Nutrient dynamics in Amazon shelf waters: results from AMASSEDS

~)

Pergamon

ContinentalShelf Research, Vol. 16, No. 3, pp. 263--289, 1996 Copyright© 1995 Elsevier ScienceLtd Printed in Great Britain. All rights reserved 0278-4343/96 $9.50 + 0.00

0278--4343(95)00008-9

Nutrient dynamics in Amazon shelf waters: results from AMASSEDS DAVID J. DEMASTER* and ROBERT H. POPE* (Received 7 July 1993; accepted 24 February 1994)

Abstract--Four hydrographic cruises were conducted on the Amazon shelf as part of the AMASSEDS field program. During each cruise, approximately 55 stations were occupied and nutrients, as well as other hydrographic parameters, were measured. The results of this time series sampling program indicate that the nutrient concentrations in the riverine end-member (silicate = 144/~mol kg -1 , phosphate = 0.7/~mol kg -1, nitrate = 16/~mol kg-1 , ammonium = 0.4/~mol kg- l , and urea = 0.9/~mol kg-1) remain relatively constant, despite a two-fold seasonal variation in river water discharge rate. Of the major nutrients (nitrate, phosphate, ammonium and silicate), nitrate shows the greatest seasonal change in riverine end-member concentration with a high value (23 ~mol kg -1) during the March cruise (rising river discharge) and a low value (12/~mol kg-1 ) during the November cruise (falling river discharge). Nitrate is the dominant nutrient form of inorganic nitrogen throughout most of the river/ocean mixing zone, however, in the outershelf area, where nitrate has been depleted by biological production, this nutrient occurs at concentrations comparable to the other nitrogen species (ammonium, nitrite and urea), which are at levels < 1/~mol kg-1 . Nearshore, high turbidity inhibits phytoplankton production because of light limitation, whereas on the outershelf, nitrate appears to be limiting growth more than silicate or phosphate. Nutrient uptake was observed during all four cruises, however, nearly all of this production must be regenerated in shelf bottom waters, because very little of the biogenic materials are buried in the seabed (silicate burial <4% of flux to algal blooms; ~ 10% burial of biologically available inorganic nitrogen reaching the river/ocean mixing zone; and <3% burial of phosphate flux to shelf environment). Clearly the Amazon shelf is not an efficient nutrient trap. Initial estimates of primary production on the Amazon shelf suggest that algal blooms are sustained by regeneration to a large extent (up to 83%, 69% and 59% for N, P and Si, respectively) as well as by riverine and upwelled sources. Nutrient budget calculations have been used to establish the dominant external source of nutrients to the algal blooms occurring on the outer shelf. Based on flux core measurements, diffusive nutrient fluxes from Amazon shelf sediments are very low relative to riverine supply rates (silicate flux out = 1.3% of riverine flux, the nitrate plus ammonium flux is essentially zero, and the phosphate seabed flux shows removal of - 2 % of thc riverine flux). Inventories of naturally occurring 21°pb were used to estimate the onshore flow of subsurface water onto the Amazon shelf. The radiochemical data indicate that the flux of water onto the shelf may be as much as five to ten times greater than the annual flow of the Amazon River. The nutrient flux from this shoreward movement of ocean water (originating at a depth of 60-100 m water depth) accounts for about 80% of the externally supplied ammonium, 52% of the externally supplied phosphate, 38% of the externally supplied nitrate, and 17% of the externally supplied silicate reaching the outer shelf, with the remainder of the nutrient fluxes coming from the river. Therefore, the outershelf algal blooms are supported to a significant extent by the shoreward flux of nutrients from offshore, subsurface waters.

*Department of Marine, Earth and Atmospheric Sciences, North Carolina State University, Raleigh, NC 27695-8208, U.S.A. 263

