeed a real phenomenon.
3. Experimental stressing of natural waters vlth outboard
r_irinc engines influenced the periphyton communities under
study. Althouch no s i t;r.i f i cant variations were seen in the
species associations of the photoautotrophic organisms
(diatoms), differences were seen in the production of
orpanic biomass and chlorophyll a. When annlied in a ratio
as the autotrophic index the dara reflected more hecertrophic
periphyton cornroinities under treatment conditions. The
speculation beinp that there was greater nicrobial activity,
i.e.. bacteria and funri, in the stress sections of The
northern field study. Experimentation should be conducted
ro determine if this is the result of selective enrichment
vhere the hydrocarbons of the exhaust emissions provide a
carbon source for bacteria, and :'unr;i. This aspect should be
studied i;nd?r continuous flow bioassay conditions where natural
water wouid be treated with outboard engine exhausts and fed
through the bioassay system.
-------
U. The zooplankton community demonscraced no effects observ-
able at this level of study, i.e.. y.ross population monitoring.
It is felt then, that the zooplankton could be studied on a
oore individual manner. such as their robustness, fecundity
and fertility. Cut analysis on the Rotifera. in a stressed
situation, should also be undertaken to determine if the
ratio of small alj-a (i.e.. Chi ore 1 la sp.) to suspended
deer leal natter is concomitant with populations of unstressed
lakes.
5- The benthic macro!nvcrtcbratcs should be studied In 1 i>• hc
of their preferred habitat; the sediment. It should be
determined, in a continuous flow bioassav unit usinp, natural
lake water as the diluent, if the population of Chirononidae
denxinstrate a reduced ability to exchange Rases (0? CO2 etc.)
due t:o hydrocarbon build up. Biochemical analysis of the
body tissues should then be perfomed to determine variations
in lead, and the various hydrocarbons- associated with the
flesh of the organises. In the bioassav unit, the duration of
the larval staj;c should also be determined. Species of nacro-
invercebrates less anatomically amorphous should also be
observed and chemically analyzed on an anatomical basis:
i.e.. t'.ills, body -overinj', (chitin). and viscera. The icthyo-
population should bo studied with particular reference to
their f.ills and the possible build up of the heavy r.etdls.
and lead. A simple bioassay-anima1 dissectior coupled with
atonic absorption sncctrophotometric determinations would
yield information in this area.
6. There is some uncertainty concerning the build-up of lead
in the bottom sediments as a result of outboard engine opera-
tion jsinp, leaded fuel. Conscquen 11 y , further study should
be nide of the mechanisms of transport by which any signifi-
cant concentration could occur, c.n. sedimentation. However,
this potential environmental problem should he eliminated in
the near future with the advent of non-leaded fuels and the
development of two-cycle outboard engines capable of operating
on such fuels.
7. The fate.of hydrocarbons in the aquatic environment is
still not well understood. The d.3ta presented in this
report si'i-.^esrs the need for further study in the area.
Such studies should focus or: the physical, chemical and
biological degradation of hydrocarbons in the aqueous
and sediment" phases as veil as the mechanism of hydro-
carbon transport between the two phases. This aspect
should be looked at for "ior.p-tcrm effects". I'rc-hably.
p.ot as important per se for outboard motor emissions as
much as aidinr in ascertaining the environmental impacts
of the total hydrocarbon. input from man's activity on
natural water systems.
7
-------
8. An investigation to assess the effect of outboard
motor operation on the carbon dioxide L>LHt'.et should
be conducted in carbon limiting lakes. Eir.phusis should
be placed on the effects of carbon enrichment on primary
productivity and plant biomnss. Carbon dioxide pathways
(e.g. nass transport of carbon dioxide via water circula-
tion, dissolved carbon dioxide emissions, atmospheric
diffusion of carbon dioxide) should be traced and quanti-
fied with the use of a MDIR carbon dioxide gas analyzer.
9. Experiments should be designed to discern between the
effects of motor emissions and the effects of nixing
and stirring on lakes such as those experienced in the
southern study.
8
-------
SECTION III
INTRODUCTION
GENERAL
This summary presents the results and conclusions of a three
year research project designed to ascertain the pollutional
effect of two-cycle outboard engines emissions on the aquatic
environment. The study entitled "Analysis of Pollution
from Marine Engines and Effects on the Environment" (EPA
Grant No. R-801799) was directed by personnel of the
Industrial Waste Treatrnent. Research Laboratory, Edison, New
Jersey (a division of the National Environmental Research
Center, United States Environmental Protection Agency,
Cincinnati, Ohio). Financial support for the project came
from the United States Government and the Boating Industry
Association. The field investigations were conducted by
Environmental Control Technology Corporation of Ann Arbor,
Michigan and Environmental Science and Engineering,
Incorporated of Gainesville, Florida. The laboratory
investigations were performed by the staffs of the Environ-
mental and U'ater Resources Engineering and the Automotive
Engineering Laboratories of the University of Michigan,
Ann Arbor. Michigan.
This summary is a digest of a considerable body of data
collected over the course of the project. A complete
compilation of data is provided by each of the above
research groups, in separate reports to the United States
Environmental Protection Agency.
OBJECTIVES
The main objectives of the study were:
1. To determine the effects of two-cycle outboard engine
emissions on the aquatic ecosystem. This included not
only determining whether any detrimental effects occurred,
but also pin-pointing that portion or portions of the
food chain which night be most seriously affected.
2. Quantitative and qualitative characterization of the
exhausts from two-cycle outboard engines. Particular
emphasis being placed on those exhaust components which
tend to remain in the aqueous phase. The charac'/erization
studies included variations cue to engine horsepower, age
and maintenance, as well as, manufacturers design.
Q
-------
SCOPE OF VORK
Basically- the project was divided into two distinct parts,
namely, a laboratory study and a field study.
The laboratory studies were conducted under controlled
conditions in order to optimize the collection and characte-
rization of components in outboard notor emission water
(OME water). The laboratory approach was necessary, in
this regard, as it was the only practical way of obtaining
the data and it reduced or eliminated many of the analytical
problems associated with the characterization of the OMK
water in a natural v-?ter system.
The field studies, on the other hand, were designed to
ascertain the environmental impact of two-cycle outboard
motor emissions on the chemical and biological quality of
natural lake systems. In this instance, it was felt that
laboratory scale studies would be inaporopriate due to the
inherent difficulties that would be associated with extro-
polating the data to "real world" conditions. In like
manner, field studies of larp,e lake systems was prohibitive
from both an economic and tim? standpoint. Therefore, it
v;as decided to perform the field studies on small ("mini")
iakts. This "mini" lake concept allowed for the field
investigations to be carried out on water bodies of manage-
able size. This approach also permitted simulation of long
term stressing by tvo-cycle outboard engines to be accomplished
over a relatively short period of time.
The field studies were conducted in "mini-lakes" (0.5 to
U acres - 0.2 to 1.6 hectare)in both a northern and southern
climate. All "mini-lake" systems, both control as well as
"stressed" systems, had no prior boating activity on them
and receive no other pollutiona! inputs other than the
imposed stressing from two-cycle outboard engines operated
under a predetermined schedule. The northern study dealt
with the pollution impact of crainless two-cyc?.c outboard
engines using leaded and non-leaded fuels. The southern
study investigated the pollutional impact of drain versus
drainless tvo-cycle outboard engines using leaded fuel.
Basically, the testing procedures were the same for both the
northern and southern studies, except that the northern study
underwent an enforced rest period due to ice cover during
the winter, while che southern lakes were stressed year
round.
10
-------
The northern field study was carried our. in two one-half
acre (0.20 hectare) ponds near Saline, Michigan. These
ponds, part of the State of Michigan's Fisheries Research
Station, are owned and maintained by the Department of
Natural Resources - State of Michigan. Each of these
ponds was divided by means of aluminum sheeting to provide
two test ponds, each with its integral own control (Figure
1). These ponds or "mini-lakes" varied between four and
nine feet (1.22 and 2.1U meters) in depth and have been in
existence for over ten years.
The southern "mini-lakes", situated west .of Archer, Florida,
were three separate lake systems located in close proximity
to one another. The two test, or "stress" ponds were three
to four acres (1.2 to 1.6 hectares) in size while the
control lake was approximately twelve to fifteen acres
(4.8 to 6 hectares). Each of these :-outhern "mini-lakes'*
varied between zero and twelvj f<-et (0 to 3.66 meters) in
depth, with an average depth of approximately six feet
(1.83 meters).
Water levels in all "mini-lakes" systems, both north and
south, were controlled strictly by evaporation and rainfall.
There were no natural surface or .sub-surface water flows to
the systems. In the northern "mini-lakes". water levels,
hence water volume, were hel-i constant by periodic war fir
input when evaporation and seepage exceeded rainfall. In
the southern study water levels varied seasonally as there
was no way to withdraw or acd water to the "mini-lake"
systems. "Mini-lake" water levels in the southern field
studies varied between two feet (0.61 meters) below and
three feet (0.9 meters) above the original water level
datum at t!ie outset o: the study.
Stress inn Levels
The first consideration in the field studies was to determine
die -3 mount of two-cycle outboard stressing that war- to be
i':pp.ied to the test "mini" lakes. The amount of outboard
stressing to be applied to each test system was approached
by asctrtaining,' as closely as possible, the number of boats
that resonab1.;,' could be expected to occupy a given surface
..re.'i of water under opt i.r.u.": use condition. Optimum use
conditions being, l "real-world" situation where the follow-
ing conditions exist:
1 i
-------
WORTHED FIELD STUOY - TEST FACILITY
SOFT.
60 FT.
50 FT.
60 FT.
POND I
(NW- LEADED
STRESS)
KET.IROJLATIN I
— PUMP3
— AIR SYSTEM —
FOR D£- ICING WALL
-------
a) diversity of boating activity taking place
b) lake surface large enough to have boats operating
over a wide spectrum of horsepower loadings
c) lake should be at maximum usage capacity
(saturation boating). Saturation boating boing
that situation where from a safety, aesthetic
and recreational viewpoint the presence of
more boats on the lake would cause:
1. a problem of boating safety
2. reduce one of the legiraate use
activities presently taking place
3. reduce the recreational attractive-
ness of the lake
Thus, a literature review was undertaken to determine the
spatial (surface area) requirements per boat based on a
specific boating activity. In 1967, Threinen^ reported
on his findings of boating activity on Lake Geneva.
Wisconsin- A review of Threinen's work indicated that the
boating situation on Lake Geneva during a peak holiday
weekend period (July 4th) was at or near optimum. This
boating situation together with the size of Lake Geneva met
all the necessary qualifications set forth previously, to
define optimum use conditions. Therefore, the work of
Threinen on Lake Geneva, Wisconsin was used as the base to
determine what level of outboard notor stressing would be
used in the field studies.
Schematic flow sheets of the instantaneous boat population
and specific boating activities as developed from Threinen's
work on Lake Geneva*- are shown in Figures 2 and 3. Figure
2, which shows the average instantaneous boating activity
from 1000 to 1800 hours, indicates that there is essenti-
ally total saturation with respect to notorboat usage during
this period.
Using the data from these two figures (Figures 2 and 3), and
making the following assumptions for average brake horse-
power (bhp) output for the various boating activities:
water skiing - 40 bhp
boating - 10 bhp
trolling - 5 bhp
a total brake horsepower input of 61,800 hours for the peak
13
-------
Figure 2
SIMULATKD MOTOKBOAT 1SSI ON LOAftlNC
Spatial Considerat 1cm - l-ifce (Ifni'V.i, Wlscon1 n
Peak weekend period (July U, 1967)
1000 hours to 1800 hours
Lake t'.onevo
(5100 Acres of Water)
2064 hectares of Water
3000 local registered boats
300 local unregistered boats
300 out-of-state boats
percent, in use at given tine
rent 31
oTiTT
600
i
t
502
I
I
100
-600-
i
private
2-'. 00
300
300
1000
I
102
I
3C0
I
50.T non-ttobilc
!
300 units
I
SO.* r.>hi 1 e
300 units
water ski inn (50?O
boating (502)
recommended spatial
need
150 un i t s
30 acres per unit
(I 2 . 14 hectares)
150 units
3 acres per unit
(1.21 hectares)
1 acre per unit
(0.0-'.0? hectares)
300 acres
(121 hectares)
I
5250 acres.
i500 ::c res
(1821 hectares)
^50 acres
(182 hectares)
14
-------
Figure 3
SIMULATKD MOTOKaOAT EMISSION LOAD I N'j
Spatl.il Consideration - l-ike Geneva. Wisconsin
Peak Weekend - non-^eak hours
0600 Co 1000 hours and 1800 co 2000 hours
600 unite
901 fishing.
