Particle size, light scattering and composition of suspended particulate matter in the North Atlantic

Deep-Sea Research, Vol. 34. No, 8. pp. 13(11 132% 1987. Printed in Grcal Britain.

{)198 0149/87 $3.111}+ ILll0 © 1987 Pergamon Journals l.td

Particle size, light scattering and c o m p o s i t i o n o f s u s p e n d e d particulate m a t t e r in the N o r t h Atlantic M. J. RICHARDSON* (Received 30 June 1986; in revised form 26 November 1986; accepted 16 December 1986) Abstract--Suspended particulate matter characteristics were studied on the Iceland Rise and in the western North Atlantic southeast of New York. Comparison of suspended particulate matter (SPM) and light-scattering in the Iceland Rise area suggests a second-order response due to particle size and/or composition or a nonlinear response with increasing concentration. Correlations of light scattering to SPM concentration also differ regionally. Particle size distributions are more peaked in the nepheloid layer than those in clear water. The nepheloid layer samples have a mean modal size between 3 and 9 lam, interpreted as being primarily due to resuspension and advection of sediment into the region in the bottom boundary layer. Apparent densities are relatively high, though density differences I;etwcen SPM in clear water and the nepheloid layer are not distinguishable in the Iceland Rise area. Apparent densities increase in the nepheloid layer in the western North Atlantic where 1.1 g cm 3 adequately separates clear water from nepheloid layer samples. Compared to clear water, the nepheloid layer in both regions includes lower percentages of small coccoliths and increased clays and mineral matter. These compositional variations are more dramatic in the western North Atlantic region, due to dissolution of carbonate at the seafloor, and later resuspension of the clays and mineral matter into the nepheloid layer.

INTRODUCTION

THE distribution and redistribution of marine particulate matter are of great importance m interpreting the biological, chemical and geological processes acting in the deep sea. Particulate matter from the productive surface layer descends through the water column slowly by individual particle settling and rapidly by fecal pellet or marine snow transport, providing organic-rich food to the benthic communities. Absorption of dissolved species and scavenging of chemical elements by the particulate matter may affect the distribution of chemical tracers and radioisotopes throughout the water column. Ultimately, the material in flux to the seafloor may become part of the sedimentary record. Presently, a primary source of particulate matter to the deep ocean is the surface waters. Rivers and, in cold climates, glaciers, deliver high suspended loads of terrigenous material as well as dissolved nutrients necessary for the proliferation of plankton. Phytoplankton and zooplankton dominate the particles present in the surface waters. Their skeletons and pellets sink, comprising a large fraction of the material in transit through the water column. Other components, such as volcanic and wind-blown detritus, are regionally important constituents of the suspended matter. * Texas A&M University, College Station, TX 77843, U . S . A .

1301

302

M.J. RI(HARI)SON LIGHT- S C A T T E R I N G

#/£D

LOG i ....

l

04

f

o, 8

[

/ •~

2

I ~

/

OOC

&

.... .

CLEAR W~TER ~MINIMUM

b. NE PHELO~O

2000

\\\\\\\\\\\\X\',

Fig. 1. Idealized nephelomcter proliles lrom the L-DGO/Thorndike deep-sea nephelometcr Cross-hatched area represents the near-bottom nepheloid layer, t.lnits are a ratio of scattered light to direcl lighl.

Particles settling from the surface waters are constantly subjected to the physicai. biological and chemical processes of aggregation, disaggregation, decomposition and dissolution. These processes are responsible for changing the characteristics of the particulate matter during descent through tile water column to the seafloor. Suspended material is also derived from the seafloor. Only infrequently does freshh deposited material become permanently incorporated into the sedimentary record at it,, first site of deposition. More commonly, animals living in and on the seafloor reproccs> and transport the sediments through feeding and bioturbation. Deep-sea currents can ai~ resuspend and transport the surface sediments. Both of these mechanisms, animals and currents, reintroduce previously deposited material into the water column. This recur rent resuspension can result in vast redistribution of material. Due to dissolution, decomposition and consumption in the water column, there i:, generally a decrease in the concentration of suspended particulate matter (SPM) with depth. The depth of minimum concentration of SPM is termed the "'clear-water minimum" (BIscAYE and ErrrREIM, 1977: Fig. 1). Below this level, there is often ~m increase in the concentration of SPM (the near-bottom nepheloid layer) due to resuspen-~ sion of sediment from the seafloor (HEEzEN et al., 1966; BETZER and PH SON, 197t: EITFREIM and EWlNG, 1972; FEELY, 1975; BREWER et al., 1976; BISCAYE and EITTR~:I,V, 1977). By sampling in clear water and the nepheloid layer below, the increasing infuence of resuspended sediment can be examined. I,IKGIONAL

SETTINGS

The principal study region for this work is the Iceland Rise. By comparing suspended particulate matter on the Iceland Rise with the continental rise southeast of New York site-specific results can be differentiated from wide-spread phenomena.

Particulate matter in the North Atlantic

1303

Iceland Rise

The south Iceland Rise (Fig. 2) is directly in the path of the Norwegian Sea Overflow Water. This water flows southwestward through the region directed upslope from the contours of the East Katla Ridge by 30 ° to 45 ° (Fig. 3; STEELEet al., 1962; WORTttINGTON, 1970; SHOR et al., 1977). Significant transport of Norwegian Sea Overflow Water, with current velocities on the order of 20-30 cm s-~ (SHoR, 1979), has been documented through this region. The detailed study area, the Katla Ridge province (Fig. 3; MALMBERG, 1974) in the south Iceland Rise, consists of two sedimentary ridges, the East and West Katla Ridges (SHoR et al., 1977). The East Katla Ridge, striking NNE, has smooth parallel contours along its eastern flank from 1200-2100 m. Along this flank, the overflow water flows as a geostrophic current, roughly parallel to the regional contours (STEELE et al., 1962). East Katla R!dge is separated from West Katla Ridge by a narrow canyon. The West Katla Ridge, oriented approximately north-south, has a wide, blunt nose and is separated from a large sediment drift to the southwest by a second canyon. The south Iceland Rise is a region where both a surface source and a resuspended input of particulate matter are expected. Biogenic material in the surface waters and terrigenous material from volcanic and glaciofluvial sources are probable primary particulates (i.e. those of surface origin). These sources are likely to have peak inputs during the spring and summer months, during the period of sampling for this study. A large input of resuspended material is probably related to the strong flow of Norwegian

80 °

6 0 '='

40 °

20 °

Fig. 2. Generalized western boundary, bottom-water circulation in the North Atlantic. Study areas south of Iceland and southeast of New York, denoted by the box and triangle, respectively, are located in the path of a strong bottom current.

1304

M.J. RICIIARDSON

62"N

6t*N

19"W

t8 °

o

CURRENT

•

CTD/NEPHEL/HYDRO STATIONS

METERS

17 °

16ow

CURRENT

VELOCITY

=; : ; ; o ~o ,o

Fig. 3. Bathymetry of the Iceland Rise study area. Contours arc m corrected meters, (:urrcnt measurements were made akmg thc northern transect across the eastern flank of the East Katla Ridge. Current velocities show westward to southwestward flow with an upslope component. The southern transect of stations extends from the nose of the West Katla Ridge into thc Iceland Ra~in~

Sea Overflow Water through the region. The bottom current may be sporadic (STEELt~ ~'~ al., 1962), but it has been observed in both winter and s u m m e r (CREASE, 1965, L~:~ and ELt£rr, 1965; SHOR, 1979). A mean velocity in the current core > 2 0 cm s-~ was recorded during this study (SHOR, 1979).

