Early diagenesis of organic matter in the water column and sediments: Microbial degradation and resynthesis of lipids in Lake Haruna

Off. Geochem. Vol. I I, No. 4, pp. 251-264, 1987 Printed in Great Britain. All rights reserved

0146-6380/87 $3.00+ 0.00 Copyright © 1987 PergamonJournals Ltd

Early diagenesis of organic matter in the water column and sediments: Microbial degradation and resynthesis of Hpids in Lake Haruna KIMITAKA KAWAMURA*, RYOSHI ISHIWATARIand KAZUKO OGUKA Department of Chemistry, Faculty of Science, Tokyo Metropolitan University, Fukazawa 2-1-1, Setagaya-ku, Tokyo 158, Japan (Received 16 September 1986; accepted 9 February 1987)

AlWtract--Particulate matter, sediment trap, and surface sediment samples collected in freshwater Lake Haruna were studied to understand early diagenesis of organic materials in the water column and in bottom sediments. The samples were analyzed for biomarkers, including aliphatic and aromatic hydrocarbons, fatty alcohols, saturated and unsaturated fatty acids, ~- and co-hydroxyacids, and a,co-dicarboxylic acids. Decreases in concentrations of autochthonous saturated C,2--C~, fatty acids and polyunsaturated C,s acids relative to TOC occurred with the settling of organic matter to the lake bottom, whereas the amounts of terrestrial saturated C20-C30acids remained almost constant. Conversely, the concentrations of monounsaturated fatty acids, branched chain fatty acids, and #- and m.hydroxyacids, which are probably produced by microbial activity, increased. These results indicate that preferential degradation of algal lipids accompanies microbial resynthesis of iipids during settling, however, terrigenous lipids are relatively stable. Key words: early diagenesis, particulate organic matter, sediments, fatty acids, hydroxyacids, alcohols, hydrocarbons

INTRODUCTION

Planktonic algae are important primary sources of lipidsin sediments. However, sedimentary lipidsare differentfrom algal lipidsin terms of the following points. (I) Algae contain abundant polyunsaturated fatty acids (Hitchcock and Nichols, 1971) whereas sedimentary lipids contain lower amounts of polyunsaturated acids (Farrington et al., 1977; Matsuda and Koyama, 1977; Van Vleet and Quinn, 1979; Kawamura eta/., 1980; Cranwell, 1981a, 1984). (2) Algal lipids are mostly solvent extractable but significantportions of sedimentary lipidscannot be extracted with organic solvents.They are present as bound and tightly bound forms (Farrington et aL, 1977; Cranwell, 1981a; Kawamura and Ishiwatari, 1981b, 1984a).(3) Surface sediments contain lipidsof microbial origin (Boon et al., 1977a; Perry et al., 1979; Volkman et al., 1980; Cranwell, 1981b; Kawamura and Ishiwatari,1982).These comparisons indicate that algal lipids,during settlingin the water column, undergo rapid decomposition and transformation by microbial activities,and that bacteria and other microorganisms contribute lipidsto sedimentary organic matter. Early diagenesis of organic matter in the water column has been studied by using sediment traps (Crisp et al., 1979; Wakeham et al., 1980a; Tanoue and Handa, 1980; Wakeham, 1982; Lee and Cronin, 1982; Wefer et al., 1982; Gagosian et al., 1983; De Baar et al., 1983; Gardner et al., 1983; Meyers et al., 1984). De Baar et al. (1983) reported that the rate of *Present address: Department of Chemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, U.S.A.

not loss of carboxylic acids in the North Atlantic Ocean increases with the number of double bonds and decreases with the number of carbon atoms. They considered that branched fatty acids as well as some monocnoic acids are more persistent, probably due to enhanced microbial synthesis during settling which counteracts degradation. Meyers et al. (1984) also reported that lipid components associated with sinking particles undergo substantial microbial degradation during passage to the bottom of Lake Michigan. Although microbial degradation of algal materials in the water column has been seriously considered, bacteria-derived compounds such as ~-hydroxyacids have not been studied in previously conducted sediment trap experiments. To understand better diagenetic changes of organic matter in the water sediment interface, we collected water particulate matter, sediment trap, and bottom sediment samples in a freshwater lake (Lake Haruna). The samples were analyzed for a variety of lipid class compounds (including ~-hydroxyacids) of different origin: algal, bacterial, terrestrial and anthropogenic. Lake Haruna is a mesotrophic lake with no river inflow and the sedimentary organic matter is believed to be mostly of autochthonous origin (Ishiwatari et aL, 1980), The lake sediments are diatomaceous ooze, thus, most of the inorganic materials are amorphous silica of diatom origin, and terrigenous minerals are very minor components (less than 8%, Kawamura et al., 1980). The bottom sediments are under highly reducing conditions (the color is black) and no worms and other animals that cause bioturbation were present. Organic geochemistry of Lake Haruna sediments has been reported previously (Ishiwatari et al., 1980; Kawamura et al., 1980).

251

252

KIMITAKA KAWAMURA et al. EXPERIMENTAL

Sample collection In 1980, a sediment trap experiment was conducted in Lake Haruna (area: 1.23 km2: maximum depth: 14 m). Two sediment traps (14.5 era diameter × 45 cm length) were set at a depth of 11.5 m in the center of Lake Haruna (total depth: 14m) from Oct. 23 to Nov. 16 (25 days). No poison was added to the traps. Trapped material (sediment trap: ST, 2.88 g dry weight) was frozen, transported to the laboratory and stored at -20°C prior to analysis. Water samples were collected at the same location on Nov. 17 at depths of 0, 4, 8, and 10m and were immediately filtered using Whatman GF/C filters to isolate particulate matter (P). Surface (0--10 era) sediments were also collected at the center of the lake using an Ekman dredge. The top surface layers (nepheloid layer) of the collected sediments were immediately separated by decantation and the suspended particles were separated on GF/C filter. This fraction is called hereafter fresh sediments (FS). Both particulate and fresh sediment samples were frozen in the field and transported to the laboratory. Details of sample collection and handling are presented by Ogura et al. (1985) and Table 1 summarizes measurements of carbon and nitrogen contents in the samples.

Separation of lipids Particulate matter (15-21.5mg dry weight), sediment trap (361 nag) and fresh sediment (385mg) samples were extracted three times with benzene/methanol (6:4) using a homogenizer. The extracts (unbound lipids:UB) were saponified with 0.5 M KOH/methanol. The extracted samples were saponified with 0.5 M KOH/methanol (containing 5% water) to separate bound lipids (B). The saponified residues, from which UB and B lipids were removed, were subjected to a harsher saponification at 190°C for 2 hr with 2 M KOH solution in a Pyrex tube to separate tightly bound lipids (TB) which were defined as those released by thermal treatment with KOH of pre-saponified sediment (Kawamura and Ishiwatari, 1982, 1984a). The alkaline solutions (containing lipids) were extracted with n-hexane/ethyl ether (9:1) to separate neutral components and then acidified to pH 1 with concentrated HCI, followed by extraction with n-

hexane/ethyl ether (9:1) to separate acidic components. Neutral components were fractionated into aliphatic hydrocarbons, aromatic hydrocarbons, and alcohols on a silica gel column (Kawamura and Ishiwatari, 1984b). The acid fraction was methylated with 14% BFflmethanol and the methyl esters were divided into a monocarboxylic ester fraction and a hydroxy plus dicarboxylic ester fraction by silica gel column chromatography. The former fraction was further divided into saturated and unsaturated esters by AgNO3-SiO: column chromatography. The latter fraction was treated with a TMS reagent (BSA) prior to gas chromatographic (GC) analysis. Details of the separation are given by Kawamura et al. (1980) and Kawarnura and Ishiwatari (1984b).