540 units
102 ooblle
(trolling)
60 units
1 acre per unit
(.40^7 hectares)
W0 acres
(219 hectares)
3 Acres per unit
(1.21 hectares)
I
180 acres
(73 hectares)
720 acres
(297 hectares)
15
-------
day at Lake Geneva was calculated. Since an average of
one gallon (3.785 liters) of fuel is consumed for 6 bhp-
hours,2 this represents a daily fuel use of 10,300 gallons
(33,986 liters), or approximately two gallons per acre per
day. (18.7 liters per hectare per day).
It is obvious that the vast majority of lakes do not receive
such extensive boating usage on a daily basis. However, it
was decided for the purpose of this study to assume that
the maximum weekly stressing level of a hypothetical lake
saturated by boating activity wculd be represented by
two days of peak activity (18.7 liters per hectare per day)
and five days of one-half peak activity (9.35 liters per
hectare per day) of that shown for Lake Geneva. The total
maximum weekly outboard motor stress,for a lake saturated
by motorboat activity, was thus calculated to be nine gallons
of fuel burned per acre (84.15 liters of fuel burned per
hectare). Finally, for the northern study area, a boating
season of eight months was assumed, while in reality the
normal yearly boating season is only four to five months in
duration, which represents a total yearly stress of
approximately 300 gallons of fuel burned per acre (2806 liters
per hectare).
These calculations (yearly stess levels of 300 gallons of
fuel burned per ace), however, provides no information as
to the stressing lwels ii-. terms of water volume, since no
restriction has been placed on the depth of the lake. Obviously
certain boating activities would be limited without sufficient
water depth. A twelve foot (3.66 meters) average lake depth
was assumed for purposes of placing the stressing level in
term.-. of water volume. It was felt that this depth would be
the minimum required for such nor.-restricted boating activity,
and a lake of this depth would most likely be the first to
experience stress, if any. due to two-cycle outboard motor
operation. The yearly stress — based on an eight month boating
period and a lake with an average depth of twelve feet (3.66
meters)-- is thus 75 liters of fuel burned per million liters
of water. The daily stress over an eight "jnth boating period
would then be approximately 0.33 liters or fuel burned per
day per million liters of water.
This ;.evel of stressing is shown graphically in Figure U. Also
shown in this figure arc the levels reported by English, et.
al_- */f for fish tainting and acute fish toxicity of 3 and 500
liters of fuel burned per million liters of water respectively.
It is evident from the intersection of the stress line for a
twelve foot (3.66 r.ieters) lake with the reported toxicity
level:- • that:'the chances for observing effects from two-
cycle outboard motor exhaust emissions would be minimal.
i j
-------
FIGURE 4
Design Stress Levels
Acute Fisli Toxicity - 96 l»r. TLm (500 litors/10 liters)
Stressing Level for Northern Study
(3 tine " maximum rate)
Maximum Stressing Level for 12 ft
~ Lake (calculated)
STRESS TIME
MONTHS
-------
Consequently, the calculated maximum stressing schedule es-
tablished for this study was tripled. Thus, the stressing
level used in :he field studies for :;h;s project was approxi-
mately one liter of fuel burned per day per million liters
of water -- see Figure 4.
18
-------
SECTION IV
LABORATORY STUDIES
ENGINE OPERATION AND TESTING
A total of 12 different engines representing a variety of
two-stroke configurations, were studied. They ranged in size
from 3.6 hp to 105 hp and represented a spectrun of engines
from four manufacturers: Outboard Marine Corporation, Mercury,
Chrysler and Tecumseh Productions. Crankcase designs included
closed, recirculating, and drained types. Ignition systens
represented both conventional and high-voltage capactive-
discharge (CD) system. Two engines (Chrysler 3.6 hp and Eska
7.0 hp) wer^ air-cooled, the rest were water-cooled. Carbura-
tion varied both with manufacturer and engine size.
The engines tested (excluding two "maintenance study" engines
tested after use in associated northern field studies), to-
gether with a number of pertinent specifications arc shown
in Table I. Each engine was instrumented and fitted with
exhaust gas saaple probes.
The larger engines (35 hp to 105 hp) were rnojntcd on a special
test stand and coupled to a 120 hp General Electric dynamometer.
Smaller engines (3.6 hp to 18 hp) were operated in a test
tank and were loaded using test propellors.
Mass flow of fuel was obtained by measuring initial and after-
test fuel weight. Because each test was run at a constant
speed and Load, ir. was nor necessary to continuously monitor
fuel rate. Any change in engine operation was readily detected
from the continuously monitored exhaust emission data. Engine
speed was measured with an optical electronic tachometer on
the test-tank engines.
Three engines were operated with external crankcase drainage
systems installed. The engines selected were similar to three
of the drair.less engines tested, an 18 hp Evinrude, 35 hp
Chrysler, and a 50 hp Mercury. One of the engines, the 35 hp
Chrysler was an older model which was normally drained. Both
t'ne 18 hp Evinrude and j0 hp Mercury were 1972 model engines
vhich were converted from a recirculating or drainless type
engine to a drained type.
Representative and known fractions of exhaust were obtained
with special sampling probes constructed of 1/4 inch O.D,
stainless steel tubing and designed as static probes with a
number of sample holes on the circumference and spaced longi-
19
-------
TABLE *
TEST KNGI Nl-,S AND SPECIFICATIONS
Di splace-
Brand Model Serial No. oC merit Gear Cooling Crunkcaso Fuel/
Njiiic Rated hp t> RPM Number Number Cvl . (cu. in.) Ratio System Drain Oil Initio
Chrysler
105
5000
.1057 IIC
9 994
4
90.55
15/26
Water
Recycled
50/1
Mercury
50
5 3 00
500E
3.17050]
4
43. 8
1/2
Water
Rccycled
50/1
Chrysler
35
4750
3 501115
1053
2
35.9
1 3/21
Water
Recycled
50/1
Chrysler
12.9
5000
1 2 2 HA
3725
2
13.62
14/22
Water
Recycled
50/1
Mercury
7 . 5
5500
75
319 0951
2
10.9
1/2
Water
Recycled
50/1
Eska
PO
7 . 0
5G50
1747A
153145
1
7.5
14/21
Air & Water
Closed
24/1
o
Chrysler
3 . 6
4 500
32MB
G 34 G
1
CO
»—<
14/21
Air i Viator
Closed
16/1
Evi nrude
6
4 500
G202DJ
03899
2
884
15/26
Water
Recy -:led
50/1
Evinrude
18
4500
18202R
E03098
2
22.0
12/21
Water
Recycled
50/1
Chrysler
15
4 7 50
355IIC
6336
2
35.9
13/21
Wjter
Drained
50/1
Evi nrude
] 8 *
4 500
18202R
E03 098
2
220
12/21
Water
Dra ined
50/1
Mercury
50*
5300
5 00K
3170501
4
43. fl
1/2
Water
Dra ined
50/1
*Modificd to a drained engine
-------
tudinally along the length. Particular attention was given
to the location of the probe in the ongine exhaust system
to insure representative sampling of the average composition
from all cylinders in the cuIti-cyIinder engines and to
minimize tine and spatial resolution problems. The location
chosen was the position farthest down the exhaust collector
before any water was added to the exhaust stream.
The sample line between the engine and condensate apparatus was
maintained at 350°F. At this elevated temperature no conden-
sation of water and only limited hydrocarbon condensation
occurred on the tubing walls. This temperature was still low
enough to prevent sigrIficant reaction of either hydrocarbons
or CO with the oxygen present.
The exhaust sample flow and analysis system, shown schenatically
ir. Figure 5. allowed three separate modes of analytical opera-
tion, non-dispersive-infrared (NDIR) analysis of gas phase,
flan:e ionization detector (FID) hydrocarbon analysis, of gas
phase or extended condensate collection. Both the NDIR and
FID analysis systems could be purged with room air and cali-
brated with span gas without affecting the flow through the
condensate or liquid sample collection apparatus. All of
the analyzers with the exception of the hydrocarbon flame ioniza-
tion detector (FlD)were connected in series because they
arc non-destructive to the sample.
A specially constructed subtractive column analyzer was used
in conjunction with the FID. This device permitted separation
of the gaseous hydrocarbons in the exhaust sample intc prin-
cipal family components: paraffins, olefins, and aromaries.
Indolene 30 was used as the test fuel. All of the fuel used
was from the same refinery batch, thus insuring consistent
hydrocarbon fan.ily composition. Quicksilver outboard motor
oil was used as the lubricant in all of the engines. It was
mixed wirh the fuel in the ratio specified by each manufacturer.
(See Table I).
The engines were run only under steady-state conditions, after
being first broken in according to the manufacturer's recom-
mendations. A simulated boat load curve was used to establish
the proper speed-horsepower relationship. Generally an out-
board engine follows approximately a 2.5 order load curve on
a planing hull.
GAS-PHASE EMISSION'S
Carbon Monoxide
Carbon monoxide is a moderately toxic compound with limited
solubility in water. Averaee carbon monoxide emissions in
21
-------
ENGINE
ro
no
HEAT TAPE
ROTAMETER
O n-
FLOW
CONTROll.
—§—
FILTER
PROBE
PUMP
n
PUMP
jg FILTER
PUMP
ACETONE MERCURY
WATER 8 DRY ICE MANOMETER
No CALIBRATION No
GAS
ROTA
METER
PUMP
FILTER
SUBTRACTIVE FID
COLUMN
2 ww2
ANALYZERS
NDIR HC
Fig, s Schematic Diagram of the Sample Collection and
Analysis System.
-------
percent, varied from 4-1/2 percent at 1000 rpms to a high
of 6-1/2 percent at 3000 rpms. The results of percent CO
emissions at 4500 and 5000 rpms have not been reported in this
"average" data analysis because data at this speed were taken
from only a few engines. No trend in CO emissions is evident
from these results. The CO emissions are high compared to those
observed from current four-cycle automotive engines.
Carbon Monoxide is basically a function of the engine air/fuel
ratio. The high CO emissions of the two-cycle marine engines
are attributable to a rich mixture ratio The air/fuel ratio
used on all of the test engines, generally between 10 and 13
to 1 depending on carburation requirements for the various
engines, was richer than the stoichiometric ratio of approxi-
mately 14.8 to 1. This richness is necessary to assist
with internal cooling and lubrication, and provide smooth
operation even with substantial internal exhaust residual
dilution which is a problem of all two-cycle, crankcase
scavenged engines.
Another factor influencing carbon monoxide emission is the
trapping efficiency of the engine, which is defined as the
ratio of the mass of fuel and air trapped in the engine to that
which is furnished to the engine. The lower the trapping effic-
iency, the greater the o.uantity of unburned fuel and air in
the exhaust and, for a given mixture ratio, the lower the CO
emissions. The unburned fuel/air in essence dilutes the ex-
haust products.
"arbon Dioxide
Carbon dioxide is not considered a pollutant in normal atmos-
pheric concentrations, however it is a very significant
constituent in gasoline engine exhaust. The average carbon
dioxide emission results from the various test engines ranged
between 5,-1/4 percent at 1000 rpm to 7-1/2 percent at 4,000
rpm in an.almost linear fashion. As with CO, the concentration
of CO2 is strongly a function of the air/fuel ratio used in the
engine. Trapping efficiency also influences CO2 concentration.
In general,'-., CO2 concentration should be inversely proportional
to carbon monoxide concentration, although this is not evident
when comparing the average CO and C00 emission data for a
number of engines.
Hydrocarbon Emissions
Only the flame ionization detector (FID) hydrocarbons will be
discussed since this nurr.ber is representative of the total
hydrocarbon concentration rather than the partial hydrocarbon
23
-------
fraction shown by the non-dispersive infrared (NDIR) analyzer.
Average total hydrocarbon concentration in the gaseous exhaust
emissions data from all test engines at each speed varied from
a high of 7.75 parts p
-------
105 hp Chrysler, and one of the smaller engines, a 6 hp
Evinrudc, show that Che concentration emissions are
similar. Thus, the major factor causing higher mass
emissions from large engines is the greater air and
fuel flow rate. Brake horsepower is clorely related to
mass flow rate. In general the mass rate increases with
both speed and horsepower.
Hydrocarbon Family Analysis
The exhaust hydrocarbons were separated into their primary
family constituents (paraffins, aromatics and olefins)
with the aid of the subtractive column analyzer.