Western North Atlantic The area chosen for comparative study of the characteristics of particulate matter was the continental rise southeast of New York in the western North Atlantic (Fig. 2). Hydrographically, this region is influenced by the deep northeastward-flowing Gull Stream G y r e System (WORTIIINGTON,1976; LUYTEN, 1977; LAINE and HOLLISTER, 1981 ). and the southwestward-flowing Western Boundary Undercurrent (Fig. 4; HOLLISTr.'R, 1967; Z~MMERMAN, 1971; R I C H A R D S O N , 1977). Through geologic time, deep c o n t o u r following currents have shaped and developed the continental rise of eastern North America (HEEzEN et al,, 1966; FIELD and PILKEY, 1971; EGTREIM and EwIN~, 1972: HOt,LBTER and HEEZEN, 1972). This continental boundary is dissected by many canyons which may have disgorged vast amounts of terrigenous debris into the deep sea during

1305

P a r t i c u l a t e m a t t e r in t h e N o r t h A t l a n t i c

45 °

40 °

55 °

50 °

80 °

75 °

70 °

65 °

60 °

55 °

25 ° 50 °

Fig. 4. L o c a t i o n s o f h y d r o g r a p h i c s t a t i o n s in t h e w e s t e r n N o r t h A t l a n t i c . S t a t i o n s w e r e t a k e n f r o m the b a s e o f t h e c o n t i n e n t a l s l o p e to the l o w e r c o n t i n e n t a l rise. A r r o w s i n d i c a t e t h e a b y s s a l flow o f the W e s t e r n B o u n d a r y U n d e r c u r r e n t a n d t h e G u l f S t r e a m G y r e .

lower stands of sea level; presently, however, most of the sediment transported by rivers is trapped in estuaries and does not escape the continental shelf (MEADE, 1972). Some terrigenous material may be supplied at mid-water depths by horizontal transport from the continental slope (DRAKE et al., 1972; PIERCE, 1976; GARDNER, 1987). The introduction of biogenic material from the surface waters decreases rapidly from the highly productive, nutrient-rich slope waters to the unproductive Sargasso Sea. Particle input to the water column from the seafloor by resuspension and advection of sediments in this region, is possible from either the Gulf Stream Gyre System or the Western Boundary Undercurrent.

METHODS

A total of 32 hydrographic stations and 24 nephelometer stations were occupied on the Iceland Rise during June and July 1977 (Fig. 3), and seven hydrographic stations were occupied in the western North Atlantic along a transect southeast of New York in the summer of 1976 (Fig. 4). Each hydrographic station employed 6-12 5 or 301 Niskin bottles for collection of SPM. Samples were collected from the lower portion of the water column from both clear water and the near-bottom nepheloid layer. Reversing thermometers were used to obtain accurate records of sample depth in mid-water; a pinger was used to detemine the height of the bottom bottles above the seafloor. Simultaneous nephelometer profiles using an L-DGO-Thorndike nephelometer (THORNDIKE, 1975) were taken at most Iceland Rise stations to ensure that samples were from a known position relative to the nepheloid layer (if present).

13()6

M.J. R[c,~+u~SON

Concentration c~/ S P M A 250 ml aliquot was taken from the Niskin bottles immediately on retrieval for size distribution determination of SPM. The remaining water was filtered with an in-tin,: wtcuum filtration system through a preweighed 47 mm diameter Nuclepore filter wi~h 0.4 ~m pores. Filters were washed 1() times with filtered distilled water to remove salt and reweighed upon return to the laboratory to determine the mass of particles. For the Iceland Rise. water beneath the spigot was filtered onto a separate filter to collect the "dregs" (GARDNER, 1977). For the western North Atlantic, the entire bottle was filtcr~:,d onto one filter by tilting the bottle to remove all the water. Concentrations reported a,, +corrected +concentration include the "dregs" material. Combined errors in filter weighing,, and volume readings amount to ~ 9% for concentrations of 20 jag 1-j and 4--~..... ,+ 1(1() i.tg I i. Size distribution Size distributions of SPM were determined at sea with a Model T A 11 Coulter Countu~ with methods described by S~IELI~(>Nand PARSONS (1967) and modified bv Rlc't~Am~s(;', and GARDNI~R (1985). Coulter-(7~mnter dala were recorded in terms ot either tol~ti number or total w)lume of particles counted and were subdivided into logarithmicall 3 increasing size grades, each grade representing a doubling of volume. A 50 Jam apertmu and a 2.0 ml sample size were used in this study to measure the size distribution ~,t particles having equivalent spherical diameters from 1 to 20 Jam. Light scatteri~g Light scattering was measured with a L - l ) G O - T h o r n d i k e nephelometer (I'HoRNt~IKJ 1975), designed to measure relative forward light scattering from 8 ° to 24 ~. A ratio is formed between the intensity of the scattered light, E, and that of the attenuated direct light, ED, to form a scattering index, E/E¢,. Values are reported in terms of Iog(E/Ez~} Nephelometer readings in a few instances have been calibrated by sampling and filtering water for SPM from nephelometer lowerings (BEAROSI+E't'et al., 1970: BAKHa C; al.+ 1974; CARI~.R et al.. 1974; S'rERNBZR(i et al., 1974). The L-DGO/Thorndikc nephelometer has been calibrated on the Blake-Bahama Outer Ridge, the Hattera~ Abyssal Plain ( B B O R - H A P ) and the eastern North American lower continental rise (LCR) by B~sc.aY~ and EITI'Rt-IM (1974) and by GARDNER et al. (1985) on the Nov~ Scotian Rise ( H E B B L E ) and by RICHARZ)SONet al. (1987) in the Vema Channel (VC)i+~ the South Atlantic. Somewhat different calibrations for this instrument were obtained b~ these investigators. Compositional analyses Composition of the SPM was studied by photographing sections of filters with a scanning electron microscope. Five stations in the Iceland Rise area and two along the western North Atlantic transect were selected for quantification of the composition ot SPM. A sample from the nepheloid layer and one from clear water for each station were photographed for optical identification of the SPM. Four or five random photographs were taken in each of four sections of a filter to determine homogeneity of the SPM or~ the filters. Additionally, 'dregs' filters from two of the Iceland Rise samples were photographed and counted to allow a valid comparison with the western North Atlantic samples, whert~

Fig. 5. SEM photographs of suspended particulate matter. Samples la and lb are taken from clear water (a) and the nepheloid layer (b) from the western North Atlantic. Note the decrease in percentage of small coccoliths and increase in percentage of clays and mineral matter between these two samples. Samples 2a and 2b are 'dregs" filters from the Iceland Rise. Note the aggregates of coccoliths and diatoms and the large mineral grains. Approximate 10pro scale is given below each photograph.

Particulate matter in the North Atlantic

13(}9

the entire sample was collected on one filter, and to compare with the above spigot sample to examine the differentiation of the SPM while settling in the water samples (GARDNER, 1977). Twelve classes of particles were identified and counted: small coccoliths (<4 larn), large coccoliths (>4 lam), centric diatoms, pennate diatoms, dinoflagellates, organic matter, plankton fragments, fecal pellets, aggregates, clays, mineral matter, and unidentified particles (Fig. 5). Particles smaller than 2 ~tm were not included in the counts since positive identification was impossible. Aggregates are sometimes difficult to detect on filters, because drying of the samples removes water and collapses organic matter which may act as a binding agent. In clear-water samples, collapsed aggregates were readily detected because there was little material on the filters (Fig. 6). However, for nepheloid layer samples, filters were usually covered with material, in some cases more than a single layer, which made estimation of aggregates more difficult. Chi-square analysis of the compositional data was performed: (a) to determine whether the random photographs represented subsamples of a homogeneous population of material on the filters; (b) to determine whether compositional differences between clear-water samples and nepheloid-layer samples for individual stations were statistically significant; (c) to determine whether the clear-water samples from all stations were statistically different from each other and whether the nepheloid-layer samples from all stations were statistically different from each other; and (d) to determine whether collectively the clear-water samples statistically differed from nepheloid-layer samples.