GC and GC-MS analysis Aliphatic hydrocarbons were analyzed with a Shimadzu-LKB 9000 GC-mass spectrometer (GC-MS) on a glass column packed with 1% OV-1 on Chromosorb W (AW, DMCS). The column temperature was programmed from 100 to 290°C at 8°C/rain. Quantification was performed by mass fragmentography (m/z 57) using authentic nC~b-nC36 alkanes as external standards. Aromatic hydrocarbons were analyzed on 1% OV-17 packed columns and were determined by mass fragmentography at m/z M + (molecular ions) using authentic polynuclear aromatic hydrocarbons. The column temperature was programmed from 180 to 290°C at 8°C/rain. Saturated and unsaturated fatty acid methyl esters were analyzed by a Shimadzu 6A GC on 1% OV-I and 3% DEGS glass columns, respectiveiy, co-Hydroxyacid methyl esters were determined as TMS ethers on a OV-I column using a GC-MS. fl-Hydroxyacid methyl esters (as TMS ethers), a,o~-dicarboxylic acid dimethyl esters, and fatty alcohols (as TMS ethers) were analyzed on the same column by mass fragraentography using ions at m/z 175, 98 and M-15, respectively. RESULTS

Normal Ci5_33 hydrocarbons, eight polynuclear aromatic hydrocarbons (PAHs), normal C12_32saturated fatty acids, branched chain (iso/anteiso) Ct3, Ct4, Ct5 and Cj7 saturated fatty acids, C~6 and Ct8 mono-and poly-unsaturated fatty acids, fl-hydroxy Ct0-30 acids, og-hydroxy C~2_2s acids, C9-2s

Table I. Carbon and nitrogen contents in the particulate, sediment trap and fresh ~ l i m e n t samples collected from Lake Havana (from Ogura et at. 1985) Samples Particulates 0 m 4m 8m 10 m Sediment trap Fresh sediment Surface Icdimcnt (O--lOom)

Carbon Nitrogen (mg/g dry weight) 243 300 274 267 107 102 96.8

36 50 46 44 14 12 9.2

C/N ratio (weight) 6.7 6.0 5.9 6. I 7.5 8.6 10.5

Early diagenesis of organic matter in the water column and sediments

253

Table 2. Concentrations of lipid class compounds in unbound (UB), bound (B) and tightly bound (TB) fractions separated from particulate, sediment trap and fresh sediment samples of Lake H a r u n a Concentration (mg/g carbon) Sample Particulates 0m

4m

UB B TB UB B

8m

TB UB B

10 m

TB UB

n-HC

PAH

n-SFA

br-FA

UFA

p-OH

co-OH

~,co-Di

n-ALC

Phytol

0.54 --0. I I

0.0056 --0,0031

46 16 1.1 54

2.2 0.78 0.16 1.3

35 1.7 -49

0.95 1.3 0.25 0.48

0.53 ND ND ND

0.49 ND ND ND

0.89 0.20 0.045 0.33

9.5 0.45 ND 15

. ND

0.34

7.2

. --

0A0

5.2

.

. 0.35 .

.

.

. 0.0044 .

. 0.48

.

. 67 .

. 0.0033

[3

.

.

TB

.

.

.

. .

.

.

. 43

.

.

.

.

. 2.2 .

. 1.2 .

.

.

. 44 . 26

.

.

. 0.55

.

. ND

.

.

.

. ND

.

.

.

.

. 0.81

.

.

.

.

.

.

.

.

.

.

Sediment trap UB B TB

0.26 ---

0.0072 0.0052 0.00066

4.6 9.0 2.2

1.0 1.3 0.87

8.2 7.5 0.42

0.25 1.7 1.2

0.075 2.4 0.17

0.028 0.50 0.056

0.21 0.42 0.16

0.26 0.61 0.009

UB B TB

0.18 ---

0.0082 0.0025 0.00069

9.3 5.1 2.2

1.7 1.3 0.68

11 5.2 0.49

0.64 2.8 0.89

0.42 1.9 0.12

0.059 0.56 0.049

0.38 0.57 0.13

1.93 0.79 0.009

Fresh sediment

n-HC: n-hydrocarbons; PAH: polynuclear aromatic hydrocarbons; n-SFA: n-saturated fatty acids, br-FA: branched chain fatty acids; UFA: unsaturated fatty acids; ~-OH:/~-hydroxyacids; co-OH: ¢o-hydroxyacids; ,~,w-Di: ,,,w-dicarboxylic acids; ALC: alcohols; - - : n o t analyzed; ND: not detected.

a,co-dicarboxylic acids, normal C~2_30fatty alcohols and phytol were detected in the unbound, bound, and tightly bound fractions separated from particulate matter, sediment trap and fresh sediment samples. Hydroxyacids and dicarboxylic acids have not been previously reported in a sediment trap sample. Table 2 gives the concentrations of these compounds.

Aliphatic hydrocarbons

ably brought to the lake from anthropogenic sources by atmospheric dry and wet deposition. Anthropogenic contribution was also suggested to n-alkanes in particulate samples by low CPI (odd/even ratio) values (1.12-1.86, av. 1.44 +0.37). However, sediment trap and fresh sediment samples show higher CPI values of 6.06 and 2.73, respectively, suggesting less anthropogvnic contribution. Although this is mysterious, one possible explanation is that abundant anthropogenic hydrocarbons with no odd/evvn pre-

Mass fragmentograms (m/z 57) of the unbound aliphatic hydrocarbon fraction are shown in Fig. 1 for the sediment trap and fresh sediment samples. n Normal alkane distributions are characterized by odd/even predominance with two major peaks at Cl~ SEDIMENT TRAP and C29. The high abundance of nC~7 alkane in the particulate matter and sediment trap samples suggests a significant contribution from algae to the n-alkanes since n C l 7 is a dominant alkane of bluegreen algae (Geipi et al., 1970). However, the amount of n C l 7 alkane decreased from sediment trap to fresh sediment samples, whereas normal C27, C~ and C31 alkanes which probably originated from terrigenous plant waxes and soil organic matter (Ishiwatari et al., 1980) are almost constant, suggesting a selective degradation of shorter-chain (algal) over longerchain (plant wax) alkanes. An unresolved complex mixture (UCM) of branched and cyclic hydrocarbons was also present in the samples. They appeared as a hump on the gas chromatogram (Fig. 1). UCM hydrocarbons probably originate from anthropogenic sources, including incomplete combustion of fossil fuels. Because UCM 151 w , w i ~ ) 1 a i ) .~ ¼ i ) S~OiIS! hydrocarbons are reported in urban atmospheric CARBON NUMBER samples: aerosols (Simoneit and Mazurek, 1982), dry Fig. 1. Mass fraBmentograms (m/z 57) of unbound aft. fallout (Matsumoto and Hanya, 1980) and rainwater phatic hydrocarbons in the sedimenttrap and surface (fresh) sediment samples of Lake Haruna. (Kawamura and Kaplan, 1983, 1986), they are prob-

254

KIMITAKA KAWAMURA et al. 1

,

1

1

1

1

1

I

OM

/0 M

4M

ST

F-~E -.4 ~J

g

8M

2

3

4

5

6

7

FS

8

1

2

3

4

5

6

7

8

Fig. 2. Relative abundance of polynuclear aromatic hydrocarbons (PAHs) in water particles (0, 4, 8, 10 m), sediment trap (ST) and fresh sediments (FS) of Lake Haruna. For the names of PAHs, see text. dominance have been temporarily supplied to the lake water before the sampling probably through the atmosphere. Aromatic hydrocarbons

The following polynuclear aromatic hydrocarbons (PAHs) were detected in the sediment trap sample: No. 1, anthracene/phenanthrene; No. 2, fluoranthene; No. 3, pyrene; No. 4, benzophenanthrene; No. 5, chrysene/l,2-benzanthracene; No. 6, benzofluoranthene; No. 7, benzopyrene (a and e); No. 8, perylene. These PAHs except for perylene are probably of anthropogenic origin and are brought through the atmosphere since they are present in autoexhausts (Boyer and Laitinen, 1975; Graedel, 1978) and are reported in aerosols (Gordon, 1976; Graedel, 1978) and rainwaters (Lunde et al., 1977; Kawamura and Kaplan, 1986). Figure 2 presents relative abun-

dances of these PAHs in particulate, sediment trap and fresh sediment samples. Interestingly, perylene, which was not detected in particulate matter samples, appeared in sediment trap and became the most abundant PAH in the fresh sediment sample. This point will be discussed later. Fatty acids

The particulate samples show that saturated C~4 and C]6 acids and polyunsaturated C~s acids are major components whereas monoenoic acids and C20-32 acids are minor (Fig. 3 and Table 3). Conversely, the sediment trap sample shows that polyunsaturated acids are less abundant than monounsaturated acids (Table 3) whereas C~3o acids slightly increase (see Fig. 3). The fresh sediment sample also gives a distribution pattern similar to that of sediment trap sample; however, the C20-32 acids

IO0 (d) 10 m

(o) 0 m 60 40 20 0

LI "-'

..,I

.h

'~ (b)

_.