In general the results from testing of the two-cycle out-
board engines differ significantly from those observed
with the conventional f-jur-cyclt; engines in that the
exhaust hydrocarbons appear to be more closely related to
the fuel. The major difference observed between the exhaust
hydrocarbons and the fuel was the olefin concentration,
which is moderately gre-Jter in the exhaust gas, averaging
between 20 and 30 percent of the total hydrocarbon composi-
tion whereas in the fuel the olefin fraction was six percent.
The increase in olefins is a result of "restructuring",
primarily of paraffinic hydrocarbons during combustion. Most
cf the reaction occurs in a "quench zone" located between
the hot flame and the relatively cold walls of the engine
c^rn'oustion chamber.
The aromatic fraction is slightly less than observed in the
fuel and is generally in the range of 20-30 percent of the
total hydrocarbon composition. As a class, the aromatics
appear to be relatively stable.
Paraffinic hydrocarbons generally are the most prevalent
of the exhaust hydrocarbons, comprising approximately 50
percent of the total. However, in some instances the
paraffinic fraction, such as with the 6 hp EvinruJe engine,
was nearly 70 percent of the total and very simila.- to the
fraction of paraffins observed in the tcsc fuel. Other
engines, such as the 3.6 hp Chrysler, exhibited only
approximately 45 percent paraffinic hydrocarbons under
several test conditions.
No discernible trend is evident with regard to the hydro-
carbon family breakdown'as a function of engine speed and
load.
?5
-------
CON DESSA BLE COM] 'ONENTS
General
The purpose* of this part of the study was to collect and
identify condensable fractions of outboard engine
exhaust emissions under controlled laboratory conditions,
in a canner such that the contribution of this fraction
of the exhaust could be accurately assessed. As noted
previously, condensable fractions are those most likely
to be transferred from submerged exhaust emission to
the water column.
To this end, engines were mounted on test stands, as
described in a previous section of this report. The exhaust
was passed through the collection and gas-phase analysis
systen shown schematically in Figure 5. Collection of
condensate samples required three hours of continuous
engine operation at each rest condition. The system was
arranged so that a constant 2 cu. ft./hv. (0.56 cu.m/hr)
was passed through the condensation section. It was
possible to calculate condensables in the total exhaust
by measuring the hydrocarbon flux with and without the
collection apparatus in operation.
The condensation system consists essentially of an efficient
cold water condenser and a dry-ice cold finger trap. At the
end of each run both condenser systems were washed down with
a fixed volume of CHC13, and the condensates and washings
from th-'i two receivers combined. Due to the large condensing
surface area and the low temperatures used in collecting
the condensate, it is virtually certain that this system
condenses at least as much - and orobably significantly
more - of the exhaust than would a column of water in a lake
or other water body under nonral engine usage. Consequently,
the amounts reported herein should represent maximum potential
pollution loads rather than average values.
The composition of the condensable organic portion of the
exhaust has been shown to consist of three fractions:
unburr.ed gasoline, partially oxidized hydrocarbons (i.e.
phenols and carbonyl compounds) and unburr.ed oil^. Further,
the unburned fuel is composed of three fractions, aromatics,
olefins and paraffins. Since it has been established that
aromatic compounds constitute the most inporr;ant of these
in terms of acute toxicity to aquatic life^, primary
attention was directed to the qualitative and quantitative
analvsis of this fraction.
26
-------
The condensates were analysed for arcmatic hydrocarbons,
olefins, phenols, carbonyl compounds, paraffinic hydrocarbons,
and total condensable organic material. The respective
analytical techniques and methods used are outlined in detail
in the University's final report to the Environmental Prot-
ection Agenc^. Only the results and discussions of these
analyses arc presented in this summary report.
Aromatic Hydrocarbons
Samples of the fuel used in the engine tests were examined
first by infrared spectroscopy and then by gas chromatography.
Identification of the major peaks was done by mass spectral
analysis and confirmed by comparing the retention times of
the pure components with those of the naior peaks in the
exhaust chrorcatograrn. A comparison of the ratio of toluene
to other aroraatics in the condensate and in the fuel showed
chat in most cases the ratio is lower in the condensate, sug-
gesting that either toluene is preferentially burned or that
the higher boiling aroraatics are preferentially condensed,
or both.
Binuclear arocatic hydrocarbons were also detected in randomly
selected samples of the condensable fraction. In particular,
naphthalene and its two isomeric methyl derivatives were
identified by comparison of their retention times with those
of the pure compounds. In six samples analyzed, the three
binuclear aroraatics ranged from 1-2 percent of the total
mononuclear aroir.atic content of the condensate. This level
is somewhat higher than that in the raw fuel (0.5 percent
binuclear aromatics).
Olefins
Olefins were determined by amperometrie titration using a
standard ASTM Method . Phenols interfered and w-^re removed
by extraction with 0.1N NaOH. Olefins are reported as cvclo-
hexene (molecular weigRt 86).
Phenols
Phenols verc determined on the base extract from the olefin
determination by ultraviolet spectroscopy. Dje to the
high background absorbance hovevc-r. it was necessary to
take the difference in absorbance at-,290 nri.llinierons
between an acidic and basic solution'.
27
-------
Carbonyl Compounds
Carbonyl compounds were determined by direct infrared spec-
troscopy on the organic phase oi: the extracted condensate.
No attempt was made to distinguish between the various
types of carbonyl compounds; the peak at about 5.9 microns
was measured and compared with known solutions of butyralde-
hyde. Bar.e extraction lowered the absorbance in a random
selection of samples by 7 - 12 percent, indicating an acid
content of roughly 1/10 of the carbonyl signal.
Paraffinic Hydrocarbons
Paraffinic hydrocarbons wsre not determined directly: however,
paraffins were estimated by obtaining the total condensable
hydrocarbons, subtracting the major constituents (olefins
and aromatics) and assuming the remainder to be paraffins.
Total Condensable Organic Material
Determination of the total amount of organic material con-
densed from the exhaust is complicated by the fact that con-
siderable water is formed during tne combustion process and
condensed in the traps. For one particular engine, the
normal addition of CHCl^ to the condensate was eliminated.
Instead, the weight of total condensate (water plus organics)
was recorded. A known weight of tetrahydrofuran was added
such that the two phases initially present were completely
miscible. Water was determined by Karl Fischer titration
and t:he weight of organic material determined by difference.
RESULTS
Exhaust Condensate
Results of the exhaust condensate analysis are presented
in Table II. The last two columns entitled Total Condensable
Mononuclear Aromatics were calculated as follows: the
TCMA quantity in grams/hr was calculated by multiplying the
observed condensable mononuclear aromatics in mg/eu ft. by
the total exhaust in cu ft./hr and dividing by 1000 to convert
mg to grams; the TCMA quantity in grams/kg was calculated
by dividing the grams/hr value by the fuel consumption rate
(kg/hr) to give grams of condensable mononuclear aromatic
per kg. of fuel. Condensalles concentration was measured
in duplicate runs on the first few engines tested. The
reproducibility vas very good ar.d therefore it was decided
that single condensate collections would suffice thereafter.
28
-------
Table II
SUMMARY OK CONDENSATE ANALYSES
Engine
Fue 1
Total
Concentrations
of Condensable
Total Condensable
and
Consumption
Exhaust
Substances:
(mp,/cu.
Ct.)
Mononuclear Aroma-
speed
Rate
(cu.ft. /
Mononuclear
Olefin*
Phenol •'
Carbonyl
tics
(rpm)
(kg/hr.)
hr. )
Aromat ic
(g/i'V. )
(g/kg fuel)
Chrysler
3.6 hp (air-cooled)
1000
0. 322
111
98
lo. 0
-
-
10.9
33.9
2000
0. 5S9
190
113
20.3
-
-
21.5
36. 5
3000
0.815
281
55
15.3
-
-
15.5
19.0
4000
1.060 .
407
47
10.2
1.53
10.8
19.1
18.0
4500
1. 287
482
55
10. 3
1.26
10.2
26.5
20.6
Mercury 7
.5 hp
1000
0.5559
205
105
19. 3
-
7.0
21.5
38.5
2000
1.074
383
70
14.1
-
5.8
26. 8
25.0
3000
1.628
849
68
-
-
6.8
57. 7
35.4
4000
2. 220
905
36
-
-
7.2
32.6
14. 7
Chrysler
12.9 hp
1150
0.900
313
145
18.8
1.15
9.5
45.4
50.4
2000
1.432
495
159
19.2
1.07
14 5
78.7
55.0
3000
1. 770
656
73
11.6
1.57
9.5
47.9
27.1
4000
2. 952
1070
119
15.0
1.82
18.3
127
43.0
5000
4. 298
1575
132
17.2
3. 37
12.5
208
48.4
Evinrude
18 hp
1500
1.821
599
217
36.0
4.70
-
130
71.4
2000
2.400
810
165
13.7
1.65
11.8
134
55.8
3000
3- 740
1220
77
13.0
1.43
12.5
93.9
25.1
4000
5. 249
1818
140
15.8
1.80
-
255
48.6
4900
7. 398
2478
142
12.0
1.05
19.5
352
47.6
Evinrude
18 hp - drained
1500
1.724
581
229
20.3
1.48
26.8
133
44.8
2000
1.926
692
123
17.3
2.08
30.0
85.1
44.2
3000
3.049
1064
107
17.6
1 .92
14.0
114
37.4
4000
5. 873
1905
165
17.3
1.08
13.3
314
53.5
5000
6. 783
2347
160
16.3
0.80
9.8
376
55.4
-------
Table II (continued)
SUMMARY OF CONDENSATE ANALYSES
Engine
Fuel
To La 1
Concentration
of Condensable
Total Condensable
and
Consump tion
Exhaus t
Subs tances:
(mg/cu. f t
¦ )
Mononuclear Aron-.ati.es
Speed
Ra ce
(cu. ft:
./
Mononuclear
(rpm)
(kg/hr.)
hr.
)
Aromatic
Olefin*
Phenol*
Carbonyl
* (g/hr.) (g/kg fuel)
Mercury
50 hp
1500
3.477
1266
104
14.2
0. 53
6.7
132 38.0
2000
3. 540
1370
73
9.8
0.81
7 . 3
100 28.2
3000
6. 581
2326
88
9.3
0. 5<:
5.2
205 31.2
4000
9. 82S
3643
63
7.3
0. 52
6.5
230 2 3.4
Mercury
50 hp - drained
1500
3. 779
1376
85
-
-
5.8
117 31.0
2000
3. 784
1460
70
-
-
6.8
102 27.0
3000
6. 787
2404
73
-
-
5.5
175 25.8
4000
10.390
3824
78
-
-
7.7
298 28.7
Chrysler
105 hp
1000
4.472
1550
150
20.0
-
28.6
233 52.1
1500
5.655
2094
86
19.7
-
19.0
180 31.8
2000
6. 557
2513
91
19.5
2.05
62 . 2
188 23.7
3000
9.666
3663
76
14. 5
1.85
66.0
278 28.7
4000
16.636
6369
74
16.5
-
29. 3
467 28.1
Evinrude
18 hp (field
engine .is
received)
1000
1.460
475
123
-
-
8.8
58.4 40.0
2000
2 . 195
760
56
-
-
8.1
42.6 19.4
3000
3.493
1163
45
-
-
7.2
52.3 15.0
40C0
5.073
1302
43
-
-
9.0
77.5 15.3
Evinrude
18 hp (field
eng
i.ne new
plugs)
1500
1. 870
-
71
-
-
7.2
-
2000
2.140
7 36
73
-
-
10.1
53.7 25.1
3000
3.695
1247
49
-
-
7.8
61.1 16.5
4000
5.073
-
40
-
-
7.8
-
4500
7.000
-
45
-
-
6 .4
-
-------
Most of the data, therefore, was obtained in this way.
As can be seen from Table II, condensable aroma tics amount
to from 1.5 to 7.1% of the fuel fed <15 to 71 g/kg) with
most values between 2 and 5%. An estimate of the total
amount of hydrocarbons which would be condensed in normal
use situations can be arrived at by reference to Table III
and Figure 6. Aromatics and olefins were determined and
the difference between the total weight of condensable
organics and the sum of olefins and aromatics was assumed
to be paraffins. As can be seen from Figure 6 and Table III,
over the range of speeds studied, the amounts of olefin and
paraffin present in the total(gas stream) exhausu varied
markedly; nevertheless, both the percent aromatics in the
condensed phase and the percent of the total hydrocarbons
condensed in the cold water trap varied between 20 and 257c.