RESULTS

SPM concentration

The present study of SPM across the bottom-current axis south of Iceland and along a transect across the western North Atlantic margin is an attempt to compare SPM characteristics from areas widely separated, but perhaps influenced by the same deep current system. Iceland Rise. Profiles of concentration of SPM were obtained in the lower 1000 m of the water column along two transects across the axis of the bottom current from the crests of the East and West Katla Ridges (1200 and 1500 m) into the Iceland Basin (2500 m) (Fig. 3). Spatial differences in the profiles are generally related to the water depth. In water depths >2000 m the profiles are characterized by low concentrations (=40 lag 1-~) in mid-water depths with sharp increases in concentration in the lowest hundred meters of the water column (Fig. 7). Shallower than 1800 m, concentrations in mid-water are greater (=70 pg 1-1) within thicker nepheloid layers (up to 300 m) (Fig. 7). The character of the SPM profile changed dramatically over a 10-day period when Sta. 29 was reoccupied (Stas 40 and 85). Mid-water concentrations had a range of ---4070 lag 1-1 to nepheloid-layer concentrations varying from 90 to 300 tag I-L. Western North Atlantic. A single hydrographic section was made across the continental rise southeast of New York. Nepheloid-layer corrected concentrations on the slope and upper rise did not exceed 100 lag 1-1, while above the lower rise and abyssal plain, nepheloid-layer concentrations consistently exceeded 100 lag 1-L (Fig. 8). Mid-water maxima in particle concentrations in slope and upper rise waters are interpreted as being caused by material advected horizontally outward from the slope.

13 ]()

M . ,I. Ri(tiARD,',;CIN

ICELAND RISE S OUTH_ERN SECTION

ICELAND RISE NORTHERN SECTION

o] /. ]

SUSPENDED PARTICULATE CONCENTRATION (pgll)

,ooo£

•

SUSPENDED P A R T ICLJ[ A7 CONCENTRATION (#g, IZ

i

~5OO2000-

SUSPENDED PARTICULATE CORRECTED CONCENTRATION (~g/i)

-•

SUSPENDED PART CULAi{ CORRECTED CONCEN~q&],ON

~c

500 T,

,oooL ~

5OO 2000

OIST/INCE {kilomelers)

Fig. 7. Cross-sections of suspended particulate [nailer for the Iceland Rise region. Dols indlcatc sample locations. Corrected concentrations include the 'dregs' (GARDNER, 1977). Highest

concentrations in the near-bottom w;iter occur in the bottom current axis, along the flanks of the ridges.

Light-scattering observations To obtain a continuous vertical profile of SPM, light-scattering measurements, which can be calibrated in terms of SPM concentrations, were made on most of the hydro,graphic stations in the Iceland Rise region (Fig. 9). The nephelometer profiles exhibit features similar to those of the SPM. Lowest clear-water values (0.34 log E/E1)) occurred over the basin; highest near-bottom values (1.43-1.55 log E/EzJ occurred along the ridge flank and in water depths > 1800 in. The character of the nephelometer profiles changed substantially from the crest of the ridge eastward into the basin. The ridge-crest profiles showed variable light scattering throughout the water column, generally in excess of 0.5 log E/Er~. At one station on the ridge crest, no clearly defined nepheloid layer was observed. To the east, along the ridge flank (1600-2200 m) nepheloid layers were up to 700 m thick whereas clear-water values decreased below those from the ridge crest, The nepheloid layers in the deepest stations in the basin were thin (<100 m), but showed a very sharp gradient from clear water, Profiles taken within the canyon between the two ridges are somewhat erratic. A general increase in light scattering from mid-water to the seafloor is observed, but the

1311

Particulate matter in the North Atlantic

STATION

NUMBERS

K58 D 58

710

715

743

718

730

1000

30 2000

-

7~1

• *

A (/) (E tlJ I--

<~0~

"r 3000 to c~

4000

SUSPENDED

5000

PARTICU LATI"

MATTER

P-9 / ~!

Fig. 8. Cross-section of suspended particulate matter along the continental rise southeast of New York. Concentrations are corrected for "dregs'. Note the occurrence of some high concentrations in mid-water which may reflect advection o1' material from the slope. Modified from GARDNER (1985).

nepheloid layer is not as well-defined as along the ridge flanks (Fig. 9). Light scattering is fairly high (>0.6 log E/Eo) for these profiles throughout the entire water column. The two reoccupations of Sta. 28 (2000 m) show that the nepheloid layer varies from 400 to 700 m in thickness and is also highly variable over the time scale of 10 days (Fig. 9).

Size distribution analyses Particle size distributions are most often displayed in terms of number, volume, or weight of particles per logarithmic size grade. Since this work deals with volume or

u~ ~i

i

z 0

1

I--

o

z >-

~L

"~f~,, ......

o

i ?i •

I

~Z

z f rr w

,~

I--

':'

/cu2

c

',.

"ot~ d

< \ ".'~'k

HJd3(7

\

:

7~

.:/,~

~ ~o

1


:f

p~ u

,,,

-~

Z

ILl F--

rr~ 0

<0

o o

o o

~

1313

Particulate matter in the North Atlantic

weight concentrations rather than numbers of particles, all of the data will be displayed in terms of volumetric measurements. Iceland Rise. Two stations (25 and 39) demonstrate the differences in the percentvolume size distributions between clear water and the nepheloid layer. Station 25 is located along the northern section in the Iceland Rise region at 1804 m, in the axis of the bottom current. A nephelometer profile obtained at the same site (Fig. 9) was compared with SPM concentrations from the water bottles verifying which bottles were within the nepheloid layer. The size distribution of particles at 1780 m, within the nepheloid layer, shows a peaked distribution with a modal spherical diameter of 5.6 ~tm (Fig. 10). In clear water at 790 m at the same station we observed a more typical fiat distribution with equal volumes of material from 1 to 20 ~tm. Similar features are seen at Sta. 39 at 2163 m, farther down the ridge flank (Fig. 10). Within the nepheloid layer, at 2145 m, a peaked distribution is observed, similar to the nepheloid-layer particle size distribution at Sta. 25, with a coarser modal size (=9 Urn). In the clear water, at 1205 m, there is an irregular, fiat distribution, similar to Sta. 25 (Fig. 10). Western North Atlantic. Size distributions also were compared for clear and nepheloidlayer waters in the western North Atlantic. Since nephelometer profiles were not available at the hydrographic stations, the nepheloid layer was judged to begin at the sharp increase in concentration in the near-bottom waters. The general results of the comparison made between clear-water and nepheloid-layer samples are similar to those from the Iceland Rise. The nepheloid-layer samples (e,g. Sta. 718, 4462 m) have peaked distributions with the modal size at 3 Urn, whereas clear-water sample distributions have flat distributions (Fig. 11). The nepheloid-layer 'dregs' samples exhibit the peaked-type distribution even more strongly than do the standard nepheloid-layer samples (Fig. 11). STATION 25 t804m

STATION 67 4522m

STATION 39 2463m

7OO m 20

20

2O

I

I

4

16

I

I

I

I

4

I

1 I

I

t6

I

L 4

I

I 16

,o

0

I

I

I

t

4

DIAMETER

i

I

16

(,urn)

I

O

I

I

I

I

4

i

I

16

D I A M E T E R (,urn)

i

0

I

I

4

DIAMETER

16

(,urn)

Fig, 10. Coulter Counter volumetric size histograms from Stas 25, 39 and 67 in the Iceland rise region. See Fig. 9 for the location of the clear-water and nepheloid-layer samples with respect to the light-scattering profile. Note the lower variance of the nepheloid-layer sample. Station 67 does not show a well-defined nepheloid layer.

1314

M.J. RI('HARDSON

STATION 7 t 8 4485m t453

rn

CLEAR WATER 20

l I

I

4462

I 4

l

I 16

m

-4

NEPHELOID LAYER 20

-4 I,~

~0

% 0

!

I

i

20

0

I

I

4

I

16

4462 m NEPHELOID LAYER "DREGS"

I

I

I 4

DIAMETER

i

I

I

16

(,urn)

Fig. 1l. Coulter Coulltcr volumetric siz,., hi,~tograms Iron] Sta. 718 in the' wc~tcrn Norlh Atlantic. Samples are from clear water, the nepheloid layer and the 'dregs" water from a nepheloid-layer sample. Note the lower variance of the nepheloid-layer standard and 'dregs" samplcs.