( e ) Sediment Tr'op

4m

6O <

4O

~-

O

8O

.,

II.

=l

(c ) 8 m

,,

¢'1~:1 iJ :

40

( ~ ) Fres~ ~ . ~ i m e n t

i C18:n

20 0

. . . . |1 12 14 16 1B ; D 22 ;14 26 2 6 3 0 : 1 2 Corbon

I , J :l . • _I 12 14 16 18 2 0 22 ~

, _. _ 26 2 6 3 0 32

Number

Fig. 3. Relative abundance of fatty acids (unbound form) in the particulate, sediment trap and fresh sediment samples taken from Lake Haruna. Solid, dotted and dashed bars indicate straight-chain saturated acids, branched-chain acids and unsaturated acids, respectively. Cis:, (n = 1-4) means sum of mono- and polyunsaturated C~s acids (see Table 3).

Early diagenesis of organic matter in the water column and sediments

255

Table 3. Unsaturated fatty acids (unboundfraction) in the particulate, sediment trap and fresh sediment samples of Lake Haruna Fatty acids (/~8/gdry sample) Particulates

0m 4m 8m 10 m

Sediment trap Fresh sediment

230 240 340 220 550 670

440 610 560 230 160 270

1950 4820 3650 2050 39 66

[~- and ¢o-hydroxyacids Both straight and branched (iso/anteiso) chain ~-hydroxy Cj0-es acids were detected in the samples, as shown in Fig. 4. In general, each fraction shows that straight chain acids with even carbon numbers are more abundant than those with odd carbon numbers, whereas branched acids with odd carbon numbers are more dominant than those with even carbon numbers, except for particulate samples in which branched Cl4 is more abundant than nCi,. /~-hydroxyacids are most abundant in bound fraction. Particulate matter samples give maxima at brows (UB) and at brCl4 (B and TB) whereas both sediment trap and fresh sediment samples show peaks at brC~7 (UB), brCis (B) and nC~6 (TB). Higher molecular weight ~-hydroxyacids ( > C ~ ) are minor components in all the samples.

•o[ ~

o

4840 4470 5500 3490 254 399

590 730 990 900 31 53

oL,o~-dicarboxylic acids Chain-length distributions of C~_~0a,co-diacids are characterized by even/odd predominance in the Cns-~ range and a weak odd/even predominance in the C~n, range, as shown in Fig. 6. Two maxima were observed at Cn6 and C,,: however, Cts was more abundant than C~6 in TB fraction of the sediment trap sample and in UB and TB fractions of the fresh sediment sample. In particulate matter samples, diacids were not detected except for 0 m, where C~5-3o diacids are present in UB form but not present in B and TB fractions. On the other hand, diacids in

/~C14

ii) ,, ,I

2300 2900 2360 1330 59 58

B

~i

,o

3400 5690 4820 2820 62 65

Distributions of o~-hydroxyacids in particulate, sediment trap and fresh sediment samples are shown in Fig. 5. Ci6 is the only ¢o-hydroxy acid detected in 0m particulate sample, oJ-Hydroxyacids were not detected in 4, 8 and 10 m particulate samples. On the other hand, a homologous series of o~-hydroxy Cjz_~ acids were found in sediment trap sample, which shows a strong even carbon predominance with maxima at Cj6 and C24- These acids are mostly in B form (91%). The TB fraction, which is the second important form for o~-hydroxyacids, contains only Cnz-ls acids. Similar distributions were also observed for the fresh sediment sample, except that C~ is most abundant.

became more abundant. Branched chain C~s acid appears to increase in the order P < ST < FS (Fig. 3). Tightly bound (TB) saturated fatty acids were detected in the samples, in addition to UB and B forms. The TB forms comprise only minor portions (2%) of total fatty acids in the water column (0 m) but they are not insignificant in sediment trap and fresh sediment samples, where they comprise ~ 13% of total saturated acids.

UB

190 460 380 210 8 12

)1i.~i..,

ii

I il;~,.

P='tk~te Om

t

..

|,• d. i_

Sllmlnt trop l

i

0

"f

.......

I

,,

l, ,il.... Fresh

BO

)

sedrn~t

40

o

1 ~

12 14 Is ~

Cort)on

Fig. 4. Distribution of ~-hydroxyacidsin the particulate (0 m), sediment trap and fresh sediment samples taken from Lake Haruna. UB: unbound form; B: bound form; TB: tightlybound form. Straight line means normal acids whereas dotted line means branched acids.

KIMITAKA KAWAMURA et al.

256

~o0 Ptwtio.~t e 0 m (UB)

-I ,, 0

T

0

Se=W~.t t ~

(S)

,l

I

,

I

3

30 Fr'es~ seek~mt (S)

10

0

J, II

,,

12 14 ~e 1is 20 22 24 2s m

Carbon N u m b e r

Fig. 5. Distributions of ,~-hydroxyacids in the particulate (0 m), sediment trap and fresh sediment samples from Lake Haruna. UB: unbound form; B: bound form. sediment trap and fresh sediment exist as UB, B and TB forms and more than 80% of diacids are present in bound form whereas small portions are in unbound (ST: 5%; FS: 9%) and tightly bound forms (ST: 10%; FS: 7%).

30 (a)

Particulate

0 m

UB

20

~0

,l,i, i[illl,,

0 in

30 { b ) Sediment trap

U

< ._u 20 c1

B

~0

{_

0

U

/5

0 30

,,l,l,,

, if,l, ,I,, (c) Fresh

sediment

B 20

o

..,~o 12.,,,14 16 ,I,I, , ,l,, 18 2o 22 24 26 2g 30

Carbon N u m b e r Fig. 6. Distributions of,,,cu-dicarboxylicacids in the particulate (0 m), sediment trap and fresh sediment samples from Lake Haruna. UB: unbound form; B: bound form.

Alcohols

Normal fatty alcohols show an even/odd predominance with a peak at C1:, C~8 or C20 (Fig. 7). Most of the alcohols for the 0 m particulate sample exist as UB form (79%). On the other hand, sediment trap and fresh sediment samples showed that more than half of n-alcohols are present in B form and their distributions are different from those of particulate samples (see Fig. 7b). Unbound n-alcohols in the sediment trap and fresh sediment samples are characterized by bimodal distribution with peaks at C,6 and Cu (ST) or Cn (FS) and by an even carbon number predominance, whereas B and TB alcohols show a unimodal distribution with a peak at C:0 and with an even carbon number predominance (see Fig. 7b). It is of interest to note that n C:0 alcohol was the dominant species released from surface sediment of Lake Biwa during hydrous pyrolysis (Kawamura and Ishiwatari, 1985c). Phytoi was identified in the samples. Its concentrations were tentatively calculated by using nC:0 alcohol-TMS ether on gas chromatogram (TICM) as shown in Table 2.