The range of family composition of total exhaust shown in
Figure 6 for the smallest engine (a 3.6 h.p. Chrysler)
encompasses almost the exact range found for the largest
engine (105 h.p.) as shown in Table II. Table II indicates
that the total condensable aromatics for the 3.6 h.p. engine
are about average for all engines. Thus to find total
condensable hydrocarbons, one can multiply the total conden-
sable aromatics by 4 or 5. To determine how much of the total
condensable hydrocarbons are found in the cold water trap
one can divide the total condensable hydrocarbons by a factor
of 4 or 5. It is felt that the amount condensed in the cold
water trap (1-3° C) best approximates the amount that would
be condensed in a water column, as the temperature of
condensation in this trap is much closer to that of water in
a lake or pond chan is that in the dry ice trap. One arrives,
by means of these estimates, at a value of about 1.5-7 for
that percent of fuel-_fed to the engine which would be condensed
in the water column.
If 3000 r.p.m. is selected as the speed at which most boat
usage occurs, and the values of total condensable aromatics
as percent of fuel is averaged over all engines tested in thi<=
study, a figure of 2.5% condensable aromatics is obtained.
Therefore, the conclusion may be reached that the "average"
engine will contribute about 2.5% of its fuel to the water,
exclusive of drainage, during most of the time it is in use.
Influence of Maintenance
Two engines, an 18 hp Evinrude and a 35 hp Chrysler, were
investigated for the influence of maintenance on emission
performance. These engines were of similar model and horse-
power to two new engines tested, but were engines that had
31
-------
80
z
9 60
H
CO
I
8 40
g?
20
0
FUEL 1000 2000 3000
R.P.M.
C = CONDENSATE ANALYSIS
G = GAS STREAM ANALYSIS
ORDER FROM BOTTOM = AROMATIC, OLEFIN,
PARAFFIN
Fig. 6 Composition by Hydrocarbon Families of Condensate
and of Exhaust Gas Stream, Superimposed over that
of Raw Fuel. Engine: Chrysler 3.6 HP
3?
-------
TABLE III
CONDENSABLE MATERIAL IN EACH STAGE AS WATER AND
ORGANIC MATERIAL (CHRYSLER 3.6 HP ENGINE)
% of Total
Total Wt. Total % Condensable
Sample (g) Organics (g) Organics Organics
1000 r.p.T,
Stage I (1-3°C) 34.52
Stage II (-65"C) 3.97
2000 r.p.m.
Stage I <1-3°C) 15.68
Stage II (~65°C) 3.58
3000 r.p.m.
Stage I (1-3°C) 16.12
Stage II (-65°C) 2.81
0.47 1.4 18
2.18 55.0 82
0.50 3.2 19
2.16 60.5 81
0.43 2.6 24
1.34 47.7 76
33
-------
been used extensively in associated northern field studios.
The 18 hp engine had bec-n operated for 81 hrs. and the 35 hp
engine for 67 hrs. in the field. The intent was to run the
engines as received, perforc normal isaintenance and then rerun
the engines again to compare emission performance. These
engines were subjected to the sane analyses as the new engines,
including gas-phase and condensate analyses.
The maintenance procedures resulted in only very ninor changes
in gas abase emission portom-ince as is shown ill Table II for
the IS hp Evinrude. The naximum variation in the air/fuel
ratio was only 4 percent. This is not a significant varia-
tion and supports the findings that there is little dif-
ference in emission performance.
The only trend toward a difference was exhibited by the 18
hp Evinrude engines, for which the field engine gave lower
amounts of condensabIcs than did the laboratory engine. This
small difference rcay be accounted for by the fact that there
are sone differences to be expected even within a given engine
nodel.
Crankcase Drainage
Three engines were investigated with internal crankcase drain-
ago rather than the recycling system used on all current
engines. The engines selected were the 35 hp Chrysler. 50 hp
Mercury and 18 hp Evinrude.
A relatively wide variation was observed between the three
test engines. The Chrysler in particular exhibited substantial
drainage (11 percent) at the low-speed test conditions. How-
ever, as with all of the engines, the drainage tended to
decrease with an increase in engine speed. Decreased drain-
age with increasing speed was expected because the greater
air flow at higher speed results in more crankcase turbulence.
No explanation is evident for the drainage differences
observed between the engines other than it is related to
particular design characteristics.
A r.axiniu.- drainage rate . of approximately 300 grarr.s per hour
was observed at the 1500 rpn tent condition with the Chrysler
and a mini nun rate of 39 grams per hour at &000 rpm with the
18 hp Evinrude engines. In teres of the percentage of fuel
used, the drainage observed ranged from a maximum of II per-
cent at 1500 rp^i for the Chrysler engine to less than one per-
cent for the Mercury 50 hp engine at the 3000 rpro test
condition. At low speed substantial variation was observed in
34
-------
the results from run to run for a given engine. For example,
with the Hercury 50 hp engine at 1500 rpm a drainage rate
of 19 7.0 gm/hr. was observed as the speed was increased
to the test speed, whereas a drainage rate of 3.7 gras/hr.
was observed when the speed was decreased to this rest speed.
This suggests that a hystersis effect is present such chat if
one would approach a given speed from one side a different
crankcase drainage would be expected than if one approached
the speed from the other side. This is probably related to
the slight differences in combinations of spark advance and
throttle settings required to maintain a given performance
level. This difference is amplified by the steep slope of
the drainage curve evident in the low speed range of the
engine.
The gaseous emission performance of the "drained" engines
compared favorably with that observed from the "drainless"
engines. No significant differences were noted in the 18
hp Evinrude and 50 hp Kercury "drained" or "drainless"
engines as shown by the data in Tabic II.
The oil composition of the crankcase drainage was about 20 -
30 percent. Since the ratio of oil to gasoline in the fuel
was 1:50 (2 percent) crankcase drainage roDresents a 10-15
fold increase in oil content over the mixture fed to the
eng ine.
Evaporation Studies
Once the exhaust products are condensed in the water column
there exist several mechanisms by which they can be removed
from the system. One of the most important of these is the
process of evaporation. In this particular study specific
attention was paid to the evaporation of the aromatic com-
pounds present in OME water.
Solutions containing outboard engine exhaust were prepared by
running a small engine (1-1/2 hp) in a 55-gallon drum for a
period such that an organic phase did not separate in the
receiving vator. Solutions of pure aromatic components were
also prepared by dissolving these compounds in water and
stirring briefly in a closed volumetric flask. Concentration
of exhaust products and of pure compounds was monitored by
observing ultraviolet absorba.nce of the test solutions at
250 millimicrons under controlled conditions of temperature,
initial concentration, turbulence and surface to volume ratio.
As the ultraviolet absorbance of total outboard exhaust
35
-------
product decayed with time to a final non-zero value (see
Figure 7), the rate studies were corrected for this value
so as to represent the disappearance of volatile aromatic
components only.
As expected, agitation in the forra of stirring or aeration
markedly increase.*! the rate of evaporation. Detailed
evaluation of the evaporation kinetics of the OME water
indicate that the eronatic component removal may be described
by a first-order rate expression. The half-life for volatile
aromatics uniformly dispersed co a depth of 1 meter in a
quiescc-rtt body of wrter at a temperature of 20°C was determined
co be approximately 11 days. Rate of loss wc-jld course be
lowesc in a quiescent vacer body. This condition of no
turbulence would be most unusual in a natural lake situation;
thus, the half-life of 11 days is highly conservative. A
fairly rapid disappearance of volatile aronatics from the
condensable phase can therefore be expected in the natural
varer systems. Half-life values of an order of magnitude
( :' 1 day) less rhan the quiescent value has been observed in
the Laboratory under aerated conditions (see Figure 7).
Fish Toxicity Studies
Fish toxicity studies were undertaken to determine the
effect of OME water op. the fish populations. Because of the
fairly rapid evaporation of toxic components (described in the
previous section), conventional static fish toxicity studies
were deemed unsuitable, and a dynamic testing system® was
devised whereby fresh solutions of the toxic components studied
were continuously fed to test aquaria.
Outboard engine exhaust condensate was prepared by running a
small outboard engine (1-1/2 hp Johnson) in a 55-gallon drum
filled with water. This condensate water was then diluted
with dechlorinated tap water to p.ive a ranp.e of desired
concentrations. Carbon monoxide, a product of the fuel
burning process was evaluated as a potentional source of
fish mortality in both the dynamic test system and in separate
static tests by bubbling L'he pure p,as through the test water.
Mortality was expressed as TL;.< values for various ueriods of
exposure, TLm being the concentration at which 50 percent
mortality occurred.
TL;.f's for xylene and toluene are p.iven in Table IV. Table V
yjives TL>;'s for outboard enr.inc exhaust water. The TLm
values for the exhaust components is lower than that for the
pure aromatic compounds but of the same order of magnitude.
It should also be noted that the 96-hour TLm value for the
pure aronatics and the condensate are two to three orders
36
-------
00
OO QUIESCENT
A A AERATED
80
60
40
20
20
DAYS
40
30
DiSAPPEARENCE OF OUTBOARD EflGlNE EXHAUST PRODUCT
from Aerated and Quiescent Aqueous Systems at
Room Temperature.
37
-------
TABLE IV
Median Tolerance Limits for Goldfish
Expose J to Aromatic Hydrocarbons
Compound
Toluene
Xylene
TL^ Values (ppm)
24 hr. 48 hr. 72 hr.
41. 6
30.6
27.6
25.1
25.3
20.7
96 hr.
22.8
16.9
TABLE V
Median Tolerance Lii.iits for Goldfish
Exposed to Outboard Engine Exhaust Water
Fue 1
Exposure Time
Non-leaded
TU,
>1
24 h r.
212"
(1/4720)**
[12 F**
48 hr.
194
(1/51«0)
ill]***
72 hr.
185
(1/5410)
[lajw-v*
96 hr.
168
(1/5960)
[8 ]*•'-*
Leaded
226
(1/4420)
204
(1/4920)
192
(1/5220)
171
(1/5860)
*values given in terms of gallons of fuel burned per million
gallons of water
**values in parentheses given as the volume ratio of fuel to
dilution water
***vnlues in brackets express aromatic fraction of OME water
in terms of ppm of toluene
38
-------
of magnitude higher than those aromatic levels found in
test pond used in the associated northern field studies
which had been subjected ro outboard usage at very high
levels (Table VI). Carbon monoxide, even at near-satur-
aticn levels, did not produce fish mortality. Thus it i
concluded that outboard exhaust is net implicated in
acute toxicity under normal boating conditions.
TABLE VI
UV-Analytical Data fov Saline Test Ponds (Fall 1971)
(Results calculated as ppm (Vol.) toluene)
Dev.. 5
Oct. 12
Oct. 20
Pond £1
(non-leaded)
0.2 7
0.24
0.08
Pond $2
(control)
0.09
-
0.07
Pond ;?3
(leaded)
0.19
0.20
0.12
Pond i':h
(control)
0.09
0.11
0.09
39
-------
srcn: \ v
PIT:! JI SlinilliS
>JORTIIP.RN PUT!) STUMPS
Stress 1.cvc* Is
The engines used in tlie field stressing were provided by the three
mior two-cycle outboard engine manufacturers, Outboard Hirine
Corporat]on, Mercury Marine Corporation, and Chrysler Corporation,
from standard production models (see Tabic VII). Pcfore use in the
field each engine was "aged" by running the engine for approximately
50 hours. This 50 hour period is equivalent to one year of normal
operation Ihe engines employed in the field studies over the
three year test pe:iod ranged in size between 2 and 50 horsepower
and as stated previously were a mixture from the three above mentioned
mnufacturers. The engines were rotated on a random basis over the
three year test period. At no time during the three year period were
any of the engines tuned or maintained.
In general, over a week stress period, the engines were operated at
1/4 speed (10f)0 to 1500 rpns) for 25 percent of tlie timo and 5/4 speed
(3000 rpns) for 75 percent, of the time. Tlie time and speed settings
were agreed upon values used to represent normal boating operations.
Ten weeks after the lakes were divided outboard engine stressing be-
gan. The stressing schedule for the entire test period is shown in
Table VIII. This table surmarir.es tlie total fuel inputs into both
the Ic-.ided and non-leaded test sections and the engines used during
the study.
The 'uel utilized throughout the study was Pndoline 30 for the leaded
test section (this gasoline contains 5.1 grams of t.ctraethyl lead per
gallon) rind ,'ndoline clear ("no lead) for the non-leaded test section,
roth of these gasolines arc produced by the Standard Oil Corporation
to rigid specifications and are commonly used in engine testing pro-
grans where fuel characteristics arc deemed necessary.