Compositional studies Particles on SEM photomicrographs were counted in this study to determine the difference in composition between the clear-water material and the nephetoid-laycr material. Iceland Rise. Five stations were selected for compositional analysis in the Iceland Rise region as being representative of the various hydrographic conditions (25, 31, 39, 67 and 76; see Fig. 3). Station 25 is in the axis of the bottom current; Sta. 39 is off-axis along the same transect. Stations 31 and 76 are in the interior of the basin. Station 67 is located on the nose of the West Katla Ridge and exhibits no nepheloid layer. Two samples were taken from each station, one in clear water at mid-depths and one in the nepheloid layer or within 50 m of the seafloor where a nepheloid layer was not present. The results of the particle counts (Fig. 12, Table 1) for these stations show that by number, small coccoliths (Fig. 6) are the largest component of the samples (up to 62% i

Particulate matter in the North Atlantic

COMPOSITIONAL VARIABILITY OF S U S P E N D E D ICELAND RISE

60

1315

PARTICULATE

O

o • ~ *

o

I

MATTER

Clear Nepheloid Station 67 Dregs

k~

40

I o*

eo* 20

O*

O*

~:

o-

O*

o~o*

~o, ,

+,

O0

• "~.. O*

O* 0

i

i

i

~.~

COMPONEN TS

Fig. 12. Compositional variability of suspended particulate~natter from the Iceland Rise. Counts were made of the different components from scanmng~lectron m~croscope photomncrographs and normalized to total number of particles counted. Identifiable fragments were included m the counts. Greatest differences between clear-water and nepheloid-layer samples are in small coccoliths, pennate diatoms, and mineral matter.

pennate diatoms, mineral matter, and clays follow in order as the next largest fractions of the samples. The results of the chi-square analysis (statistics after CRow et al., 1960) on the compositional data (Table 2) indicate that at the 95% confidence level, the random photomicrographs of the filters are representative subsamples of a homogeneous population of material. This result allows summing the counts from many photomicrographs to make additional comparisons. The second chi-square test found (at the 95% confidence level) that the clear-water and nepheloid-layer samples from individual stations are statistically different. The differences observed are that nepheloid-layer samples have fewer small coccoliths and more pennate diatioms, clays, and mineral matter than do clear-water samples (Fig. 12). These same differences are observed whether or not particles from the 'dregs' water are included. The 'dregs', however, do have fewer small coccoliths, less organic matter and mineral matter, and more large coccoliths, centric diatoms, and pennate diatoms. The smaller percentage of mineral matter in the 'dregs' samples is difficult to understand. Perhaps the mineral matter is associated with organic aggregates of low density. Another chi-square test showed (at the 95% confidence level) that the clear-water samples were statistically different from each other and the nepheloid-layer samples were statistically different from each other. Perhaps this variability between stations is related to the proximity to Iceland and resuspension of mineral matter by the bottom current. The two stations in the basin generally have less mineral matter and fewer clays and aggregates than the station along the ridge flanks.

~

~

~

~'

~

~,~

E

U

Table 2.

Chi-square analysis of compositional data from the Iceland Rise and western North Atlantic Clear (C) Nepheloid (N)

Station-bottle

Tests for homogeneity of filter samples 25-4 C 25-1 N 39-B1 C 39-B5 N 32-4 C 32-1 N 67-B6 C 67-B1 C 76-B6 C 76-B1 N Tests for standard vs dregs samples 25-4 C 25-1 N 39-1 C 39-5 N Tests for clear-water vs nepheloid-layer samples 25 C/N 39 C/N 32 C/N 67 C/C 76 C/N Same including dregs 25 C/N 39 C/N 718 C/N 734 C/N Tests for all clear-water samples All clear C Tests for all nepheloid-layer samples All nepheloid N

Degrees of freedom

Calculated chi-square

95% Confidence interval

27 27 27 27 27 30 27 27 24 30

35.16 39.83 25.83 23.98 31.95 28.93 19.29 41.30" 25.50 14.15

40.11 40.11 40.11 40.11 40.11 43.77 40.11 4(I.11 36.42 43.77

10 11 10 10

32.70* 70,03* 103,12" 73,96*

18.31 19.68 18.31 18.31

11 10 10 9 10

38.62* 50.16" 131.90" 16.52 24.07*

19.68 18.31 18.31 16.92 18.31

11 10 8 8

48.34* 64.36* 37.46* 145.50"

19.68 19.68 15.51 15.5l

45

197.28"

61.63

33

326.15"

45.69

* Significant at 95% confidence interval. COMPOSITIONAL

OF SUSPENDED P A R T I C U L A T E WESTERN N O R T H A T L A N T I C

VARIABILITY

MATTER

60o Clear • Nepheloid

40-

2O 8

• •

6 • 0

i

i

i

i

o

o

°I°

8

o

i

i

i

i

COMPONENTS Fig. 13. Compositional variability of suspended particulate matter from the western North Atlantic. Note the pronounced differences in percentages of small cocoliths, clays and mineral matter between clear water and the nepheloid layer.

1318

M. J, RKHARDSON

Western North Atlantic. The dominant components in the western North Atlantic samples are small coccoliths, clays, and mineral matter (Fig. 13; Table 1). The variations seen between clear-water and nepheloid-layer samples are also most pronounced fl~r these components. The percentage of small coccoliths drops sharply from clear-water ~o nepheloid-layer samples, whereas the percentage of clay and mineral matter rise dramatically (Fig. 13). A chi-square test verified the clear-water and nepheloid-layt~r samples are statistically different at the 95% confidence level (Table 2), DIS(

( SSI()N

Correlation of light-scattering measurements with SPM concentrations A correlation of light-scattering measurements with concentrations of SPM has been used for several years to calibrate L-DGO nephelometer profiles in quantitative terms (BISCAYE and E~VrREIM, 1974, 1977). In this study a comparison between clear-water and nepheloid-layer samples in the Iceland Rise area was made as well as a regional comparison between the Biscaye and Eittreim curves and the combined Iceland Rise data.

•

i

i

///

/

20 !

~

ioyer -...j.....~ o

£.5 ~ [.0 %

•

. / ~ 5

i

, /"

•

oQ

.o"//-

•

•

/d

©

,

(o .,4

o •

I

0,5 l

Clear Nepheloid

i

L

i

t

J

0,5

__

I

1,0 LO~

1

1

J

~,5

E/Q

Fig. 14. C o r r e l a t i o n o f light scattering and standard concentration lk~r the I c e l a n d Rise are'~,. T h e dashed line is the regression line for the clear-water s a m p l e s . T h e dashed curves arc the 9 5 % conlidence limits for clear-water concentration v a l u e s . T h e solid line is the r e g r e s s i o n line for the nepheloid-layer samples. T h e solid c u r v e s a r e the s a m e as for the c l e a r - w a t e r samplc~,. T h e regression lines for clear water and the nepheloid layer arc statistically different at the 9 5 % confidence level,

Particulate matter in the North Atlantic

1319

Analysis of the regression lines for clear-water and nepheloid-layer data shows a statistically significant difference at the 95% confidence level (Fig. 14). This can be interpreted in two ways: either the nephelometer responds differently to the particles in the nepheloid layer vs those in clear water, or the relationship between light scattering and SPM concentration is nonlinear. Theory suggests nephelometers respond to particle characteristics such as size and composition in addition to concentration (JERLOV, 1968). It has been shown here that the size and composition of suspended material differ between clear-water and nepheloid-layer samples in the Iceland Rise region. Both of these factors are likely to influence the response of the nephelometer; however, it is not yet possible to detemine the magnitudes of their influence. The considerable scatter about the regression lines for the Iceland Rise data may originate from the basic difference in the volume of water examined by each method. The SPM concentration data were obtained with 5 or 30 1 Niskin bottles, which sample over a vertical distance of 0.5-1 m. The nephelometer, on the other hand, measures light scattering that is integrated over a vertical distance of 25 m. Moreover, in a region of active currents like the Iceland Rise, temporal variability and patchiness in the concentration of particles in near-bottom water could cause differences between the point values of the Niskin bottles and the integrated values measured by the nephelometer 0.5 h later.