DISCUSSION

Sediment traps collect particles sinking in the water column. However, if resuspension of the bottom sediments occurs, they also collect the resuspended particles. Resuspension of bottom sediments has been studied in the slope region of southern Lake Michigan using sediment traps (Chambers and Eadie, 1981; Eadie et aL, 1984). Although Lake Haruna has no inflowing river, there is a possibility that the bottom sediments were resuspended by waves, and a circulation of lake water after the thermocline disappeared from the water column toward the end of the sediment trap deployment period. Therefore, prior to considering early diagenesis of organic matter in the settling process, we estimate the possible resuspension of bottom sediments. Assuming a sedimentation rate of 0.65 mm/year (which was estimated for the top 50cm sediments from a dated volcanic ash layer) and dry sediment mass of 0.16g/cm 3 for surface sediments of Lake Haruna (lshiwatari et al., 1980), local sediment accumulation rate (SAR) is calculated to be 0.29 g/m2/day on dry basis. In comparison, the mass flux obtained from the single sediment trap experiment is 3.5 g/m:/day. The mass flux value during the experiment is 12 times higher than the SAR. This difference suggests that (1) trapped materials are mostly (92%) from resuspension of bottom sediments, if annual and seasonal variation of the mass flux is negligible, or (2) mass flux was enhanced in the season (fail) of the experiment and/or the sedimentation rate was underestimated. Supplemental sediment trap experiments indicated that seasonal and annual changes in mass flux are significant in Lake Haruna: mass flux

Early diagenesis of organic matter in the water column and sediments

257

(o) Particulate samples Om

UB

4m

UB

2O 10

, ,'1,'1, I,,I

[.I,l,lll,ll,,.,.,

I,I li.,

,I,I Ill i i,,.,.,

Om

'5

2O

io h

0

J

B

I

8m

= Om

TB

UB

lOrn

UB

20 10 0

| |1

. I

12 14 16 1B 20 22 ~4 26 215 30

12 14 16 1~ 20 22 24 26 2B30

Carbon

(b

Number

Sediment trap

Fresh

lO

sediment

UB

uB

8 6 4 A

2 o

E

,.,.l., ,,l,l.l.l.

'il, I.l,l,l,l,,,.

,,. ,i,1,,.,. B

I.I,

O"

I e"

,

I .

I,, 1,1.,.,, TB

TB

O"

..,.I,i.l,,.,

12 14 16 1 8 2 0 2 2

......

24 2 6 2 6 3 0

•

,_,,I,II1,,

.

.

.

.

.

12 14 16 le, 20 22 24 26 26 30

Carbon Number

Fig. 7. Distribution of n-fatty alcohols in (a) the particulate samples, and (b) the sediment trap and fresh sediment samples, taken from Lake Haruna. UB: unbound form; B: bound form; TB: tightly bound form. increased from summer (0.9g/mZ/day) to fall (1.9 g/m2/day) in 1981 (unpublished data). However, these values are still much higher than the SAR. If major portions of trapped materials came from resuspended bottom sediments as suggested above (92% is resuspension), molecular compositions of sediment trap sample should strongly reflect those of the fresh sediment sample. However, the distributions for many compounds in the trapped materials are distinct from those of fresh sediment sample. For example, n-alkanes in sediment trap sample show a predominance of C~7 whereas the fresh sediment sample shows less abundant C~7 (see Fig. 1). PAHs

also show different distribution patterns between sediment trap and fresh sediment samples: benzofluoranthene is the dominant species in the sediment trap whereas perylene is the most abundant in the fresh sediment sample (Fig. 2). Other examples can be seen in the distributions of C~-C~2 fatty acids (Fig. 3), bound p-hydroxyacids (Fig. 4), oJ-hydroxyacids (Fig. 5) and unbound alcohols (Fig. 7b). These results do not agree with the above estimated resuspension (92%). Hence, we believe that resuspension value (92%) obtained from mass flux and sedimentation rate is largely overestimated. The sedimentation rate used for the above calculation was

258

KIMITAKA KAWAMURA et al.

probably underestimated, although there are no data for the present-day sedimentation rate, only an average for the top 50 cm. In order to better estimate possible resuspension into the sediment trap, a linear mixing model using water particles and fresh sediments as end members can be applied for different organic compound classes. If trapped materials (represented as ST) are combination of settling particles (represented as P) and resuspended bottom sediments (represented as FS), contributions from resuspension (X) and particulate matter ( 1 - X) can be calculated as follows: [ST] = [P]. (1 - X) + [FS]. X in which [ ] represents concentrations or ratios for different types of components in the sediment trap, particulate matter and fresh sediment samples (see Tables 1 and 2). In this model, diagenetic changes (degradation and bacterial resynthesis) of organic matter in water-sediment interface are assumed not to occur. If this model is applicable to the present situations, X should range between 0 and 1. Table 4 gives resuspension contribution factors (as %) calculated for various components. The results for C and N contents indicate that 97% of C and 94% of N in the trapped materials are from resuspension of the fresh sediments, and the rest (3% of C and 6% of N) come from water particles (Table 4). Because of degradation of organic matter in the water column, these values are probably overestimated. In fact, when C/N ratio of organic matter is used, the calculation gives lower value: 54% of organic matter in the sediment trap sample originate from resuspension and 46% from particulate matter. However, this may be still overstated since C/N ratios of algal detritus are known to increase during microbial degradation (Otsuki and Hanya, 1972); the initial C/N ratios of organic matter collected in the sediment traps should have been lower than the observed ratio (C/N = 7.5). For individual compound class, resuspension factors widely ranged from 28 to 93%, and 103 to 210% for selected compounds (Table 4). The resuspension factors exceeding 100% suggest a production of those compounds during the mixing processes. The big fluctuation (28-93%) among the compound classes suggests that there are compli-

eating processes: selective degradation during diagenesis in the water column or fractionation of organic matter during resuspension. Since decomposition of organic compounds in the water column, which we believe most important, is not taken into account in the mixing model, the resuspension factors obtained from biodegradable compounds such as fatty acids, alcohols and n-alkanes may be overestimated. PAHs are more realistic to the model because they are refractory to short-term biodegradation. However, non-perylene PAHs showed a resuspension of 210%. This abnormal value could be caused by a seasonal variation of the PAH inputs through the atmospheric deposition, which is the major source of non-perylene PAHs as stated before. Perylene is an ideal compound for the model, because inputs of perylene to the lake from the atmosphere and surrounding soils are negligible and it is formed in sediments. The mixing model with perylene gives a resuspension factor of 28% (Table 4). Although this is the most reasonable estimation, the value (28%) may be still overestimated because perylene might have also been formed in the traps under reducing conditions as discussed later. Therefore, we think that the maximum resuspension is 28% and more than 72% of the sediment trap materials came from sinking particles. It should be noted that perylene was not detected in any particulate matter samples in spite of its abundant presence in the surface sediment (Fig. 2), suggesting that resuspension of the bottom sediments into the upper water column is not significant.

Formation of perylene during early stage of diagenesis Perylene has been found in many recent sediments (Orr and Grady, 1967; Aizenshtat, 1973; Ishiwatari and Hanya, 1975; Laflamme and Hites, 1978; Wakeham et al., 1980b; Louda and Baker, 1984) and is considered to be produced from its precursors (perylenequinones) by their reduction under anoxic environments (Orr and Grady, 1967; Aizenshtat, 1973). Gschwend et al. (1983) reportred an exponential increase of perylene concentration with depth in a recent lacustrine sediment core, and proposed an m situ formation of perylene. Prahl and Carpenter (1979) reported the presence of perylene in plankton

Table 4. Resuspension factors (%) for various components obtained from the resuspcnsion model on an amumption that trap materials originate from water particles and bottom sediments and that there is no dialgenetic change in organic matter Component C N C/N ratio n-HC Perylene Non-perylene PAH n-SFA br-FA

Rcmmpension (%) 97 94 54 58 28

(210) (103)

Component mono-UFA poly-UFA ~-OH co-OH Di-acids

n-ALC Phytol

Resuspension (%) 7I (106) 39 (109) 93

46 (130)

71

For abbreviation, see Table 2. The data used for calculation are from Tables l and 2.