Standard two-cycle engine oil was added *o the gasoline at the recom-
mended T"i t i o of one part to fifty parts gasoline (1:50).
Methodology
In general all chemical and biological sampling was performed by rou-
tine standard procedures. A detailed description of sampling proce-
dures and frequency together with tlie chemical and biological analytical
40
-------
Table VII
Engines Used During Tent Period
Manufacturer
Model No.
Engine Serial No.
Engine BHP
Model Year
Motor Type
Mercury
15002
3238485
50 hp
1972
drainless
Mercury
15002
3238629
50 hp
1972
drainless
Mercury
10402
3297340
4 -hp
1972
drainless
Mercury
10402
3297276
4 hp
1972
drainless
Evinrude
18202
08215
18 hp
i972
drainless
Evinrude.
18202
08246
18 hp
1972
drainless
Evinrude
2202
01036
2 hp
1972
drainless
Johnson
2R72
3530006
2 hp
1972
drainless
Chrysler
354HD
21581
35 hp
1972
dra inless
Chrysler
354HD
21588
35 hp
1972
drainless
Chrysler
82HB
11132
8 hp
1972
drainless
Chrysler
82HB
11152
8 hp
1972
drainless
-------
TABLE VIII
STRESS LEVELS
A. Non-Leaded Test Section
Date
Engine
Stress Rate
Cummulative Stress
gallons burned
gallons burned
per million gallons
per million
per day
gallons
1971
22
Sep - 4
Oct
2
Hp
Evinrude
2.4
28.8
4
Oct - 18
Oct
4
Ho
Mercury
2.A
62.4
18
Oct - 1
Nov
8
Hp
Mercury
2.4
96.0
1
Nov - 15
:iov
4
Hp
Mercury
2.4
129.6
15
N'ov - 29
Nov
e
Hp
Chrysler
2.4
163.2
29
Uov - 1
Dec
2
Hp
Johnson
2.4
168.0
1972
1
May
-
6
May
8
Hp
Chrysler
0.3
169.5
8
May
-
13
May
18
lip
Evinrude
0.5
172.0
15
May
-
20
May
35
Hp
Chrysler
0.7
175.5
22
¦Hay
-
27
May
4
Hp
Mercury
1.0
180.5
27
May
-
12
Jun
50
Hp
Mercury
1.0
196.5
12
Jun
-
26
Jun
2
Hp
Johnson
1.0
210.5
26
Jun
-
10
Jul
8
Hp
Chrysler
1.0
224.5
10
Jul
-
24
Jul
18
Hp
Evinrude
1.0
238.5
24
Jul
-
7
Aug
35
Hp
Chrysler
1.0
252.5
7
Aug
-
21
Aug
4
Hp
Mercury
1.0
266.5
21
Aug
-
4
Sep
50
Hp
Mercury
1.0
280.5
4
Sep
-
18
Sep
2
Hp
EvinruJs
1.0
294.5
18
Sep
-
2
Oct
3
Hp
Ch rysler
1.0
308.5
2
Oct
-
16
Oct
18
Hp
Evinrude
1.0
322.5
16
Oc t
-
30
Oct
35
Hp
Chrysler
1.0
336.5
30
Oct
-
13
Nov
4
Hp
Mercury
1.0
350.5
13
Nov
-
27
Nov
50
Hp
Mercury
1.0
364.5
27
>ic V
-
1
Dec
2
Hp
Evinrude
1.0
368.5
1973
15
May
_
1
Jul
18
Hp
Evinrude
1.0
421.0
1
Jul
-
20
Aus
4
Hp
Mercury
1.0
4 78.0
42
-------
techniques-emploved arc presented in the final report on this phase
of the project to the United States Jinvironrcental. Protection Agency.
Results and Discussion
Hiologicnl
Pcriphyton Corinunity - Artificial substrates Kcrc USC(j to sample
tlie pcriphyton community of the pond systems during 1972 and 1973.
Standard si~e glass n.icroscope slides (25 rim wide x 75 m long x ] irm
thick) were fastened to a supporting device which held them submerged
at a depth of 10 to 15 cm below water level. These slides were per-
mitted to incubate for periods of 14 to 3D days. Slides were visually
observed daily at tlie sampling location and were collected when growth
appeared to be at a maximum before material began sloughing fro™ the
artificial substrate J-\ At the time of collection, replicate slides
from each station were taken for species identification, gravinetr.ic
hionass determinations and nigrc.it extraction.
Pcriphyton diaton richness and snecies diversity were higher in all
four ponds during the 1973 collection period than during 1972. Rich-
ness and species diversity of "paired" ponds (1 versus 2 and 5 versus
4) were colored hoth for the shallow M Teet - 1.22 raters) and deep
(9 feet - 2.75 meters) portions. A paired t-test , 95 percent confi-
dence level, was used to ccnpare statistical significance. The re-
sults indicated no significant differences in richness or species
diversity occurred during the study period between pond - pairs.
Ash-free di*}' weight (organic) hionass (g/n-) and production (g/;n-/day),
as weil as chlorophyll a (i:ig/n-) a"d chlorophyll a production f.ng/m^/
clay) of the pcriphyton community were higher in all four ponds during
the 1973 test period than, during the 197? study period. On TOSt of
the collection dates, organic production wis less in the stress ponds
(1 and 3) than in the control rends. Organic production in I'cnd 1
(non-leaded fuel stress action) was less than Pond 2 (control section)
on fifi.C) percent of the collections. Pond 3 (leaded fuel stress sec-
tion) yielded organic production values less than Pond 4 (control sec-
tion) 66.(i percent of the time. V.hen taken as a group., the stress
ponds fl and 3) had organic productions tess than controls (Ponds 2
and 4) on two of every three collections (M.6 percent). As wi-th-—.
mean organic production, TCan chlorophyll a production showed a simi-
lar pattern. Chlorophyll a production in I'cnd 1 was less than Pond 2
on SS.9 percent of the collection dates while 77.8 percent of the time
chlorophyll a_ product ion in Pond 3 was less than in Pond 4. Regard-
less of these observations, no statistical significances, using a
43
-------
paireel t-tcst, 95 percent confidence limit, was found between "paired"
ponds (1 versus 2 and 3 versus 4).
Mean autotrophic index values, ash-free weight fg/n?) calculated for
chlorophyll fg/i:^)
the 1972 and lf>7j? period indicate that the test stress ponds (1 and 3}
were more heterotrophic than the control ponds (2 and 4). J-ven though
mi-an autotrophic index values \.)
t!un in the control ponds (2 and 4) in 7f percent of th^ collections
during 1972 and 1973 these differences were statistically significant
in only one case (Pond 1 greater tlian Pond 2). >.'o statistical difference
in autophic index levels between Ponds 1 and ? were observed during
the study period.
Phytoplnnktcn ronmun.ity - Species abundance and occurrence were studied
in 1971, 1972 and 1973; sarnies were calculated evict weekly. Dominant
species and bioniass were exanincd in 1971 while dominant species and
groups, species richness and population similarity w?re examined during
the 1972 and 1973 test periods. fairing 1972, short tern population
variations were studied over a five day period. Throughout the study
prinary production was neasured hy l^carbon fixation. Puring 1973
chlorophyll was rrensured and related to 1 "'carbon production as a pro-
ductivity index. A special cxnerinenr ir, 1975 was undertaken to study
the immediate effects of outboard engines on nhvtoplankton in the vicin-
ity of the test - engine docks frefer to Fieiire 1). In this experiment
phytoplnnktcn cell counts, chlorophyll and ^''enrhon production were de-
temined before, during and after engine operation.
Shifts in phytop'lankton species richness (S/ \/~X and S-1/ln.V, where S =
total nunber of species and X = total mmher of individuals were quite
sinilar in pond pairs (1 and 2, 3 and 4). Observed differences during
the test period were not statistically significant b->sed on a two-tailed
paired t'-tvst (95 percent confidence level). Ccnparison of mean numbers
of p.hytoplankton individuals of Pond 1 versus Pond 2 and Pond 3 versus
I'ond 4 for 1972 and 1973, together with sinilarity coefficient analysis
indicates that the initial supposition of sinilar populations in pond
"pairs" was valid. I'owever, during 1973 difference in species richness
between Ponds 3 and 4 approached signi f icance. These "near" differences
were likely the results of a hlocn of blue-preen algae and shifts in the
association of green algae which occurred on. opposing dates. Since both
pends (3 and 4) exhibited these species pulse-phenorpena (bloons), al-
though on an asynchronous tine scale, these variations inust be called
r-ii. iral and not the result of outboard engine stressing. In sunnary
there were not significant variations in phytonlank ton species rich-
ness caused by outboard engine operation in th.e lour test ponds under ¦
study.
-------
Four comparisons of phvtoplankton population sinila.it;. between ponds
indicated a high similarity of species associations lx:tvecn all poncis.
Pond 1 fnon-leaded stress) being similar *". Pond 2 (control and fV>nd
3 (leaded stress) being similar to its control (Pond J). Further
conparisons also showed the stress ponds (1 and 3) were similar as
were the control joncis (2 and 4). The extent to which these compar-
isons show population similarity indicates that the stress inn by two-
cycle outboard engines did not affect the composition of species as-
sociation in the ponds.
Analysis of the phytopiankton populations showed no significant effects
due to the outboard engine operation on the abundance a-ul occurrence of
phvtoplankton species associations. The population similarity coeffi-
cients and species rictuiess indices reflect similar populations in all
ponds from the onset of the study.
Phvtoplankton productivity (^ca^hon rcth.od) showed Pond 1 (non-leaded
stress) to be less productive tJi:i:i Pond 2 (control) during to
197.3. The leaded stress pond (3) 1 ibewise had less production \hu»n
its control (Pond 4) during the three year collection period. A tv.o-
tailed paired t-tcst (95 percent confidence) showed there c!iffercr.cos
not to be significant except in 1972, when the productivity in Pond 1
was significantly less than Pond 2.
Phvtoplankton chlorophyl 1 a measurements taken during the 1973 period
of study showed no significant differences between Tend 1 (ron-leaded
stress) and its: control (Pond 2). Tt.?se measurements did show a sta-
tistically significant difference however between Pond 3 (leaded stress)
and Pond 4 (control). Visual obsei-vations of the ponds suggest noriodic
bloom conditions in Pond 4 during 197." which would explain t~ie inci-case
in chlorophyll a levels in Pond 4. 'ITic variation in chlorophyll a be-
tween Ponds .3 and 1 is most probably due to the natural cyclic patterns
of biological communities in isolated bodies of water.
Phvtoplanktcn productivity index values (calculated ri'om ^carbon p ro -
ductivity and chloroohyU a values derived from independent subsarmies
of the 1973 composite collections) shewed nc signi fica.H: difference
when Pond 1 was compared with Pond 2 and 5 was coma red with Pond 4.
The resulting phytoplankton productivity index data of 1973 showed
no inhibitory or toxic properties caused by outboard engine stress-
ing over an extended period of time.
In midr.uinner of 1973 preliminary cxrx-rirents were undertaken m =ruoy
whether there was photosynthetic inhibition in the inrvdiat;.'»-\:icL..ity
of operating outboard engines. This study .included cell-counts,
-------
chlorophyll a measurements , and ^carbon production before, during,
and after short-ten:) fone hour) engine operation. liven though phy-
toplankton organisms were submitted to a momentary thermal shock
( T of about 10OC) preliminary results indicate no significant ef-
fects on primary production in the imedlate vicinity of the engines.
Icopl.-uikton Comunity - loonlankton dimples were collected on a rou-
tine basij during the 1972 and IP7IS stress periods hy use of a
"Wisconsin" style plankton ret.
'Hie tcii.i number of net-zooplankton collected during 1973 was greater
i:: nil ponds when compared to the 1072 .-rudy period. However, tho
same seasonal trends and natural fluctuations were obscned in both
years. The two "paired" ponds (1 and 2 and 3 and 4) were approximately
one month out of phase when compared on a "total individuals" rcco.vcr-
ed basis. In general, the species richness of the rooplankton com-
munity was lower in 1973 tlian ir. 1972. This effect carried through
out all ponds and was concomitant with population trends.
Application of the paired "t" statistic.il test to the 1972 and 197}
zooplankten data showed no surni ficant differences in nooplanktcm
species richness (S/ N and >-l/lp.\") when Pond 1 (ncn-loaded stress)
was compared with Pond 2 (control) or when Pond 5 (leaded stress) was
compared with its associated control (Pond 4).