3"51 | [ 3.01

Y=0.82x+ 0.81 IC Y=O.88x+ 0.96IC* Y=l.09x+ 0.35HEBBLE* Y=1.24x+ 0.07HEBBLE

Iz°f

/

i= } 8

1.o1 l"

/

HEBBLE /HEBBLE"

/II

Ic L ~ ('~HEBBL / I e

vc-

Y=1.19x+ 0.13BBOR-HAP

0'.5

1'.0

,'.5

2'.0

2'.5

Log E/Eo Fig. 15. Correlations of light scattering and concentration for the Iceland Rise (IC), the Nova Scotian Rise (Hebble) (GARDNER et al., 1985), Vema Channel in the South Atlantic (VC) (RK'IIARDSON et al., 1987) and the western North Athmtic ( B B O R - H A P and LCR) (Bl,~cA~ I and EITTREIM, 1977). Asterisk indicates concentration including 'dregs' material.

M. ,l, RI( IIARDSON

1320

l"lt[)[~" 3.

C('cIM.vqttllr(" r('gr~:~io~l t'qtldliO~t.~ for /i~hl-.~cdllcrill,~ t',', ('Otl('Ctlll'tlliOH t OlYdtlliOH~

BBOR-HAP (from BIS('AYE and ErI-IRFIM. 1974) LCR (lrom BIS('AY! and F, IFIRFIM, 1977) ('lear water

IogY=

l . l O l o g . V ~ 0 13(r

!)~Jl)

k)g.k ~ 0 5 ( ) ( t

u:~4i

log}'-1.0

log Y = 0.63 log X ~ ().t>4 ( r , ,S), = O. 106 log Y = 0 . 3 6 l o g X ,- 1.35 (r .S), - 0.055 I o g Y - 0 . 6 5 1 o g A " ~.0.9~(r

(this study) Nepheloid layer (this study) All data (this study)

~).~.2i i) ",21 !~70i

Y = concentration (~g I ~): ,V = light-scattering (t:/Et,): r = correlation coefficient: S;, :::: standard error o! the regression coefficient, b. Test for similarity between clear walcr and nephcloid layer lines th b, O,(G- 0.36 ~- . . . . . . . . . . . . . . . . . . . . . . . 51~, t~. 0.0124

21.77.

Reject the hypothesis at the 5% signilicance level that the two regression lines arc the s a m e Statistics after CROW et at. (196(!L

Although the Biscaye and Eittreim curves (Fig. 15) yielded good approximations oi SPM concentration, their calibration curves differ from those of this data set. suggesting that calibrations differ regionally. Additional comparisons can be made using data from the Nova Scotian Rise ( H E B B L E ) (GARDNt~R et al., 1985) and the Vema Channel in the South Atlantic (VC) (RI('ItARI)SON et al., 1987) (Fig. 15). The H E B B L E curve, using concentrations uncorrected for 'dregs" (GARDNER, 1977), is very similar to the B B O R ..... H A P curve of BIS('AYt:: and ErrrRHM (1977). However, in comparing the calibration curves it must be remembered that some comparisons were made with concentration data corrected for 'dregs" and others were not. The different calibration curves may be due to a nonlinear response of the nephciometer, or to a difference in the nature of SPM in the several regions. A comparison of the SPM off Iceland to that along the continental rise off New York showed that more clays and mineral matter occur at the latter location.

Size distribution attalyses Particle size distributions l'rom clear water and the nepheloid layer irom south ot Iceland were compared to those from the western North Atlantic to determine whelher variation in the size spectra were site specific. Volumetric histograms. Histograms of normalized particle volume vs equivalent spherical diameter depict the variations in the distributions between clear water (flat) and the nepheloid layer (peaked) (Figs 10 and 11). These variations in particle size distribt|tions between clear water and the nepheloid layer are noted in both areas and may be due to either differences in the state of aggregation or the composition of the suspended material. Filters of SPM were examined to determine which alternative was more likeb. The percentage of aggregates which have survived the sampling processes does not change substantially between clear water and the nepheloid layer for either region, but the composition of the SPM does change. Both regions show an increase in percent and total concentration of the clays and mineral matter in the nepheloid layer compared to clear water.

Particulate matter in the North Atlantic

1321

¼

-..I 10 ~j

io 2 aI

o

i

i 4

t6

0

'I

4 DIAMETER

Cumulntlve

slope N

(,urn/

~0,000

t000

- 3 0

!6

LU

jSIop

eN-25

~j

128

Dif ferentlOl

s~ope: 3.0

I b

_ Dif ferenh(~l overoll slope =

3.0

o.... !,C',

4~_/Cumulative overall slope =

~00

t000

3E~

s ope

*0

\ "\\I

LU Sl°pe~-4 25

~1~

\

~9

2b

tc t

4

DIAMETER

¢6

(.urn)

I

Io

,

,

DIAMETER

,

(,urn)

,6

2c

~oo

I

I

I

4

DIAMETER

I

~6

gum)

Fig. 16. Particle size distributions and associated parameters. Graph la represents a particle size distribution with equal volumes of material in logarithmically increasing size grades• Graph lb shows the associated differential and cumulative number distributions. Graph lc shows the associated volume distribution normalized by the size grade interval. Graph 2a illustrates an idealized nepheloid layer, volumetric distribution of lower variance. Graph 2b, similar to lb, demonstrates the variability of slope through the distribution, but overall having slopes very similar to those in lb. Graph 2c, similar to lc, shows both positive and negative slopes through the distribution, with the overall slope similar to lc. Nepheloid-layer particle sizes are always different from those in clear water. In the few stations where nepheloid layers are absent, the size distributions in near-bottom waters retain a clear-water characteristic, a flat distribution with high variance (Fig. 10). This suggests that nepheloid layers are sites of introduction into the water column of material with a unique modal size. As resuspended material increases, particle distributions b e c o m e more peaked. Differential volume distributions. Marine particle size distributions are usually expressed as the slope of the cumulative number distribution (BADER, 1970; CAaDER et al., 1971; BmJN-COaq:AN, 1971; SHELDON et al., 1972; WELLERSHAUSet al., 1973; McCAvE, 1975; BRUN-COaq'AN, 1976; LERMAN et al., 1977; McCAvE, 1983). A slope of minus three is indicative of a distribution with equal volumes of material in logarithmically increasing size grades. In this study, h o w e v e r , mass and v o l u m e distributions are more useful. For this reason, the normalized differential v o l u m e (volume of material in each size range normalized to 100%) of particles is used in interpretation rather than the cumulative number distribution. For comparison with other studies, a slope of minus one for normalized differential v o l u m e distribution is equivalent to a slope of minus three for a cumulative number distribution (Fig. 16).

1322

M.J. RI(IIARDSON

Slopes from the normalized differential volume distributions were plotted vs depth for all stations from the Iceland Rise region (Fig. 17). Although the nepheloid-layer and clear-water points were differentiated on the plot, the data are too scattered to reveal any depth-dependent or clear-water vs nepheloid-layer distinctions. Examining the slopes of individual profiles (Fig. 18) shows that clear-water distributions are well -200 (:bear

• NepheloJd

-~so i r

-too, +

.~,.;+

:,oLk..+,.

. ,~°,°

,

,

• : : c';fo'

-05

0 ]

[

-

£ .

.

.

.

500

.

:

~

[

t000 t500 DE/:'TH {m )

5

_L

J

2000

2500

Scatter plot of the slopes ol the curves of the normalized differential particle volume vs particle diameter, with depth for the Iceland Rise region. Slopes of two idealized distrihutinns arc shown in Fig. 16.

Fig. 17.

STA 6

F

| 790

25

STA

CLEAR

39

WATER 6

m

-

,1205 m

/

5

4

0 IhO

0 5C

! :)2'

50

0 O0

NEPHELOID

!

'(

LAYER

+7 8 0 m

b.

400

0,50

2145 m

5

3

"~.

t

J 2 L__L 0 O0

J

-J.--

~ . - - - - L . - -

l

LOG

~

2

DIAMETER

I

-__

I

I

_

~

~__J.__~._+

(,am)

Fig. 18. Individual profiles of normalized differential particle volume vs particle diameter for the clear-water and nepheloid-layer samples plotted in Fig. 10. Clear-water samples are seen to be well represented by a single line through the data. Nepheloid-layer samples are better represented as a two-slope distribution.