Early diagenesis of organic matter in the water column and sediments net and sediment trap samples collected from Dabob Bay, Washington. They concluded that in situ formation is not a major factor and erosion introduces perylene into the Dabob Bay. The presence of perylene in the sediment trap sample of Lake Haruna indicates that (1) perylene could be formed very rapidly in the water column if a reducing environment existed, or, (2) perylene is transported in the trap by resuspension of the bottom sediment or by erosion of soils around the lake. If the erosion is the case, perylene should also be present in water particles. However, it was not detected in the particulate samples, suggesting that erosion is unlikely. As shown in Table 5, the ratios of perylene to other PAH concentrations largely increase in the order of P < ST < FS. The same trend is also found for the ratios of perylene to TOC (Table 5); the ratios increase by a factor of 3.6 from the sediment trap to fresh sediment sample and 6.5 from fresh sediments to surface sediments (0-10 cm). These results suggest (1) a rapid formation of perylene in water-sediment interface under reducing conditions and (2) resuspension of perylene from the bottom sediments. Although a minor resuspension is suggested to have occurred during the sediment trap experiment based on a mixing model with perylene, we cannot preclude in situ formation of perylene in the sediment traps in which reducing conditions were formed by microbial consumption of dissolved oxygen (a smell of H2S was detected in the water of the traps). Production and degradation o f algal lipids in the water column

Particulate organic matter (POM) in Lake Haruna is mostly of autochthonous (algal) origin, based on low C]N ratios of POM (5.9-6.7, see Table 1) which resemble those of phytoplankton (Kawamura and Ishiwatari, 1981a). This origin is also supported by the distributions of fatty acids in POM: predominance of saturated CI4, Cle and C,8 and polyunsaturated Cis acids (see Table 3, Fig. 3), which is similar to algal lipids (Hitchcock and Nichols, 1971). A minor terrigenous contribution to POM is indicated by the presence of long-chain C2o-32fatty acids (Fig. 3), which have been used as biomarkers of higher plants (e.g. Kawamura and Ishiwatari, 1984b). Bacterial cells are probably present in POM because

259

branched-chain fatty acids and/~-hydroxyacids were detected. Zooplankton may also contribute to POM because Cts:3o,6 acid, which is a possible marker of zooplankton (Kawamura and Ishiwatari, 1981a) was detected in particulate matter samples, although its amount is quite low compared to algal-derived Cn:3~3 acid (Table 3). However, the ratios of C~s:3,o6acid to C,s:3~3 acid, which is a possible indicator of zooplankton vs phytoplankton lipids, increase from 0.073+_0.011 (P) to 0.13 (ST) and 0.19 (FS), suggesting a synthesis and/or transformation of organic compounds by zooplankton in the water column. Zooplankton-derived lipids may be transported to the bottom sediments with sinking fecal pellets (Turner and Ferrante, 1979; Prahl and Carpenter, 1979). Carbon contents of the particulate samples showed a maximum at 4 m depth in the water column (see Table 1). The maximum in the subsurface layer is probably related to an enhanced primary production in the lake. This is reasonable because algal polyunsaturated fatty acids also gave the highest concentrations at the same depth (Table 3). A similar vertical profile was obtained for phytol (see Table 2), which probably originates from the phytyl chain of chlorophyll a. However, under the subsurface layers (deeper than 4 m), concentrations of both polyunsaturated fatty acids and phytol decreased with depth. The decrease of relative abundance of these compounds in TOC in the water layers deeper than 4 m (see Table 2) indicates that polyunsaturated acids and phytol undergo selective degradation in the water column. Because these compounds contain double bonds, they may decompose and/or polymerize during sedimentation and settling. Preferential degradation of polyunsaturated acids relative to their saturated analogs appears to occur in sinking particles in Lake Haruna. As shown in Fig. 8, the ratios of unsaturated Cn acids to corresponding saturated acids (U/S ratios) decrease with depth below the sub-surface layers, indicating that these unsaturated acids are less stable than saturated ones in the water column. Similar trends with depth were observed in sediment trap samples collected in the North Atlantic Ocean (DeBaar et al., 1983), surface sediments of this lake (Kawamura et aL, 1980), and other lakes (Matsuda and Koyama, 1977; Meyers et

Table 5. Concentrations of perylene and PAHs (excluding perylene) in Lake Haruna samples ~g/8 dry weight). Unbound form is presented

Ratios of peryleneto Sample Particulate (0 m) Sediment trap Fresh sediment Surface sediment

(0-10cm)b

Perylene ND 0.093 0.318 1.18

"PAH-EP means PAHs excluding perylene. bData are from Ishiwatari et aL (1980).

PAH-EP' 1.35 0.679 0.523 1.75

TOC (rag/g) 240 107 102 57.8

PAH-EP

TOC

0 0.14 0.61 0,67

0 0.87 x 10-6 3.1 x 10-e 20 x 10-6

KIMITAKA KAWAMURAet

260

al.

Unsatulated / Satur~te0 Ratios 0.05 0.1

O01

!o

0.5

2 3

8

I

1~:

10

5

1B:O 1

5 7

c"

11~:313 IB:O

8

9 10

, t "°~.

..

Sediment trap

"-

...

•

~ 0 I..:Y.'

Fresh

sediment

(~O'

~ 0

4

Fig. 8. Ratios of unsaturated fatty acids to corresponding saturated acids in the water column of Lake Haruna as well as sediment trap and fresh sediment samples.

aL, 1980). However, the U/S ratios of Ctr:t did not decrease with depth in the water column and are low even in the subsurface water (4 m), where a maximum algal growth was suggcstcd. This result probably indicates that C~:~ acid is not mostly from algae, rather it comes from microbial sources (Volkman et al., 1980). The U/S ratios for Ctr:l, as well as C:8::, largely increased from the particulate matter to sediment trap sample and slighdy decreased to fresh sediment sample. These results suggest that drastic changes in lipid composition occurred in a short time during settling, probably due to zooplankton feeding of algae and bacterial resynthesis of lipids.

Microbial transformation of lipida during settling The sediment trap sample should have been exposed to microbial attack during settling and in the trap itself because no bacteriostat was used in the sediment trap equipment. Because the period of the sediment trap experiment was short (25 days), a rapid microbial change of lipid components during sedi-

N-Saturated ...~

...

FA

. . . . . . . .

06 ,/¢

(0m) F::vv':::' : :::~:: :JC!I

i

Unsaturated

FA

mentation and settling can be recognized by comparing lipid distribution of the sediment trap sample with those of particulate and fresh sediment samples. Particulate matter, which includes living and dead plankton and is considered to be the source of bottom sediments, has probably been subjected to less degradation than the sediment trap materials due to shorter exposure time whereas organic matter in the bottom sediment should be more degraded. Changes in abundances of various lipid components relative to total organic carbon (TOC) in Lake Haruna samples are illustrated in Fig. 9. The lower molecular weight saturated fatty acids (Ct2-zg: LFAs) decrease from the particulate matter to sediment trap sample, whereas the higher molecular weight saturated fatty acids (C20-30: HFAs) remain almost constant. This result clearly indicates that LFAs are less stable than H F A s and significant parts of the LFAs are lost during settling processes. As a result of selective degradation of LFAs, abundance of H F A s appears to increase in the sediment trap sam-

=.,-[~

br- FA

~1~;.'.',..'.. • ...-.!

~

: :.:.:..::.;~.:'-

I=

i

\

t

•

...i

~04

,oo5

t

c~-c~ ..~:(-"'"~ooo-i " Po,y .," ..;':::::'c~0-c~ ~ ": ,-""

SED,. ~ o 4

~-OH

a-OH

. ~ ...-~.:.: . . . .

, ' . ' ." .- . . ; : .'.-.

1.':

I

.':..:::;::...

Fig. 9. Relative abundance (%) of lipid class compounds in the total organic carbon content, for the particulate, sediment trap and fresh sediment samples taken from Lake Haruna. POM: particulate organic matter; TRAP: sediment trap; SEDI: Fresh sediment. Other abbreviations are in Table 2.