In 1975 rrcnlankton species diversity lietween Pcr.ds 1 and 2 again showed
no significant difference. 1'ovcver, in 1973 a statistical difference in
"ooplankton species diversity was observed between ib:id 3 (leaded stress
section) and Pond -1 (control section). This species diversity signifi-
cance occurred because Pond 5 had higher total numbers of net ;oaplnr.V.-
ton than Pond 4 (7.2 percent), but more importantly, these greater num-
bers within Pond 4 collections were confined primarily to the Rrachionid
Rotifers (-1.4 percent). This seems to reduce the calculated diversity
of Ibnd 3. Rirther, according to Ruttner variations of ln to 20
percent have no significance in plankton statistics.
Prom nct-ioeplankton data collected during the stud.v period it is con-
cluded that the rrooplaiklon community in all ponds demonstrated normal
ecological dynamics throughout the stud.- period. ,\c effeC't on the
;oopl ajikton comunity can be attributed to two-cycle outboard motor
operations in the .northern test lake systems.
Ponthos fiy—--nity - r\i;uit i rat ivo henthic surveys of all ponds were
perfon-ed monthly during rhe entire stiu.y ivriod OnTl to 19~3).
These siiiTCv; entailed the collection of throe bcnthic samples from
each of the- four j^ends once a month. Statistical sarpling points
46
-------
were achieved by usini; an irru:innry alphnnumeric £rid system over
layinp each pond. Sampling points were then chosen by rnndon selec-
tion during the first study period (1971) ami by directed selection
the suhscoisent years (1072 and 1975). Ail four ponds had conparativcly
kird substrates, therefore the Ponar fmh s:unpler was chosen for
benthic collection.
11k? 1971 test data indicates a rather diverse (2.SO, 2.0-1, 1.95 and
2.02 species diversity valises for Ponds 1,2,3, and 4 respectively)
benthic faunal asserihlauc. The subsequent years (1972 an J 1973)
yielded a r.uch lower species diversity of benthic nacrofauna in all
four ponds (1.15, 1.24, 1.07 and ].20 for ponds 1,2, 5 and 4 re-
spectively in 1972 and 0,"3, 0.61, 0.f>l and 0.58 for ponds 1, 2, 5,
and 4 respectively in 1975). This shift in benthic populations
occurred in >by 1972 in all ponds.
This shift was disc to norral trcrhic succession foia-d in all txxlics
of water arid accelerated lie re because of the siue of the pond cells
and the fact that the trophic nature of the ixvtds shifted (April
through June of 1972) fron Chara sn. to the free livin:'. rh\toplank-
ten species. This :;hift in tvcT^.ic nature reduced the ricrotiahit.it
availability as well as changed the t;>-pe of habitat. Thus, or^anisns
such as fvlonata, Inhere root era and Coleontera which lived within the
Cham ccnraunitv £a\"e way to substrate dwell inc organisms such as:
the Anne] id worn and Piptcrans. These organisns could survi\"e by
burrowing into the unprotected ruds and feed upon the decaving Chara
;ind other detritus. Once this shift cccurred in early 19~2 the
l>enthic racrofauna were sinilar in all four ponds during the re-
minder of the study period. Paired "t" statistical testing (95
percent confidence linit) of the benthic data 1971 to 1975, in-
dicates no sici; ficruit difference between Ponds 1 and 2 and between
i'ends 5 and 4. Our hoard engine stressing over a th.ree year pcrioc!
did not effect the benthic ccrnimity.
'¦'ish Cominity - In the northern field fish stud}- fathead nirmow
i?ir,erhales r-rorclas) , Pduegills (I.crop.is nacrochi n:'j , unJ gold
i ish fwavassii'iS nuratns) were stocked in the ponds (Test as well
as control sections) prior to stressing. In the first year (1971)
l he fish were allowed to roan in all ponds. Inuring the second, and
third years of the study sc~e of those fish were placed in live
boxes for subsequent flesh testing studies. The fish placed in
these live boxes ( .'5 fish r«?r hox^ verc allowed to oat food native
ro their respective pone! habitats rather than artificial foods.
."ish. flesh taste studies were rerfmred when the gallons of fuel
binned per mi11ion gallons of water levels were 1.4, 1.5, 2iJ% 4.0,
4.2, 11.2, 7<-..9, and .1.10,5 These studies were per foiled on fish
-------
fron l*oth the non-loaded arid !o:uW fuel stress section?, Ponds J
and 3, resivct ive ! v. i-'ish handled in the rannrr fror 1'onds 2
and 4' (control ponds') vere used as control fish. <>t the average over
the test pori«\l no '"off-flavor" k>s noticed on (>2.? percent of tIjC-
fish sa;-nl es in the r.on• leaded test r-"*nd and 74.3 percent nf the fish
fro:'! the leaded test ;v>r.d. .' mid lv nct?d that total dissolved
sol ids wctc rni .'-xvistiivd directly, hut rather, tie ::viicrir" of anions
and cations uluch ;;ial e up the dissolved sol ids vere measured on an
individual basis.
'Ver the three year period nhosphorus level;: did not vary significant 1
h-otveen control ;i;h! stress '.wds. It should K- noted here that the
total phosphorus levels for .ill iv^ds increased s ip.n.i f icai-.t !y during
the 1P~3 test season. This increase is rainiv in the jvMyphosplKjms
or part iculai e fraction. 'P-is rise in total phpsnhorus is post prob-
ably due to tho natural process n'~ r.e.t rie:U r.wclinp v/i t 1 - in each of
the nonds.
A p.eneral review of tW- nitnven data, eatherod during t he study, for
each separate, r-ond indicates an overall increase in organic and nit'..it
nitropen ub.en cerparins i'.'Tl levels v.it'-i ilewis. Armenia ami
nitrite ritropen data show no sum "trend" increase over the stude
¦vrinii. Apain, this cheHca! shift ire nn'var? to he a natural process
cf the ponds themselves ami not related to nut board enpine stressir.;,!.
>';atist:cal analysis ("paired "t" testinp) of all the chonical data col
'.ected I'lirii." the study showed that there v.a? significant differences
confidence leveH in five chonical co:">cr.cr.TS <'n!', hardress. sul-
fate, conductivitv and tend"* vhen Per-.' 1 fnou-leaded, stress scctin??1
enrpared ro its associated control Tend 2"i. Sip.ni ficar.ce at the
:1.V confidence level va- observed in nine ."he"1::.'!] cornonts in the
i titer ce!'r::i nihil initv , hardness, caicii", ::iapne--itr", chloride,
conduct ivi ty, dissolved i •.wicer.. tcrncviturc and leadl '-hen the leaded
ins; section Tend 31 i-'.'.s comarev n< i>nd J fconrroP.
-------
The differences reported for hardness, calciun, chloride, sulfate,
anil conductivity, although statistically significant wore not that
large when average pcan values over the three year test period arc
ccmtroJ. Variations observed in these chenical components are nost
prokiblv duo to no:v.nl variations in sonrl inc. and accuracy of chcnical
analysis. Those variations con!;! also ho ihe result of scouring of
the iiotton sedincuts in the stressed ponds by the test propeller wash.
This latter explanation is a distinct ressi1 iliry as the nean values
cf the five a!*ovc-::icnrioiic\l chenical conponcnts were higher in the
test sect ion water coltn:uis (Ponds 1 and 3) than ir. the associated
ccnt'-M section water colir-j-is (l*onds 2 and. ;1). This perturbation of
the hot ton sedincnts is more a result of physical siring (average
water depth of 0 feet) of the test facility ami would K atvpical
of normal outboard notor operations in deeper water. Thus, it is con-
cluded that the significant differences rejvtrted above cannot ho at-
tributable to two-cycle outboard exhaust emissions per so hut rather
to the physical operation of these engines in shallow water (¦£. 6 fcctl.
Significant differences in dissolved oxvgon was also observed during
1P72 when the non-leailed fuel stress section (Pond 1") was compared
to its associated control fl'rnd .?). the other Inrd, lYrnd 1 fcon-
trol) showed higher dissolved oxygen levels than its associated
stress test section flY>nd 31. A review of the phytcplankton results
for these ponds during 1!'72 indicates larger phytonlanVton populations
in Ponds 1 and J than in Pn:\i< 2 ami 3. This leads one to conclude
that the significant differences observed in dissolved oxygen in 1972
V'Cti-.een Ponds 1 aad 2 and I'or.ds 3 and -1 is probably due to natural bio-
logical activity nunc-" than an effect due to outboard engine stressing.
A temperature significance was noted between rends 3 and J. The rcan
difference in tenncraturc IaTI -..as .S°C over the three year period.
!t is lei t that this significance- can be explained by the netiioils used
to noasure terTcraturc in the field. No effect on. water terrreraturc
can, he attributed to the stress si;g hy outboard engines except in the
i'T'ediate vicinity oi the engines. Occasional statistical significance
'.as .also observed in n" and alkalinity when stress sections wore con-
rared to control sections. iTcse shifts in pi' and. alValinilies can be
j'olatod to aigal activities i-hich leads to preferent-i-al cor. sunn t ion of
carbon.arc.-s by the hiojcgical c^niuii t ies with a resulting decrease in
carbonate ali.alip.ity :ri
i.ead - l.c-ad concentrations in the water colunn were significantly higher
:n !>oth stressed ponds fl and 3) than in their rc-snoct ive control ponds
"1' and -i :. The sign, i f icant: differences in lead levels between Ponds 1
¦:i-d r can. possibly be cxnlnined bv sedircnt liotton scouring due to
:n.c propel ier wasii in the test section (Pond I). The hot ton sedincnts
-------
of the tost [Kinds (1 and 3) contain between 1" :mu lf> r^; of lead per
lhe resuspension and subsequent resolution of this sodinent r.atcrinl
would tend to raise the total lead content observed in the writer colunn
oT the tost section. \s susrcctod lead lewis in the writer colurm of
I'oiuf 3 ("leaded fuel stress sevtiorj t-ero higher than in any other
I'ccnuse of this possible hot ton sedir.ent scouring situation iv v.tis found
desirable, fren a statistical point of view to compare the leaded levels
in the writer colian of Pord 3 with the lead levels in Pond 1. i'oth of
these por.ds were n f fee toil by possible scour im.1. due to outhoru'd. engine
operations. Itowcvcr, only Pond 3 was ex'xiscd to stressing with leaded
fuel. I'enc.o, Ponds 1 and 3 (non-Unded an'.l leaded stress sections,
respectively) vere evaluated usin;^ :i rise-tailed mired "t" analysis.
A si pi i f icrirt difference in lend levels i-:i s oI'Scitci1 between Por.ds 1
and 3. Ihere is a 3-! percent increase in lerul concentration in Pond 3
as coi:e in lead levels is directly at-
tributable t o tlie use of outboard engines operating on leaded fuel.
'Xirhv.; the three seasons of stressing, a total of 1 , !S Kc of lerul was
inticduced to Pond 3 tbroncb the corhustion of fuel containing 3.1 cirnr.s
of lead per r.nllon. Anal'.'sis of lead content in the po.ml sedircnts and
statistical en 1 mt i on of the data does not sIioh a significant increase
in lead in the scdirents of "Pond 3 iloaded fuel stress sectionl. Jhe
roan lead concentration in. the better -edinonts of Ponds 1, 3, and -1
were 23.r, 17.9, 29.5. and"2n.l :"u/respectively. Faired s;at ist ical
coirrpar i sons between i'^nds 1 and 3, does r-'M indicate a sipr.i ficnnt dif-
ference in lead contend. of the scdirents between "he two jvnds. It
should he noted here that the snail srir-de sire rrrrd th.e hich stnndniJ
error of difference due to the standard deviation of the differences in
the snriple population nay be the reason r.o significant differences were
observed in 11:e lead content of the sedirerts.
Hydrocarbon - !V'ita fron the 1973 -tressir.v season on saturntred hvdrocar-
hofi levels in the .¦ater c.olum is presented in TaKle !v. These data in-
die.ite saturated hydrocarbons in the ho'linu wir.t nuv.'.c fccv-
respondinr to, in rcleciilnr weight 'l'1 to l-]:> n-tvirr:ffins present i'1 tdx.4
rends") in the ran<:e of n.3." to nr.vj 1 . Saturated Iwdrocnrhor.s in,the
3fin-.;r'P0C boiling point rniire fco'Tesi^n.Iinr in roJocular i-.ei>:ht to '-1?
to (-2-l n-paraffins) are present at the P.II tu fi.21 it^/1 level. It should
he noted that hot h erulsi! :ec! "7\! rulsorbcv hvdrocarhon na*oria 1 is include^
-n liie anrilysis of these writer co!ir"n srvrles hv the ana !\'t ica 1 procedure
ei-ployct). Stati stica1 ;innl i s of this J.nt;i by ri ti-.o-tai k\i p;ji>*ed "r"
tost shoii's that no s ipn i t : cant ci i f fer'en.es exist between stressed rind con-
trol ponds durint". tb.e period of t.estinr.