Particulate matter in the North Atlantic

1323

represented by single lines having slopes of approximately -0.9. Nepheloid-layer distributions as single-line representations have slopes similar to those of clear water, but much more scatter. The similarity in slopes between clear-water and the nepheloid-layer samples may be explained by idealized distributions (Fig. 16). Both idealized clear water (high variance, flat distributions) and idealized nepheloid layer (low variance, peaked distributions) have the same differential volume slopes of minus one. However, the nepheloid-layer distributions are much better expressed as two-slope distributions with a slope break at the modal point of the particle volume histograms (Fig. 18). This break in a slope has previously been reported for shallow-water data (<200 m) (BADER, 1970; BRuN-COTTAN, 1971; MCCAVE, 1975) and is presumably because of biological production (MCCAvE, 1984). For deep water, the two-slope distribution is interpreted to be due to an introduction of resuspended clays and mineral matter into the nepheloid layer. MCCAVE (1983) found similar results for particle size spectra on the Nova Scotian continental rise, whereas RICHARDSONand GARDNER(1985) found that the particle size spectra in the nepheloid layer on the Nova Scotian Rise was a signature of water mass. Apparent density One of the characteristics of suspended particles most difficult to determine is their in situ bulk density. However, measurements related to the density of SPM may be useful in providing insight into the differences between clear-water and nepheloid-layer samples. Organic-rich matter has a density close to that of seawater, whereas individual mineral grains have densities exceeding 2.5 g cm -3. However, few marine particles are composed solely of either organic matter or mineral grains. Organic-rich matter, fecal pellets and aggregates usually contain varying percentages of plankton skeletons, organic matter and clays. Contents of organic matter in SPM drop substantially from 25 to 50% in the water column to <2% in surface sediments (HEATH et al., 1977; BAKERet al., 1979; Rowe and GARDNER, 1979). This difference may be useful in identifying resuspended material, which would tend to have a lower organic-matter content and therefore a higher density. An apparent density of the ratio between mass and volume concentrations can be calculated. The mass concentration is determined by filtration and the volume concentration, by Coulter-Counter analysis. Filtration of SPM gives a measurement of dry weight of particles per volume of seawater. The Coulter-Counter gives a measurement of wet volume of particles per volume of seawater. The ratio of the two gives a dry weight of material per wet volume of material. Iceland Rise. A plot of mass concentration vs wet particle volume (Fig. 19) reveals no distinction between the apparent density of the nepheloid-layer and clear-water samples, except that nepheloid-layer samples plot in the upper portion of the graph due to their higher concentrations. The 52 clear-water samples give a mean apparent density of 2.00 + 1.39 g ,can-3, whereas the nepheloid-layer samples yield a statistically indistinguishable 2.38 + 1.23 g cm -3. Western North Atlantic. A similar comparison for the western North Atlantic continental rise shows that the nepheloid-layer particles (1.34 + 0.46 g cm -3) have an apparent density greater than that of clear-water particles (0.83 + 0.40 g cm -3) (Fig, 20). Why different apparent densities characterize clear water vs nepheloid water in the western North Atlantic and not south of Iceland probably lies in the composition of the

1324

M.J. RICItARI)SON

APPARENT

DENSITY

- ICELAND

RISE

i

Q

•

•

•

4

% 20o

% o

J

o

o•

i

•

o o~a

:i_

;

Nephelold

::

~ _

_ k ......

~

.....

i

_._ ~ _ _ A _

80

4r

WET"

VOLUME

.....

!20

a _ _ _ _ _ _1J t60

2OO

(ppb)

Fig. 19. Apparent density ol samples from (he Iceland Rise. No differentiation by apparent density is evident between clear-water and nepheloid-layer samples,

particulate matter. Both regions have statistically significant differences in the composition of SPM between clear-water and nepheloid-layer samples. However, the differences for the western North Atlantic samples are much more pronounced (Table I ) Nepheloid-layer samples show a dramatic 10-43% decrease in coccoliths, a 1(~-23"', increase in mineral matter, and a 0-15% increase in the clays and mineral matter compared to clear water. The increase in the clays and mineral matter serves to increase the apparent density of these particles with respect to clear-water particles. Composit ional differences for the Iceland Rise samples are not sufficient to reflect differences in the apparent densities.

Compositional differences Compositional differences and state of aggregation were examined to determine the likely cause of the differences in particle size distributions. Iceland Rise. Compositional studies indicate that at the 95% confidence level particles from clear water differ from those in the nepheloid layer at individual hydrographic

1325

Particulate matter in the North Atlantic

APPARENT DENSITY

WESTERN NORTH A T L A N T I C

20O 180

~'

'~60

~

t40

/ .:...~Slope/=

t20

1.1 g / c m 3

,oo I 60

""

40

°

°

°

°

o Clear • Nepheloid 0

20

40

60

I I 80 400 {20 t40 WET VOLUME (PPB )

I

160

I

I

480 200

Fig. 20. Apparent density of samples from the western North Atlantic. Clear-watcr samples have apparent densities smaller than those for nephcloid-layer samples. A line with slope 1.1 g cm 3 effectively separates the two groups.

stations. To distinguish whether these differences are due to processes causing nepheloid layers, or are evidence of dissolution, decomposition and consumption of particles with increasing water depth, Sta. 67 (Fig. 9) with no nepheloid layer was sampled at 700 and 1479 m where the total water depth is 1522 m. These two samples (at the 95% confidence level) are from the same population. These data support the hypothesis that the variability between clear-water and nepheloid-layer samples is due to the existence of the nepheloid layer, and is not simply a function of depth or depth-related processes. For comparative purposes, the percentages, rather than numbers, of particles in the different compositional classes were used to eliminate the effects of concentration differences between clear water and the nepheloid layer (Fig. 12). The major variations seen in composition were a decrease in the percentage of small coccoliths and an increase in the percentage of diatoms (particularly Rhizosolenia), clays and mineral matter in the nepheloid layer relative to clear water. The increase in these percentages are probably caused by resuspension of local surface sediments into the nepheloid layer. The decrease in the percentage (but not number, Table 1, Fig. 12) of small coccoliths reflects dilution. Said another way, the other components of the SPM in the nepheloid layer are preferentially added, resulting in the decrease in percentage of small coccoliths. However, percentages of aggregates were similar in both the clear-water and nepheloid-layer samples (Table 1) suggesting that aggregates may have been underestimated in the nepheloid-layer samples due to higher concentrations than clear-water samples. These data suggest that compositional differences between clear water and the nepheloid layer, particularly the increases in diatoms and mineral matter, may account for the differences in particle size distribution. Aggregation of particles in the nepheloid layer may be of secondary importance as LAMBERTet al. (1981) did find a chemically distinct form of aggregate particles characteristic of the bottom nepheloid layer.