Early diagenesis of organic matter in the water column and sediments pie and then fresh sediment sample (see Fig. 3). This consideration is consistent with a result of Ogura (1984) that during a 60-day incubation experiment of particulate matter collected in Lake Haruna fatty acid distribution patterns drastically changed to those of sediment trap sample. Selective degradation of LFAs is a common process occurring in surface sediments (Matsuda and Koyama, 1977; Meyers et al., 1980; Kawamura and Ishiwatari, 1984b), however, the degradation rate should be much faster during settling than after deposition because more oxygen is supplied to the water column than to the sediments. Interestingly, relative abundances of monounsaturated acids in TOC increases in the order P < ST < FS, as shown in Fig. 9. The concentration (#g/g-dry weight) of CIr:l acid also increased in the same order (see Table 3). This suggests a production of monounsaturated acids during sedimentation. Monounsaturated fatty acids are major components of bacterial lipids (Voikman et al., 1980), although they are widely present in algae and other organisms. Some monounsaturated acids in our sediment trap sample are probably produced during settling due to microbial activity, as suggested for the ocean water column (DeBaar et al., 1983). On the other hand, concentrations of polyunsaturated fatty acids (PUFAs) relative to TOC drastically decreased from particulate matter to sediment trap samples (see Fig. 9), suggesting a rapid decomposition and/or polymerization of these acids in the water-sediment interface. Degradation of PUFAs is probably caused during zooplankton feeding and by bacteria, which do not generally produce PUFAs. Alternatively, nonbiological degradations such as autooxidative crosslinking (Harvey et al., 1984) and photochemicallyinduced oxidation (Kawamura and Gagosian, 1987) may also be responsible for the degradation of unsaturated fatty acids. Microbial activities should enhance the amount of bacterial lipids. Branched chain fatty acids, which are characteristic of bacterial lipids (Kaneda, 1967), increase from the particulate matter to sediment trap samples and sediment trap to fresh sediment samples (Fig. 9), although the increase is small. However, when the ratios of concentrations of branched fatty acids to those of corresponding normal fatty acids are compared among these samples, a clear trend is observed; the ratio for C13, C15 and C~7 acids (UB + B + TB) increases from 1.2 (P) to 3.5 (ST). This result suggests that bacterial lipids are produced during settling. The relative abundance of dicarboxylic acids appears to increase slightly in the same way (see Fig. 9). These diacids are not generally present in algal lipids. They may originate either from microbial oxidation of fatty acids and co-hydroxyacids (Eglinton et al., 1968; Johns and Onder, 1975; Ishiwatari and Hanya, 1975), from water-grass (Volkman et al., 1980) or from landderived higher plans (Cranwell, 1977; Kawamura and

261

Ishiwatari, 1984b). A part of these diacids may be produced by microbial co-oxidation of fatty acids and oJ-hydroxyacids during settling. It is of special interest to note that the relative abundance of /~-hydroxyacids increases from the particulate matter to sediment trap and then fresh sediment samples (Fig. 9). This is strong evidence to support the bacterial transformation of planktonic organic matter during sedimentation, because /~-hydroxyacids are characteristic constituents of bacterial lipids (Moss et al., 1973; Mayberry et ai., 1973; Boon et aL, 1977b; Kawamura and Ishiwatari, 1982) and are not generally present in algae, except for small amounts of these acids detected in cultured algal samples (Matsumoto and Nagashima, 1984). As a result of bacterial growth during settling of algal organic matter, labile compounds such as unsaturated fatty acids and saturated C~2_19acids are subjected to preferential degradation and bacterial lipids are synthesized. However, distribution patterns of /~-hydroxyacids are not consistent among the samples (Fig. 4); branched chain Cl,, Cis and Ci6 acids are dominant species in particulate matter whereas normal Ci2, C,, and CI6 and branched Cj5 are among the most dominant/~-hydroxyacids in the sediment trap and fresh sediment samples. This may be caused by different bacterial populations in particulate (oxic) and bottom sediment (anoxic) samples. The relative abundance of co-hydroxyacids increases from the particulate matter to sediment trap samples (Fig. 9). These acids can be produced by fungi and bacteria (Stodola et al., 1967; Kester and Foster, 1963), and a sea-grass (Volkman et al., 1980), as well as higher plants (Hoiloway, 1972a,b; Cardoso et al., 1977; Kolattukudy, 1980). The increase of ¢o-hydroxyacids (Fig. 9) may suggest microbial production of these acids in the water column under oxic conditions. However, they do not increase from the sediment trap to fresh sediment (anoxic) samples, suggesting that w-hydroxyacids are not produced under anoxic conditions in the sediments. In contrast, PAHs, which are of anthropogenic origin except for perylene, can survive in the watersediment interface, where microbial change of organic matter is strongly indicated. Concentrations of PAHs, excluding perylene, are 5.6, 6.3 and 5.1/~g/g C for 0 m particulate matter, sediment trap and fresh sediment samples, respectively. These values are fairly constant when compared with those for fatty acids, especially polyunsaturated acids (see Tables 2 and 3), indicating that PAHs escape microbial degradation. Preservation of PAHs during settling has also been proposed in Dabob Bay, Washington (Prahl and Carpenter, 1979). Normal alcohols also seem to escape the microbial degradation because their concentrations (UB + B + TB) appear to be constant: 1.1 (0 m P), 0.79 (ST) and 1.1 mg/g C (FS) (Table 2). However, this result does not exclude a microbial loss of the alcohols, which may be compensated by other microbial prod-

KIMITAKA KAWAMURA et al.

262

Table 6. Percentages of unbound (UB), bound (B), and tightly bound (TB) compounds in the total concentration (UB + B + TB) in particulate (0 m), sediment trap and fresh sediment samples of Lake Haruna Lipid (% of the sum of UB + B + TB) Sample Particulate UB B TB Total Sediment trap UB B TB Total Fresh Sediment UB B TB Total

n-SFA

br-FA

UFA

E-OH

~o-OH

Diacids

nALC

73 25 2 100

70 25 5 100

95 5 ND 100

38 52 10 100

100 ND ND 100

100 ND ND 100

79 17 4 100

29 58 13 100

32 41 27 100

52 46 2 100

8 53 39 100

3 91 6 100

5 85 I0 100

26 54 20 100

56 31 13 100

46 35 19 100

67 31 2 100

15 65 20 100

17 76 7 100

9 84 7 100

35 53 12 100

For the compound names, see Table 2. ND: not detected.

uction. In fact, the distribution of n-alcohols (B and TB forms) of the sediment trap and fresh sediment samples are different from those of the particulate matter samples (UB, B and TB forms) (see Fig. 7), indicating that microbial production of n-alcohols is possibly occurring in the water-sediment interface.

Rapid formation of bound and tightly bound lipids Table 6 compares percentages of unbound (UB), bound (B) and tightly bound (TB) carboxylic acids and alcohols in total concentrations (UB + B + TB) for the particulate matter, sediment trap and fresh sediment samples. The particulate sample shows the lipid components are mostly in UB form (more than 70%), except for co-hydroxyacids which are mostly in B form. On the contrary, the sediment trap and fresh sediment samples show that the percentages of B and TB lipids increase up to 91% (B form) and 39% (TB form), respectively, depending on lipid components. These differences suggest that a rapid transformation of unbound lipids to bound and tightly bound forms is occurring in the water column and surface sediments. This is consistent with a result of Meyers et al. (1984) that bound/unbound ratios for fatty acids collected at deeper water layers are greater than those of shallower layers in Lake Michigan. Formation of B and TB lipids may be related with microbial activity. Cranwell (1981a) reported that branched fatty acids in recent sediments, which are of bacterial origin, are more abundant in B form than UB form, and considered that B lipids are associated with bacterial activity. Kawamura and Ishiwatari (1982, 1984a) found that //-hydroxyacids in recent sediments, which are also of bacterial origin, are abundant in the order UB < B < TB. These results suggest that the increase of B and TB lipids relative to UB lipids in our sediment trap and fresh sediment samples are caused during microbial attack on planktonic organic matter. TB lipids may be associated

with humic substances because a good correlation was found between the concentrations of TB lipids and humic compounds in a Lake Biwa sediment core (Kawamura and Ishiwatari, 1984a). The presence of B and "I'B lipids in the sediment trap sample indicates that these lipids are rapidly formed during sedimentation and settling processes, probably as a result of microbial activity.