-------
tap.j.i; ix
UYDKOCARBONS* (Boilinp. Point r.anRC 175-300°C) (nR/l)
Date
Pond 1
I'o nd 2
I'ond 3
Pond 4
15 May 73
0.45
0.40
0.41
0.38
1 Jun 73
0.39
0.40
0.43
0.39
26 Jun 73
0.48
0.42
0.36
0.35
2 Jul 73
0.35
0.35
0.35
0.38
23 JlI 73
0. 37
0. 37
0.37
0.36
HYDROCARBON'S* (iioIlinR Totnt K-ingo 300-400°C) (r^./l)
Date Pond i ?'ond 2 Pond 3 Fond 4
15 ray 73 0.13 0.11 0.12 0.11
1 Jun 73 0.11 0.12 0.13 0.12
26 Jon 73 0.15 0.14 0.17 0.19
2 Jul 73 0.1S 0.16 0.19 0.20
23 Jul 73 0.21 0.16 0.16 0.17
*S.-ii.-plei; arc co.w>osites of lndlvidu.il samples taken at surface and middle at
stations A and li and surface, middle and depth at station C.
"1
-------
73
73
73
73
73
73
73
73
73
73
73
73
73
73
73
73
73
73
Tabic X
Saturated Hydrocarbons in Sodinents (nig/kg)
(boil in? point range 175-300°C)
Station Pond 1 Pond 2 Pond 3 Pond 4
A
•0.1
3.2
12.8
11.1
B
6.0
15.3
3.8
<0.1
C
3.A
4.7
<0.1
2.8
A
3.4
3.8
10.5
9.1
R
0.8
2.7
4.0
3.6
C
5.4
5.0
25.8
6.7
A
7.5
2.1
4.6
1.8
B
5.9
8.7
<0.1
4.3
C
4.7
12.9
19.5
0.3
Saturated Hydrocarbons in Sedicientt: (mg/kg)
(boiling poir.t ranpe 300-400°C)
A
4.7
2.9
19.2
7.9
B
1 1 .0
6.4
4.5
3.7
r
v>
5.1
4.0
6. 7
3.7
A
9.9
1.7
4.6
7.3
3
3.1
3.4
6.5
4.2
C
L6.0
2.3
22.5
6.1
A
22. 3
1.1
1.6
7.8
B
12.3
15.9
0.2
4.6
C
5.4
7.6
8.8
5.1
52
-------
Saturated hvdrocarhon levels of the soil incuts fron .ill four ronds arc
shown in Table X. I'ydrocarhon levels in tlic 175-50(1° hoi line point ranpe
varied from 11. 1 to 25.8 nc/Ke dry v.eipht of seel incut . The overall mean
concentration of hydrocarbons in the 175-300°C hoilinj: ]X)int ran^c in
the sediment? of Poi.d 1 is about tlic same as that of Pond 2 fJ.! jng/rg
versus (>.5 riti/K«, respect ivcly"). Ilic overall mean concentration of hydro-
carbons of this ran<:o (175-500pC) in the sediments of Ponds 5 and -J
were P.n and S.O nig/Kii, respectively. in the 500-400°C hoi line point
ran^e, tlic saturated hydrocarbon levels in the sediments were 10.;5,
5.0, S.5, and 5.<> mp/Ki; for Ponds 1, 2, 5, and 4, respectively. Sta-
tistical analysis of this data indicates nD significant difference be-
tween Ton-Is 1 and 2 and 1'onds 5 and 4. Visual observations of the
data seens to indicate a '"trend" with an apparent increase of saturated
hydrocarbons in the .~0fi--!0(l°c boiling point range when Pond 1 is com-
pared to Pond 2 and Pond 3 is conpared to Pond 4. !'ov/evcr, this is
negated statistically by the small sa-rle size and the hio.b standard
error of difference due to the standard deviation of the differences in
the sample population.
sirmr.Rx nr.u) sinpirs
Treatment Levels
Motor operations of Koth rest systems was on a continuous basis for an
IS-month peri oil. ^ii a emulative basis both the drained and drainless
ermine test por.ds were closed to a treatnent of approximately 40n liters
of fuel burned per million 1itors of water over the first 12 nonths and
"00 liters of fuel burned per million liters of water over the total
18-month period. It should he roted that the on ti rial treatment level of
three times the maximum boatinj'. usage was not attained until the
twelfth ~ontb of the IS month, study as the water level receded. Treat-
ment levels continued at values of three tines maximum boating usage or
great or for the reminder of the project.
Methodol ogv
All chemical and biological sampling was perromcd by routine standnrcJ
procedures. A detailed description of sampling procedures together with
the chemical and biological analytical techniques employed is presented
in the final rcjiort to the irivi ronncntal Protection Agency.
.'¦.esi'l ts
1; i ological
i'ciiitiic Varroinvertebrates - Fenthic samples were collected '-ith a 5(»
so.uare inch ("0.1)2 square meter") l.lman dredge from August, 1971 to >*arch,
35
-------
IP75. Hio control and drained cupino test ix>nds were sampled at five
locations each month, the drainless engine test pond was sampled at four
location.-; per month throughout the study period.
A complete list of the total macroinvertebrate taxa collected throughout
the study is included in the final report. In general, the hcnthic com-
mnity structure illustrated taxanomic similarity throughout all three
rest nonds; each he inn dominated hy insect larvae prinar.il>' within the
Order Piptera. Oth.cr organisms collected during, the study period
were: Tuhificids, Anphipods, ("-donates, rphemeronterans, Nematodes,
Trichopterans, Turbel larians, !!iriid ineans , Colecpterans, Cladocerans
and I'ydracarina.
Hie number of families of macro invertebrates was lowest in the drain-
loss engine test pond. Ibwever, calculated indices on community di-
versity indicate the drained engine test jxjnd to be the lowest in di-
versity with the control. pond the highest.
A one way analysis of variance (AVOYA), '..hen applied to species corv
rcsition and density, indicates no discernable difference in the
mcroinvertcbrate populations of the throe ponds after 18 months
stressing. The ANOYA analysis was performed at the P5 percent level
of confidence.
A multiple regression anal.yais was also performed on hcnthic inverte-
brate diversity to determine seasonality or differences between ponds
due to engine stressing. There was a significant difference between
rornis with respect to their mean diversity indices. The control pond
had the highest species diversity index, the drained engine pond next,
and the drainless engine pond the lowest index. 'Ilic control pone! was
significantly higher than the drainless engine pond, hut was not sig-
nificantly higher than tlie drained engine rend. The drained engine
pond was net significantly higher than the draijiless engine r>ond. For
a riven pond there was no statistical difference hctwonn the mean specie
diversity indices over the full 19~2 Txmua1. evele, although the spring
season lad the highest indices and the winter season had the lowest
i r.d i ces.
Periphy'ron ; Pcrinhyton productivity as dry weight and ashfrce dry
weight !'g/r.-/day) was measured throughout the study (August 19, lf>71
- ."¦"arch 27, 1?75). ('.lass nicroscope slides provided artificial sub-
strates for rerinhyton colonisation. Artificial substrates such as
glass slides are somewhat selective but this-provides a technique
fcr sampling similar populations between kmkIs . Natural substrates,
i.e., macronhytos, were more abundant in the drained engine and drain-
less engine ponds than in the control pond. Vcasurenent.s of the
-------
periphyton cornunity were based on the assursption that natural sub-
strate surface ;ircn , gracing, predation, death. ami decoriposit ion
were on the s;irie order of magnitude in all three study ponds. There-
fore, the calculated organic productivities reflected actual dif-
ferences related to the natural and experimcntal stresses in each
pond. Periphyton daily growth rates, expressed as grans of carbon
fixed per square merer per :Vw, ranged fron 0 .U1 to 0.12 during tl-.c
course of the study. The highest sin»le daily growth rate occur--ec' in
September in the dniined engine ponds (0.12 g/n-). The genera], peri-
phyton growth Kites were highest during the sitrner :no:iths and lowest
during the winter.
A one wav analysis of variance test was used ?.o determine significance
of variations in the data, ''.his test connarcd productivity values from
the couthinations of control and drained ponds, control and drainless
ponds, drained and drainless nonds, and control, drained, and di-ainlcss
ponds. Although the drained en;;inc pond showed higher daily growth rates
"ijid a higher yearly average, th.e variations in periphyton productivity
between th.e three study ponds was not significant at the percent level
of confidence, lience, assuming substrate surface area was equal in all
:hr:e ponds, the operation of two-cycle outboard engines (drained and
drainless; bad no apparent, effect on the periphyion comiuiity o^'cr the
course of th.e study.
Thy top Ian!-: ton - The nhytoplanktor. connmity was studied throughout the
course of this project using species diversity, product i J t>* and hiomass
analyses. identification and counts of con^os'ite surface and hotton
sarples were rade rtonth!y. Co;--->arisop.s between ponds wove made with the
Ph«*union-Weaver species diversity index where enviro.nrontal stresses are
reflected by a decrease in diversity. Morass was estimated through
chlorophyll a_ ncasuregents which were taken on con~>ositc surface and
-•ottom sarrdes by in vivo P.unrcnct.ry once each nont.h m 107.". Phvto-
rlankton prirxsry product i vi ty was reasured us inf. the carbon teclui itjue
during 1971 and an oxygen net hod during 1972 and lf-7."S.
Hank ton cell dive-.-sit.y, recorded as cells/nl, indicated that the
Cyanophyccae fhlue-green algae! vorc r.ost abundant in all three ponds.
Anacyst is r.-srina was described as the d-ninar.t organism as it occurred
in -!6 percent o' a',1 samples fren the drainless cr.nine ponds, ?>?> percent
c: all sanrles in the drained engine por.ds , and .i.> percent of all samples
i':: the control nond. The Chlorophyceao fgreen algae) was the most diverse
croup but also occurred with the lowest abundance. "Moons (' snn colls/rl
cf green algap did not occur. The Chrysrophyceae (golden-brown algae),
vc olten rare in abundance (rare = 1-? cells,;rl and uncoruron = 6 -! g-
cells/nil. Likewise r ho JiinofiagelJates, also chaTncter ist ic of r.ost
i-lorida lares, occurred in the "rare-unco.'iiron" range r.{ abundance. l"ho
-------
R'ici ilariophyceae fdiatoms) were exceedingly rare. Several other rcaior
group? were represented by a few taxa, but were "rare-uncomon" in
abundance.
Seasonal species diversity indices were variable, although differences
in moan diversity values (2S June 1971 - 2 April 1973) were not tliat
great. A one way analysis of variance test (AXOV'A) indicated no signif-
icant difference between diversity in the three test ponds during the
1971 baseline study. Furthermore, no significant, variations in diversity
were found during 1972. itov.evcr, :Wn\'A calculations did show significant
variations (99 percent confidence level) in species diversity during 1975
with the drainless engine pond exhibiting the lowest diversity and the
control pond yielding the highest diversity. Piversity values seem to he
closelv related to the blue-preen ajgae, Anncvstis marina, which often
occurred in "hlcon" conditions. There are, however, no studies avail-
able which relate this species to environmental stress factors.
Phytoplanfcton cell density was quite variable although mean and standard
deviation values arc reported as being siriilar to earlier studies in acid-
soft water lakes. Cell diversity was lowest during simmer months and it
is felt that phosphate was a limiting factor at < 0.001 ng/1 orthosphos-
phate. The one way analysis of variance test showed no signi f ica.it dif-
ference between cell density of the three test ponds during the 1°71 base-
line study or during the 1972 test period. The AN'OVA test showed signif-
icant variations between cell density of the test ponds during 1973 at
the 99 percent confidence level. Greater ninhers of cells/nl were found
in the stressed ponds compared to the control.