1326

M.J. R,']~AJ~DS¢~N

The causes of aggregate formation in the nepheloid layer can bc assessed on theoretical grounds. Inorganic flocculation occurs by Brownian motion, local shear :~lld differential settling. EfYSl"HN and KRONJ (1962) report equations to calculate the probability of successful collisions by particles in salt water resulting in flocculation, f:r~,n their equations, and using Iceland Rise data. collisions by Brownian motion would occur once every 6 months to 2 years; collisions by local shear would occur once every 2~-~i days; and differential settling may be important for particles larger than 5.4 pm which comprise <1% by number of the suspended particles. These calculations indicate ~h~:~t inorganic flocculation is not likely to be responsible for aggregate formation in lhe nepheloid layer on the Iceland Rise. However, organic aggregation by benthic organisms may be important within the water column. Western North Atlantic. The more pronounced compositional differences between particles in clear water and the nepheloid layer in the western North Atlantic ~ e probably due to the different inputs and processes. At present, the Iceland Rise region has a high surface input of both terrigenous and biogenic components, whereas the western North Atlantic has a smaller terrigenous input. For the western North Atlantic. the clear-water samples are mostly composed of biogenic material; the ncphetoid-laycl~ samples contain a larger fraction of clays and mineral matter (Fig. 13). Changes between clear water and the nepheloid layer include a decrease in percentage of small coccoliths and organic matter and an increase in mineral matter, clays and aggregates. Water depths are much greater in the western North Atlantic than in the Iceland Rise area. In these increased depths carbonatc dissolves at the seafloor (T,\k.xJL.\s~H. Ic~7~: B~OECKER and TAKAnASm, 1978). The coccoliths observed in the nepheloid-laycv samples are frequently partially dissolved. Dissolution at the seafloor and subsequent resuspension certainly could account for the dramatic decrease in coccoliths from cle;ll~ water to the nepheloid layer. The increase in clays and mineral matter in the nepheloid layer is partially due to the decomposition and dissolution of other components at the seafloor. However, advection of nonbiogenic material into the region in the bottcm~ boundary layer is another probable cause. Since current velocities were not measured during this study, assessing the probability of local sediment input is difficult. Composition and state of aggregation change from clear-water to nepheloid-laye~ samples. Both of these may cause the differences observed in the particle size distribw tions. Mineral grains and clays increase substantially in the nepheloid layer. These components usually range in size from 3 to 8 ~tm and therefore may account for the pe~Jk of material in the nepheloid-layer particle size distributions (Fig. 11). Aggregate~ observed tended to be larger (>10 ~tm) and therefore are not as likely candidates f~, explaining the size changes observed. ('ON('I.

USIONS

Analysis of light scattering vs SPM concentration from the Iceland Rise gives three principal results: (a) clear-water and nepheloid-layer samples show different relationships of concentration to light scattering, indicating that although the first-order response of the L-DGO nephelometer is a function of particle concentration, second-order responses due to other SPM characteristics such as particle size and composition are not negligible, and/or the response is nonlinear with increasing concentration; (b) compari,

Particulate matter in the North Atlantic

1327

son of the correlation obtained south of Iceland to the Blake-Bahama Outer Ridge/ Hatteras Abyssal Plain ( B B O R - H A P ) , Lower Continental Rise (LCR), High Energy Benthic Boundary Layer Experiment (HEBBLE) and Vema Channel (VC) regression lines (B1scAYEand EITITREIM,1974, 1977; GARDNERet al., 1985; RICHARDSONet al., 1987) demonstrates that the correlation is somewhat site-specific. Clear-water and nepheloid-layer particles have different size distributions. Clear-water samples are characterized by flat high variance distributions of roughly equal volumes of material in logarithmically increasing size grades. Nepheloid-layer samples have peaked lower variance distributions with a mean modal size between 3 and 9 p.m. Expressed in terms of normalized differential volume curves, nepheloid layers are shown to have a two-slope distribution. The variations in particle size observed between clear-water and nepheloid-layer samples are interpreted as being due primarily to resuspension of sediment into the nepheloid layer and advection of material into the region in the bottom boundary layer. Determinations of apparent density from the Iceland Rise area show no differentiation between clear-water and nepheloid-layer samples. However, plots of apparent density from the western North Atlantic do show a marked increase for nepheloid-layer samples. The compositional differences for the western North Atlantic clear-water and nepheloid~ layer samples are much greater than for the Iceland Rise samples. The dramatic decrease in coccoliths and increase in mineral matter and clays between clear water and the nepheloid layer serves to increase the apparent density of these particles with respect to clear-water particles. Dissolution of the carbonate is more likely here in the deeper waters. Samples from the nepheloid layer statistically differ from clear-water samples. For the Iceland Rise, compositional differences include a decrease in the percentage of small coccoliths and an increase in the percentage of diatoms, clays and mineral matter in the nepheloid layer relative to clear water. These differences are interpreted as being due to dilution of the coccoliths by other components (diatoms, clays and mineral matter) which are readily resuspended into the nepheloid layer from the local surface sediments. For the western North Atlantic, the major compositional changes between clear water and the nepheloid layer include a decrease in small coccoliths and organic matter and an increase in mineral matter, clays, and aggregates. These differences are due to dissolution of the carbonate and decomposition and consumption of the organic matter on the seafloor with subsequent resuspension and advection of the refractory material into the near-bottom boundary layer. These compositional changes rather than increases in aggregates are probably responsible for the changes in the particle size distributions between clear water and the nepheloid layer in both regions. Theoretical assessments of inorganic flocculation by Brownian motion, local shear and differential settling in the nepheloid layer suggest that inorganic flocculation is not likely to be responsible for aggregate formation whereas organic aggregation by benthic organisms within the water column may be important. Acknowledgements--I am

most grateful to C, D. Hollister, and several commentors, P. E. Biscaye, W. D. Gardner, S. Honjo, J. D. Milliman, A. N. Shor, J. B. Southard and D. W. Spcncer. This work was supported by ONR through Contracts N00014-79-C-00-71 NR 083-004, N00014-74-C0262 NR 083-004 and N00014-75-C-0291 and E R D A through Contracts 13-7923 and 13-2559.

1328

M..1. RI,, ilAI>.i)St)N

I~,E FE P,l~ N ('l{ S BAI)I{R H. (19711) The hyperbolic distribution ot particle sizes. Journal olGeophysical Research, 75, 2822-2830 BAKt;R E . T . , R. W. STERNBERG and D. A. McM.+',NUS (1974) Continuous light-scattering profiles and suspended mattcr over Nitmat deep sca fan. In: Suspended solids in water, R. J. G1BI:JS. editor, Plenum, New York, pp. 155-172. BAKER E. T., R. A. FI IIA and K. l .kK),HASIII (I979) (heroical composition, sizc distribuliu~ and p:II'IICIC morphology of suspended particulate matter at D O M E S sites A, B, and C: Relationships with toczd sediment composition. In: Marine .geology and oceanography of the Paclfic Manganese Nodule Pr<~l,m,,+ J. L. BISCUOFF-and D. Z. PU,EI~, editors, Plenum, New York, pp. 163-201. BEARI)SI,EY O. F., H. PAK, K. CARI'JI'I~ and B. LUNI)GRIiN (197tl) Light scattering and suspcndcd particlc~, m the eastern Equatorial Pacilic Ocean..lourmd ~1' (;e~,l~lly~ical Research, 75, 2837-2845. BETZER P. R. and M. E. Q. PILSON (1971) Particulate inon and the nepheloid layer in the westcrn Not;i~ Atlantic, Caribbean and Gulf of Mcxico. Deep-Sea Research, 18, 753-761. BIS('AYE P. E. and S. L. EITTREIM (1974) Variations in benthic boundary layer phenomena; nepheloid layers m the North American Basin, In: Suspended solkL~ i;; water. R. J. GIBBS, editor. Plcmim, New Yolk+ pp. 227-260. Blsca'.+t! P. E. and S+ L. EITI RI..[M ( 19771 Suspended particulate loads and transports in the nephctoid la}
,$'cie#lce Letler~, 32, 393-4il3 BROECKER W. S+ and T. TAKAttASHI (1978)The relationship bctween [ysoclme depth and in ~itl~carbonate i{~'+ concentration. Deep-Sea Research. 25, 65-95. BI,tI+N-C()UIAN J.-C. (1971) Etudc de la granulometric tics particules marines mist_ire-; cliccruccs i \ c c tl!~

compteur Coulter. Cahiers Oceanographiques, 23, (~5-95. BRuN-Co'ITAN J.-C. (19761 Stokes settling and dissolution role model for marine particles as a tunction ol size distribution. ,lournal of Geophysical Research, 81, 1601- 16116, CARDER K. L., G. F. Bf ARDSLI "~, Jr and H. PAK (1971) Particle size distributions in thc Eastern 15,quat<~U:~! Pacific. Journal +~1Geophysical Research, 76. 50711-51177. CAtm~ R K. L., P. R. BF~'rZt¢R and D. ~ E(iOIMANN (1974) Physical, chcmical and opticat measures ',,+ suspended particulate concentrations: their intercomparison and application to the West African shell h~ Suspended solids in water, R. ,1. Gil+t~s, editor, Plenum, New York, pp+ 173-193. CRliASE J. (19651 The flow of Norwegian Sea water through the Faroe Bank Channel. Deel)-Sc,~ Researct~, 12 143-151/. CROW E. L., F. A. DAVIS and M. W. MAXHI.I l) (1961}) StalLs'tic* Manual---with examples taken p+om ordnam+ developmem. Dover Publications, Inc., New Y o r k , ]g8 pp. DRAKI: D+ E+, R. L. KOLPAK and P. J. PlS('tlt:i4 (19721 Sediment transport on the Santa Barbara--Oxnard Shelf, Santa Barbara Channel, California. In: Shelfsedimem transport, D. J. P. SwIvr, D. I-1. l)t!,\'~l am! O. H. PII,KE',, editors, Dowden, Hutchinson and Ross. Stroudsberg, PA, pp. 307-331