CONCLUSIONS

The study of sediment trap, water particulate matter and fresh bottom sediment samples collected in freshwater Lake Haruna have shown the geochemical and biogeochemical changes of organic matter associated with sinking particles. Comparison of the molecular distributions for a variety of lipid class compounds including//-hydroxyacids in the sediment trap sample with those in particulate matter and fresh sediment samples indicates that a significant microbial degradation of algal-derived organic matter occurs in the water column during settling and that, in response to this, bacterial lipids are synthesized. These microbial activities, as well as zooplankton feeding processes, cause a dynamic change in the distribution of organic compounds in sinking particles. Following are our major conclusions: (1) A preferential degradation of fatty acids occurs, namely, algal-derived polyunsaturated acids (C~s:2 and C~8:3) in the water column decompose and/or polymerize much faster than the corresponding saturated one (Cts:0), and short-chain saturated fatty acids ( < Cz0) of autochthonous origin are more labile than longer-chain saturated ones (C20-32) of terrestrial higher plant origin. (2) Relative abundance of bacterial lipids (//-hydroxyacids) in TOC increases in the order:

Early diagenesis of organic matter in the water column and sediments P < ST < FS. Microbial degradation and resynthesis of algal-derived particulate matter are important processes in controlling the distribution of lipid compounds in the water column and surface sediments. (3) Zooplankton are also suggested to be responsible to degradation of phytoplankton-derived organic matter because the ratio of Cjs:3,~ acid to Cis:3~3 acid, which we proposed as possible indicator of zooplankton vs. phytoplankton lipids, increases in the order: P < ST < FS. (4) Bound and tightly bound forms of carboxylic acids and alcohols, which are minor in the particulate matter but significant in the sediment trap and fresh sediment samples, are formed during early diagenesis in water-sediment interface. Transformations of lipids may be associated with microbial activity and formation of humic substances (geopolymers). (5) Reduction of perylenequinones to perylene is suggested to occur during settling when a reducing environment is formed as a result of microbial consumption of dissolved oxygen. Anthropogenic PAHs are stable in water-sediment interface, escaping microbial degradation. (6) Maximum resuspension of bottom sediments was estimated to be 28% based on a mixing model with perylene: more than 72% of the sediment trap materials came from sinking particles. Acknowledgements--This study was conducted as a part of cooperative work with the following people: Drs K. Fukushima, M. Ochiai, M. Shioya, S. Yamamoto, S. Yun and T. Hayashi. They are all acknowledged. We also thank Dr N. Handa for providing sediment trap equipment, Drs P. A. Meyers, B. J. Eadie, R. B. Gagnsian and J. R. Ertel for their critical and helpful comments, and Ms. M. Harvey and M. Lumping for typing the manuscript. Ttfis study was supported in part by the Ministry of Education, Science and Culture, Japan. One of the authors (K.K.) was financially supported for his post-doctoral scholarship by Japanese Promotion for Science.

REFERENCES

Aizenshtat Z. (1973) Perylene and its geochemical significance. Geochim. Cosmochim. Acta 37, 559-567. Boon J. J., De Lange F., Sehuyl P. J. W., DeLeeuw J. W. and Schenck P. A. (1977a) Organic geochemistry of Walvis Bay diatomaceaous ooze---II. Occurrence and significance of the hydroxy fatty acids. In Advances in Organic Geochemistry 1975 (Edited by Campos J. and Goni J.), pp. 255-272. Enadimsa, Spain. Boon J. J., DeLeeuw J. W., Hock (3. J. V. D. and Vosjan J. H. (1977b) Significance and taxonomic value of iso and anteiso monoenoic fatty acids and branched ~-hydroxyacids in Desulfovibrio desulfuricans. J. Bacteriol. 129, 1183-1191. Boyer K. K. and Laitinen H. A. (1975) Automobile exhaust particulates. Environ. Sci. Technol. 9, 457--459. Cardoso J. N., Eglinton G. and Holloway P. J. (1977) The use of cutin acids in the recognition of higher plant contribution to recent sediments. In Advances in Organic Geochemistry 1975 (Edited by Campos J. and Goni J.), pp. 273-287. Enadimsa, Spain. Chambers R. L. and Eadie B. J. (1981) Nepheloid and

263

suspended particulate matter in south-eastern Lake Michigan. Sedimentoiogy ~ 439-447. Cranwell P. A. (1977) Organic geochemistry of Can Loch (Sutherland) sediments. Chem. Geol. 20, 205-221. Cranweli P. A. (1981a) Diagenesis of free and bound lipids in terrestrial detritus deposits in a lacustrine sediment. Org. Geochem. 3, 79-89. Cranweli P. A. (1981b) The stereochemistry of 2- and 3-hydroxy fatty acids in a Recent lacustrine sediment. Geochim. Cosmochim. Acta 45, 547-552. Cranwell P. A. (1984) Lipid geochemistry of sediments from Upton Broad, a small productive lake. Org. Geochem. 7, 25-37. Crisp P. T., Brenner S., Venkatesan M. I., Ruth E. and Kaplan I. R. (1979) Organic chemical characterization of sediment trap particulated from San Nicolas, Santa Barbara, Santa Monica and San Pedro Basins, California. Geochim. Cosmochim. Acta 43, 1791-1801. DeBaar H. J. W., Farrington J. W. and Wakeham S. G. (1983) Vertical flux of fatty acids in the North Atlantic Ocean. J. Mar. Res. 41, 19-41. Eadie B. J., Chambers R. L., Gardner W. S. and Bell G. L. (1984) Sediment trap studies in Lake Michigan: Resuspension and chemical flux in the southern basin. J. Great Lakes Res. 10, 307-321. Eglinton G., Hunneman D. H. and Duraghi-Zadeh K. (1968) Gas chromatographic--mass spectrometric studies of long chain hydroxyacids---II. The hydroxyacids and fatty acids of a 5000-year-old lacustrine sediment. Tet. rahedron 24, 5929-5941. Farrington J. W., Henrichs S. M. and Anderson R. (1977) Fatty acids and Pb-210 geochronology of a sediment core from Bn~Ards Bay, Massachusetts. Geochim. Cosmochim. Acta 412, 289-298. Gagnsian R. B., Smith S. O. and NigreUi G. E. (1982) Vertical transport of steroid alcohols and ketones measured in a sediment trap experiment in the equatorial Atlantic Ocean. Geochim. Cosmochim. Acta 46, 1163-1172.

Gardner W. D., Hinga K. R. and Marra J. (1983) Observarious of the degradation of biogenic material in the deep ocean with implications accuracy of sediment trap flux. J. Mar. Res. 41, 195-214. Gelpi E., Schneider H., Mann J. and Oro J. (1970) Hydrocarbons of geochemical significance in microscopic algae. Phytochemistry 9, 603-612. Gordon R. J. (1976) Distribution of airborne polycycfic aromatic hydrocarbons throughout Los Angeles. Environ. Sci. Technol. 10, 370-373. Graedel T. E. (1978) Chemical Compounds in the Atmo. sphere, 440 pp. Academic Pr--,;~s,New York. Gschwend P. M., Chen P. H. and Hites R. A. (1983) On the formation of perylene in recent sediments: kinetic models. Geochim. Cosmochim. Acta 47, 2115-2119. Harvey G. R., Boran D. A., Piotrowicz S. R. and Weisel C. P. (1984) Synthesis of marine humic substances from unsaturated lipids. Nature 309, 244-246. Hitchcock C. C. and Nichols B. W. (1971) Plant Lipid Biochemistry, 387 pp. Academic Press, London. Holloway P. J. (1972a) The composition of suberin from the corks of Quercm subcr L. and Betula pendula Roth. Chem. Phys./.~p/ds 9, 158--170. HoUoway P. J. (1972b) The suberin composition of the cork layers from some ribes species. Chem. Phys. Lipids 9, 171-179. Ishiwatari R. and Hanya T. (1975) Organic geochemistry of a 200-meter core sample from Lake Biwa. If. Vertical distribution of mono- and di-carboxylic acids and polynuclear aromatic hydrocarbons. Proc. Japan Acad. 51, 436--441. Ishiwatari R., Ogura K. and Horie S. (1980) Organic geochemistry of a lacustrine sediment (Lake Haruna, Japan). Chem. Geol. 29, 261-280.

264

KIMITAKAKAWAMURAet al.