Hstinates of standing crop by pignent analysis indicated higher values in
the control j\3nd as opposed to the drained engine and drainless engine
ponds. I'ionass values were quite variable between the test ponds during
the 1971-IP?? study period. The one way analysis of variance test, r.t
the 95 percent confidence level, indicated significantly higher chlorophyll
a levels in the control pond than in the t^o stressed ponds. In applying
Fhe linear regression test to the control :iend data, there was a high
correlation between cell density and chlorophyll a. There was r.o such
functional relationship found in cither test pond. Tt should be noted how-
ever that the correlation between cell nur.hers and chlorophyll a ;iay he
influenced by cell size. As for exrailc, .\racystis marina, whicF was
found in higher abundance in the stress ponds, has an extremely s:~ill cell
si:.e and most probably a small arount of c'aloronbyl 1 a_.
The 1971 baseline data showed greater nhyteplankton productivity in the
conti-o! pond than in the treated ronds. \1 though the control por.d did
not: vary significantly from the drainless engine tx)nd, the variation be-
tween the control and drained engine pond was significant at the 95 per-
cent confidence level fAXOVA test). Significant differences between
5ft
-------
the control and the two stress ponds (PS percent confidence level)
wore seen in nhytoplankton productivity during 1972-1P73. In study-
ing the annual trends of the three test ponds, it is noted that pro-
ductivity in the control pond varied greatly fron anproxirately ft.01
gCr.~-'> to a high of 1.1 r.Qn" \ 'Hie stress ronds ringed approximately
fron O.ni gOT-^ day-l to 0.7 p.C"i:i~3 day'l ant! essentially paralleled
each other during most of the study period.
In conclusion, there wore temporal changes in species coriposition,
standing crop, species diversity, chlorophyll a^, and primary nrcduc-
tivity. These changes were most prohahly due to variations in nutri-
ents including carbon dioxide, temperature, and light, and cannot lie
conclusively correlated to outboard engine emissions or operations.
Macrophytc associations - l-val nation of the microphyte cnminity include
species composition and standing crop analysis together with grass bed
productivity and carbonate enricbrent experinents. Crass bed collections
were nadc I1}' random quadrat sampling and rrnovinp all vegetative struc-
tures.
Hie eul.ittoral and sublit'tornl :one> of the drained engine pond lrid
greater plant biomass than the other test ro:xis. Sampling frequency
was not sufficient to examine the data statistically.
The vegetat. icnal changes in all three test ponds have followed n nonral
qualitative pattern as is evidenced by species composition, frequency
and distribution. The control pond was dorinated by Hlnddcnvort
fUtricularia floridans) aird ("arret 'lush Muncus rcpens), while the drain-
less engine pond was dominated 'iy a population of Bladderwort throughout
the study period. The drained engine test pond initially sirpportcd a
nixed cornunity consisting primarily of P.ladderwort and Kater ,:int.
f!;ydrotrida carol iniana) , however, during the summer of 1972 V.'ater Mint
began dominating all TuTt the deepest areas of the pond. In May 1975
P-ladderwort was reestablishing itself within tiic drained engine test
system.
Grass bed productivity was measured in clear plastic, domes diameter
x 0.5m high,) which war-, placed over selected areas of the plant bed. Oxygen
was measured at dawn, dusk, and at the following dawn. Productivity was
calculated in terns of t!difference in oxygen levels between daylight
and dark periods. \'o statistical difference fAX'OYA at 05 percent confi-
dence) in gross productivity of the grass bed was observed between the con-
trol and the drainle:-s engine pond. There was significant difference in
productivity between the control and drained engine ponds and between the
drained engine and drainless engine ponds.
57
-------
|>.iring five of the seven months for which grass bed productivity data were
obtained, specifically, January, April, .June, July, and September, the
drained engine pond had the highest productivity, the next highest was the
drainless engine pond except during the month of June when it was the
lowest. The control pond had the lowest productivity except for the
month cf June when it was higher than the diTiinless engine pond. Aver-
aged over all the months, the highest productivity was in the drained
engine pond, next was the drainless eng'r.j nond, and the control pond
luid the lowest productivity.
F.nrichnent experiments were conducted in the control ami drained engine
ponds to assay the grass bed conrniuiity for carbon as a limiting factor.
As n!: was low (less than the carbonate-hiocarbcnate equivalence point)
and there was the probability of a non-carbonate buffer system, carbon
was suspected to be a limiting factor in the grass beds. A control
and five levels of carbonate solution (fl.5 - If pnm) were added to in-
dividual dores by capillary tubes. Oxygen production was measured be-
fore ;md during the addition of carbonates and coptparcd to the control.
No significant differences in productivity were noted by adding carbonates
to the control pond. Most likely because of the sparse vegetation that
occurred during November when the experiment was conducted. I'owever, car-
bonate enrichment did significantly increase productivity in the drained
engine pond. Mien compared, the drained engine nond In d a much greater
initial hionass than tin? control pond.
There arc indications from these studies that carbon dioxide (CO7), a
component of the gaseous exhaust emissions, may cause, at the treatment
levels employed in this study, an increase in productivity in soft-acid
watci lakes. This "treatment effect" has not been fu]ly substantiated
because only one of the tost systems (drained engine pond) showed this
response. In addition as a result of mixing and stirring, transport of
carbon dioxide as well as other nutrients into the grass hods may cause
an increase in productivity. I'atcr circulation was observed in the ponds
during tho operation of the motors. Some scouring of the bottom sediments
w;is noted in the vicinity of the motors.
Fish tainting - Samples of sun fish (I.enoni s sn.) and large mor.th bass
icrontcrus sa 1 noides) were collected periodically from each 7x2nd and
tested iiy a panel of eleven judges at the I'niversity of Flor'da Food
Sciences Laboratories. A triam.T.lar testing method was employed in
which the judges were given three sannlcs, two of which were identical.
They were asked to identify like sannlcs and to rate than 011 a scale
frcr one (very hati taste) to seven (very good taste) and to comment
as to whether any traces of "tainting" by petroleum products could le
detected.
The taste of bass and sun fi sh was iound to range from four (neither like
or dislike) to six (like moderately) in each of the three test ponds
throughout the entire study period. No significant evidence of fish taint-
in;: was observed during the study.
.>S
-------
Chemical -
Generate ehv-'MStry •- To evaluate th<\eflcct erf outboard marine engine
operation o.i the chemical syster\*tf the isouthem lakes, a relatively
broad chemical characterisation study, of the'aquatic system was under-
taken. A detailed* tfescriBtiQfr of the,chemical methods enployed is
given in the f»naj report.. Baseline data on the chemical systems in
the three poftds^is Collected durir>jv tho.flonths of Julv through f'ctolxM ,
1971.
The soutnern- "Lakes"' st#ilar to nam* freshwater l:lcrida
lakes'in .that they coS^ofiS^jh^.arxxaiits ,of dissolved solids, :»;c acidic,
and have clear w^ter'<•*«* .derfsc grass be.l zone, r.iological
activity and s torn Katffir"'i%iu!isr genera 11 y determines the nutrient level':,
and seasonal variation.ijgfljhis-.activity can c-ccount for relatively large
variations in the levels a^jroleciiltr forms of carbon, ni tr^gen and
phosphorous.'
Of twenty-five chenical^parajneters Pleasured, no effects were observed
in total dissolved solids, total suspended solids, hardness, conduc-
tivity, turbidity, pi!, dissolved oxygen, chloride, fluoride, sulfate,
magnesium and iron. Pxcessive vari:.tions in the natural background of
biochenical oxygen denand (1 to 5 rr/ll and rhenical oxygen doranci (9
to 8.s ng/1) precluded isolating the centrituitions nade by tiie outboard
marine engines. Similarly, there wore .seasonal variations in total
phosphorus, orthophosphcrus, amonia, nitrite, nitrate, pa', t iculate
organic carbon, and organic nitrogen luit no relationship with outboard
engine operation could tie found.
Inorganic carbon - Two chci.iical cornrr.'.ents showed a statist ically sig-
nific;int variation as a result of outboard engine stressing, "'can total
inorganic carbon concentrations ir. the stressed pond were S'1 to c'.tS ht-
cent higher than >n the control pond. Levels observed in ali the rands
fell v.ithin the 0.2 to 1.1 wg/1 ran^e which approach tlie l ir.it of de-
tection of the analytical method errvloyed. Hie increase in inorganic
carbon in the stressed ponds is attributed to the C.'Ot emissions of the
engines, the CCb resulting fron oxidation of unturned or partially
hurried hydrocarbon emissions, fixing of the iotton waters with the sur-
face waters in Hie linnetic :one and the increased diffusion of he-
tween the atnospheie and the lake waters due to engine operation.
Organic carbon - Tcjtal organic carbon levels '.-.ere significantly higher
m the drained online pond than in either the drainlcss engine or con-
trol ponds. A neaifrcpnccnt rati on of 7.1 r;g/l ir: the dm ined engine pond
and 28 percent highft- than that of the drainlcs* eneine ,'ond and 1" per-
cent higher than the control pond. This nay he attributed to the crank-
case drainage of dissolved organics '"hydrocnr'-ensl associated with the
use of drain-type engines as well as deconpnsition of vegetation.
-------
Motor crointion in the so'Uhem Likes increase tlic concentration of dis-
solved aromatic hydrocarbons :u*: also mixed and circulated the lake water.
Therefore, the hydrocarbon emissions, a? v.ell as, otlier exhaust emis-
sions v.vre distributed throughout the lakes. The concentration of
arorntic hviirocavhon fas nvasured by M-V analysis) increased fron
backprouy'i levels of less than (M nv./1 (measured as toluene) to levels
of 1.0 nc/l during rotor operation. Mien the rotors were not operated
for twj days, the hydrocarbon levels declined to less than 0.1 r*£/l.
Although the leaded fuel was used in this study, the environmental in-
nact of lead emissions were r.ot studied because of the large background
levels of lead («c to 90 ng Pb/Ke of dried plant tissue) observed in
tl:e too ted vegetation of the lakes. No knovn sources of leaded pol-
lut;mts v.ere found to have entered the three southern lakes prior to
the study.
oO
-------
SECTION VI
REFERENCES
1. Threinen, C., "Conflicts in the Boating Scene and Their
Bearing on Access Sites", Proceedings of the 5th National
Conference on Access to Recreational Waters - Boston,
Mass. (1967)
2. Private Communication, Dr. David Cole, Department of
Mechanical Engineering, University of Michigan, Ann Arbor,
Mi.
3. English, J.N., McDermott, G.N., and Henderson, C.,
"Pollution Effects of Outboard Motor Exhaust - Laboratory
Studies", Journal Water Pollution Control Fed., 35, 7,
°23 (1963). —
4. English, J.N., Surber, E.W. and McDermott, G.N. "Pollution
Effects of Outboard Motor Exhaust - Field Studies",
Journal Water Pollution Control Federation, 35, 9, 1121
(1963) . —
5. Blumer, M., "Oil contamination and the Living Resources of
the Sea", FAO Tech. Conf. on Marine Pollution, Rome, Italy.
Dec. 9-18, 1970.
6. A.S.T.M. (1968) Manual of Hydrocarbon Analysis. 2nd Ed.
Philadelphia, Pa.
7. Martin, J.M. Jr., e_t a_l, "Ultraviolet Determination of Tot-
al Phenol,"p. 21, Journal Water Pollution Control Federation,
30 (1), 1967.
8. Schenk, N.K. et al, "The Effect of Long-Term Exposure to
Outboard Engine Exhaust Emission on the Fathead Minnow,"
University of Michigan (ORA Project 34856), August 1970.
9. Boating Industry Association, Chicago, Illinois, personal
communication.
10. Project 15020 GPG. Review Meeting between EPA and BIA
Consultants, Gainesville, Florida. October 1971.
11. Cooke, W.B. Colonization of artificial bare areas by
microorganisms. Botanical Review. 22:613-638, 1956.
12. Vollenweider, R.A. A manual on the methods of measuring
primary production in aquatic environments, IBP Handbook
No. 12. Oxford and Edinburgh, Blackwell Scientific
Publication, 1969. 213 p.
61
-------
13 tfebe.r, Cornelius I. Biological Field and Laboratory Methods
for Measuring the Quality of Surface Waters and Effluents.
U.S. Environmental Protection Agency. Cincinnati, Ohio.
EPA-670/4-73-001. National Environmental Research Center,
Office of Research and Development. July 1973. Plankton
17 p. Periphyton 6 p.
14. Ruttner. F. Fundamentals of Limnology, Third edition.
Toronto, University of Toronto Press, 1963. p. 105-219.
15. King, D.L. The Role of Carbon in Eutrophication. J. Water
Pollution Control Federation. 42: 2035-2051, December 1970.
------- |