EINSTEIN H. A. and R. B. KRONE (1962) Experiments to determine modes of cohesive sedimcnt transport m salt water. Journal of Geophysical Research, 67, 1451-1461. E1TI'REIM S. L and M. EWINO (1072) Suspended particulate matter in the deep waters of the North Americat~ Basin. In: Studie~ in physical oceano<~,raphy, A. 1+. GORDiaN, editor. Gordon and Breach, l+olldoi~. pp. 123-167. FEInt v R. A. (1975) Major-clement composition of the particulate matter m the ncar-bottom ncpheloid lave~ <, the Gulf of Mexico. Marine Chemisto', 3, t21-156. FIF-I D M. E. and O. H. PILKE'~ ( 1971 ) Deposition of deep-sea sands: comparison of two areas ol the Carolim~ continental rise. Journal r~lSedimenta# T Petrology. 41,526-536. GAttDNER W. D. 11977) Incomplete extraction of rapidly settling particles from water samplers l+t#,lttoh)~,,!' and Oceanography, 22,764-768. (}ARI)NI'R W. D. 11987) Baltimore ('anv,,)ll as a conduit of sedinlcnt to the deep sea. I)c~7>3'ca Rewari particles in the northwest Atlantic. Marine Geology, 65, 199-242. HEATH G. R., T. C. MOORE and J. P. DAt i'IIIN (1977) Organic carbon in deep-sea sediments. In; 7"heyase +,i fossil fuel COe in the oceans, N. R. ANDERSON and A. MALAHOFF, editors, Plenum, New York, pp. 605-626. HEEZEN B. C., C. D. Hol,kls-rt~R and W. F. RUDi)IMAN (1966) Shaping of the continental rise by deep geostrophic contour currents. Science, 152, 502-508. HC.LLISTER C. D. 11967) Sediment distribution and deep circulation in the western North Atlantic. Ph+l) Thesis, Columbia University. New York, 471 pp. (unpublished manuscript).

Particulate matter in the North Atlantic

1329

HOLLISTER C. D. and B. C. HEEZEN (1972) Geologic effects of ocean bottom currents, western North Atlantic. In: Studies in physical oceanography, A. L. GORDON, editor, Gordon and Breach, London, pp. 37-66. JERLOV N. G. (1968) Ootical oceanography. Elsevier, Amsterdam, 194 pp. LAINE E. P. and C. D. HOLLISTER (19811 Geological effects of the Gulf Stream System on the Northern Bermuda Rise. Marine Geology, 39, 277-310. LAMBILRT C. E., C. JEIIANNO, N. SII.VI~RIn!R(i, J. C. BRuN-C~)rrAN and R. Clll!SSlil lq (1981) Log-norm~l distributions of suspended particles in the open ocean. Journal oJ" Marine Research, 39, 77 98. LEE A. and D. ELLETF (1965) On the contribution of overflow water from the Norwegian Sea to the hydrographic structure of the North Atlantic Ocean. Deep-Sea Research, 12, 129-142. LERMAN A., K. L. CARDER and P. R. BETZER (1977) Elimination of fine suspcnsoids in the oceanic water column. Earth and Planetary Science Letters, 37, 61-711. LLJYTEN J. R. (1977) Scales of motion in the deep Gulf Stream and across the continental rise. Journal ¢~f Marine Research, 35, 49-74. MALMBERG S. -A. (19741 A note on the deep water south of lceland--"Overflow '73." ICES Reference C . M 1974/C:32 (unpublished manuscript). McCAvE I. N. (19751 Vertical flux of particles in the ocean. Deep-Sea Research, 22,491-502. MCCAVE I. N. (1983) Particulate size spectra, behavior, and origin of nepheloid layers over the Nowt Scotian continental rise. Journal of Geophysical Research, 88, 7647-7666. McCAVE I. N, (19841 Size spectra and aggregation of suspended particles in the deep ocean. Deep-Sea Research, 3,329-352. MEADE R. H. (1972) Transport and deposition of sediments in estuaries. Geological Society of America Memoirs, 133, 91-120. PIERCE J. W. (19761 Suspended sediment transport at the shelf break and over the outer margin. In: Marine sediment transport and environmental management, D. J. STANLEYand D. J. P. SWIFT, John Wiley, New York, pp. 437-458. RK;HARDSON M. J. and W. D. GARDNER (19851 Analysis of suspended-particle-size distributions over the Nova Scotian continental rise. Marine Geology, 66, 189-203. RICHARDSON M. J., P. E. BISCAYE, W. D. GARDNER and N. G. HOGC; (1987) Suspended particulate matter transport through thc Vema Channel. Marine Geology, in press. RICHARDSON P. L. (19771 On the crossover between the Gulf Stream and the Western Boundary Undercurrent. Deep-Sea Research, 24, 13%159. ROWE G. T. and W. D. GARDNER (19791 Sedimentation rates in the slope water of the northwest Atlantic Ocean measured directly with sediment traps. Journal of Marine Research, 37, 581-600. SHELDON R. W. and T. R. PARSONS (1967) A practical manual on the use of the Coulter Counter in marine science. Coulter Electronics, Toronto, 66 pp. SHELDON R. W., A. PRAKASHand W. H. SUTCHFFE, Jr (1972) The size distribution of particles in the ocean. Limnology and Oceanography, 17,327-3411. SHOR A. N. (1979) Bottom currents and abyssal sedimentation processes south of Iceland. Ph.D. Thesis, M,I.T./W.H.O.I. Joint Program in Oceanography, 246 pp. (unpublished manuscript). SHOR A. N., D. MULLER and D. A. JOHNSON (1977) Transport of Norwegian Sea Overflow--preliminary results of ATLANTIS II cruise 94, June-July 1977. ICES--CM 1977/C:44 Hydro. Comm. (unpublished manuscript). STEELE J. H., J. R. BARRETT and L. V. WORTHINGTON (1962) Deep currents south of Iceland. Deep-Sea Research, 9,465-474, STERNBIRG R. W., E. T. BAKER, D. A. M('MANtlS, S. SMITII and D. R. MOI~RISON (19741 An integrating nephelometer for measuring suspended sediment concentrations in the deep sea. Deep-Sea Research, 21, 887-892. TAKAHASHIT. (1975) Carbonate chemistry of sea water and the calcite compensation depth in the oceans. In: Dissolution of deep-sea carbonates, W. V. SLITER, A. W. H. BE and W. H. BERGER, editors, Cushman Foundation Foraminiferal Research Special Publication No. 13, pp. 11-26. THORNDIKE E. M. (1975) A deep-sea photographic nephelometer. Ocean Engineering, 3, 1-15. WELLERSHAUSS., L. GOKE and P. FRANK(1973) Size distribution of suspended particles in sea water. Meteor Forschungsergebenisse. Series D, No. 16, pp. 1-16. WORTHINGTON L. V. (1970) The Norwegian Sea as a Mediterranean Basin. Deep-Sea Research, 17, 77-84. WORTHINGTON L. V. (1976) On the North Atlantic circulation. The Johns Hopkins University Press, Baltimore, MD, 110 pp. ZIMMERMAN H. B. (1971) Bottom currcnts on the New England continental rise. Jourltal c~/"Geophy.~ical Research, 76, 5865-5876.