Johns R. B. and Ondcr O. M. (1975) Biological diagenesis: dicarboxylic acids in recent sediments. Geochim. Cosmochim. Acta 39, 129-t36. Kaneda T. (1967) Fatty adds in the genus Bacillus I. Isoanteiso- fatty acids as characteristic constituents of lipids in 10 spe~es. J. Bacteriol. 93, 894-903. Kawamura K. and Ishiwatari R. (1981a) Polyunsaturated fatty acids in a lacustrine sediment as a possible indicator ofpaleoclimate. Geochim. Cosmochim. Acta ~ , 149-155. Kawamura K. and lshiwatari R. (1981b) Experimental diagenesis of fatty acids in a sediment: changes in their existence forms upon heating. Geochem. J. 15, 1-8. Kawamura K. and Ishiwatari R. (1982) Tightly bound ~-hydroxyacids in a Recent sediment. Nature 297, 144-145. Kawamura K. and Kaplan I. R. 0983) Organic compounds in the rainwater of Los Angeles. Environ. Sci. Tectmol. 17, 497-501. Kawamura K. and Ishiwatari R. (19Ma) Tightly bound aliphatic acids in Lake Biwa sediments: Their origin and stability. Org. Geochem. 7, 121-126. Kawamura K. and Ishiwatari R. (1984b) Fatty acid geochemistry of a 200 m sediment core from Lake Biwa, Japan: Early diagenesis and paleoenviromnental information. Geochim. Cosmoehim. Acta 48, 251-266. Kawaraura K. and Ishiwatari R. (1985) Behavior of lipid compounds on laboratory heating of a recent sediment. Geochem. J. 19, 113-126. Kawamura K. and Kaplan I. R. (1986) Biogenic and anthropogenic organic compounds in rain and snow samples collected in southern California. Atmos. Environ. 20, 115-124. Kawamura K. and Gagosian R. B. (1987) Implication of ¢o-oxocarboxylic acids in the remote marine atmosphere for photo-oxidation of unsaturated fatty acids. Nature 325, 330-332. Kawamura K., Ishiwatari R. and Yamazaki (1980) Identification of polyunsaturated fatty acids in surface lacustrine sediments. Chem. Geol. 28, 31-39. Kester A. S. and Foster J. W. (1963) Diterminal oxidation of long-chain alkanes by bacteria. J. Bacterioi. 85, 859-869. Kolattukudy P. E. (1980) Biopolyester membranes of plants: Cutin and suherin. Science ~ g , 990-1000. Laflamme R. E. and Hires R. A. (1978) The global distribution of polycyclic aromatic hydrocarbons in recent sediment. Geochim. Cosmochim. Acta 42, 289-303. Lee C. and Cronin C. (1982) The vertical flux of paxticulate organic nitrogen in the sea: decomposition of amino acids in the Peru upwelling area and the Equatorial Atlantic. J. Mar. Res. 40, 227-251. Louda J. W. and Baker E. W. (1984) Perylene occurrence, alkylation and possible sources in deep-ocean sediments. Geochim. Cosmochim. Acta 48, 1043-1058. Lunde G., Gether J., Gjos N. and Lande M. S. (1977) Organic micropollutants in precipitation in Norway. Atmos. Environ. 11, 1007-1014. Matsuda H. and Koyama T. (1977) early diagenesis of fatty acids in lacustrine sediments. I. Identification and distribution of fatty acids in recent sediment from a freshwater lake. Oeochim. Cosmochim. Acta 41, 777-783. Matsumoto G. and Hanya T. (1980) Organic constituents in atmospheric fallout in the Tokyo area. Atmos. Environ. 14, 1409-1419. Matsumoto G. and Nagashima H. (1984) Occurrence of 3-hydroxyacids in microaiga¢ and cyanohact,,q'ia and their geochemical significance. Geochim. Cosmochim. Aeta 48, 1683-1687. Mayberry W. R., Smith P. R., Langworthy T. A. and Plackett P. (1973) identification of the amide-linked fatty acids of Acholeplasma axamthum $743 as D(-)3-hydroxy-hexadecanoate and its homoiogues. J. Bacteriol. !16, 1091-1095.

Meyers P. A., Bourbonniere R. A. and Takeuchi N. (1980) Hydrocarbons and fatty acids in two cores of Lake Huron sediments. Geochim. Cosmochim. Acta 44, 1215-1221. Meyers P. A., Leenheer J. M., Eadie B. J. and Maule S. J. (1984) Organic geochemistry of suspended and settling particulate matter in Lake Michigan. Geochim. Cosmochim. Acta 48, 443-452. Moss C. W., Samuels S. B., Liddle J. and McKinney R. M. (1973) Occurrence of branched-chain hydroxy fatty acids in Pseudomonas maltophilia. J. Bacteriol. 114, 1018-1024. Ogura K. (1984) Early diagenesis of lipid materials in plankton, particulates, trap samples, and sediments (abstr.). The 1984 International Chemical Congress of Pacific Basin Societies, Honolulu, Dec. 16-21, 1984. Ogura K., Ochiai M., Kawamura K., Fukushima K., Shioya G., Yamamota S. and Ishiwatari R. (1985) Early diageuesis of organic matter in water of Lake Haruna--I. Flux of organic matter to the bottom by determination of carbon and nitrogen of sediment trap sample, particulates and sediments. Jpn. J. Limnol. 46, 297-302 (in Japanese with an English abstract). On- W. L. and Grady H. R. (1967) Perylene in basin sediments off southern California. Geochim. Cosmochim. Acta 31, 1201-1209. Otsuki A. and Hanya T. (1972) Production of dissolved organic matter from dead green algal cells I. Aerobic microbial decomposition. Limnol. Oceanogr. 17, 248-257. Perry G. J., Volkman J. K., Johns R. B. and Bavor H. J. Jr (1979) Fatty acids of bacterial origin in comtemporary marine sediments. Geochim. Cosmochim. Acta 43, 1715-1725. Prahl F. G. and Carpenter R. (1979) The role of zooplankton fecal pellets in the sedimentation of polycyclic aromatic hydrocarbons in Dabob Bay, Washington. Geochim. Cosmochim. Acta 43, 1959-1972. Simoneit B. R. T. and Mazurek M. A. (1982) Organic matter of the troposphere--II. Natural background of biogenic lipid matter in aerosols over the the rural western United States. Atmos. Environ. 16, 2139-2159. Stodola F. H., Deinema M. H. and Spencer J. F. T. (1967) Extracellular lipids of yeasts. Bacteriol. Rev. 31, 194-213. Tanoue E. and Handa N. (1980) Vertical transport of organic materials in the northern North Pacific as determined by sediment trap experiment. Part I. Fatty acid composition. J. Oceanogr. Soc. Japan 36, 231-245. Turner J. T. and Ferrante J. G. (1979) Zooplankton fecal pellets in aquatic ecosystems. Bioscience 29, 670--677. Van Vleet E. S. and Quinn J. G. (1979) Early diagenesis of fatty acids and isoprenoid alcohols in estuarine and coastal sediments. Geochim. Cosmochim. Acta 43, 289-303. Volkman J. K., Johns R. B., Gillan F. T., Perry G. J. and Bavor H. L Jr (1980) Microbial lipids of an intertidal sediment--I. Fatty acids and hydrocarbons. Geochim. Cosmochim. Acta 44, 1133-1143. Wefer G., Suess E, Balzer W., Liebezeit G., Miiller P. J., Ungerer C. A. and Zenk W. (1982) Fluxes of biogenic components from sediment trap deployment in circumpolar waters of the Drake Passage. Nature 299, 145-147. Wakeham S. G., Farrington J. W., Gagosian R. B., Lee C., DcKaar H., Nigrelli G. E., Tripp B. W., Smith S. O. and Frew N. M. (1980a) Organic matter fluxes from sediment traps in the equatorial Atlantic Ocean. Nature 286, 798-800. Wakeham S. G., Schaffner C. and Giger W. (1980b) Polycyclic aromatic hydrocarbons in Recent lake sediments---It. Compounds derived from biogenic precursors during early diagenesis. Geochim. Cosmochim. Acta 44, 415-429. Wakeham S. G. (1982) Organic matter from a sediment trap experiment in the equatorial North Atlantic: wax esters, steryl esters, triacylglycerols and alkyldiacylglycerols. Geochim. Cosmochim. Acta 46, 2239-2257.