264

D.J. DeMasterand R. H. Pope INTRODUCTION

Estuarine environments clearly are dynamic areas of enhanced biological productivity and nutrient recycling (Kennish, 1990). These environments receive nutrient-rich riverine waters, which stimulate the production and growth of marine phytoplankton. Commonly, the presumption is made that the blooms of algal material occurring near the mouths of rivers are controlled solely by the flux of nutrients coming down the dispersal system. For some of the large rivers of the world, however, shelf circulation also must be considered as an important source of nutrients (van Bennekom and Tijssen, 1976). The Amazon River is the world's largest river in terms of water discharge as well as nutrient flux (OItman, 1968; van Bennekom el al., 1978; Edmond et al., 1985). Large algal blooms, primarily composed of diatoms (Hulburt and Corwin, 1969), occur on the Amazon shelf seaward of the high turbidity zone (Milliman and Boyle, 1975; Edmond et al., 1981; DeMaster et al., 1983). These blooms can be detected readily using satellite imagery (Curtin and Legeckis, 1986; Muller-Karger et al., 1988) as well as direct chemical measurements of surface waters (DeMaster et al., 1983, 1986). The source of the nutrients sustaining these algal blooms has been debated for over two decades. Ryther et al. (1967) reported that the influence of the Amazon River was to lower the phosphate and nitrate concentrations on the shelf and consequently, decrease the fertility of the area. They also noted that the shoreward advection of subsurface water could be a significant source of nutrients. More recent work (Gibbs, 1972; Cadee, 1975; Milliman and Boyle, 1975; Edmond et al., 1981) has shown that the Amazon River is very important to the nutrient budgets of the shelf, but little quantitative evidence has been available regarding the relative supplies of nutrients from riverine and advective sources. This paper examines nutrient dynamics at the mouth of the Amazon River. The research is part of AMASSEDS (A Multi-disciplinary Amazon Shelf SEDiment Study), which is an inter-disciplinary study of oceanographic processes occurring on the Amazon shelf. By measuring the radiochemical inventory of naturally occurring 21°pb in Amazon shelf sediments, estimates can be made for the magnitude of the shoreward flow of water from offshore. This information coupled with nutrient measurements on the shelf are used to estimate the relative importance of riverine and offshore nutrient sources to the study area. In addition, this research addresses three other important aspects of Amazon shelf nutrient dynamics. Previous investigators of the Amazon River/ocean mixing zone examined only two forms of dissolved nitrogenous species, i.e. nitrate and nitrite. Our field programs measured these species as well as ammonium and urea to better define inorganic nitrogen systematics on the shelf. Secondly, most previous studies have been "single-sampling" field programs, and therefore, cannot address the issue of temporal variability in nutrient systematics on the shelf or in nutrient concentrations of the riverine end member. As part of the two-year AMASSEDS field program, nutrient dynamics were characterized on the shelf during four different stages of the river discharge (falling, rising, high and low discharge). Lastly, this research program examines the contribution of Amazon shelf sediments to the nutrient budgets of the study area. Resuspension by strong tidal currents generates high suspended solid concentrations over much of the shelf and liberates nutrients built up in near-surface sediments.

Nutrient dynamicsin Amazon shelfwaters

265

THE STUDY A R E A The Amazon River water discharge varies seasonally by a factor of two with peak flow occurring in June and minimum flow occurring in November. Along with the water discharge (mean value of 1.8 x 105 m 3 s -1) comes 1.2 x 109 tons of sediment (Meade et al., 1985) and 2-3 x 108 tons of dissolved components (Gibbs, 1972). Approximately half of the sediment discharged by the river can be accounted for in the prograding Amazon delta; the other half presumably is transported to the north or accretes in coastal deposits (Kuehl et al., 1986; Nittrouer et al., 1996). The bulk chemical composition of Amazon River water is controlled by the varied lithologies within the drainage basin as well as by diverse weathering regimes (Stallard and Edmond, 1983). Fluctuations in the flux of dissolved components within the river are in sync with the water discharge cycle (Gibbs, 1972), whereas the maximum sediment discharge period occurs in April as riverine sediments are mobilized by the rising river (Meade et al., 1985). Circulation on the Amazon shelf is a result of the complex interaction of river discharge, strong tidal currents, winds stress, and the along shore flow of the North Brazilian Current (Nittrouer et al., 1991). Gibbs (1982) emphasized the importance of shoreward flowing bottom waters on the shelf that result from the offshore movement of freshwater at the surface and the entrainment of seawater from below. A detailed discussion of Amazon shelf hydrographic data can be found in Curtin (1986a,b). Because of the enormous flow of freshwater from the river, all of the mixing between river and ocean waters occurs out on the Amazon shelf. The intensity of the trade winds changes on a seasonal basis for the study area (Picaut et al., 1985). In January and February (during low river discharge), winds are strongest (1.1 dynes cm -2) and generally blow in the onshore direction (toward the southwest), whereas in June and July (during high river discharge) the winds are weakest (0.2 dynes cm-2), blowing toward the west. The magnitude and nature of the North Brazilian Current also changes seasonally. The highest flow ( - 3 0 Sv) occurs during August (falling river discharge) and the lowest flow ( - 1 0 Sv) during April (rising river discharge; Philander and Pacanowski, 1986). Between January and June the North Brazilian Current moves northward along the coast of South America, whereas between June and December the current retroflects eastward at about 5°N (Muller-Karger et al., 1988). Two of the early systematic studies of nutrients and other dissolved components in Amazon River waters were conducted by Williams (1968) and Gibbs (1972). Soon after these studies, research concerning silicate systematics on the Amazon shelf was reported by Milliman and Boyle (1975), who noted that the uptake of this nutrient occurred in the shelf waters seaward of the high turbidity zone. Edmond et al. (1981) presented a very comprehensive study of nutrient systematics (nitrate, nitrite, silicate, phosphate, oxygen, pH and alkalinity) on the Amazon shelf. These authors concluded that regeneration of estuarine biogenic material essentially was complete for phosphorus, at least 50% complete for nitrate (ammonium was not measured), and only 20% complete for silicate. Later work by DeMaster et al. (1983, 1986) measured biogenic silica accumulation rates in shelf sediments and concluded that more than 96% of the dissolved silicate carried by the Amazon River makes it to the open ocean (despite the fact that as much as 50% of the riverine silicate can be removed initially by diatom blooms on the shelf). Most recently, Fox etal. (1986) and Fox (1989) and Berner and Rao (1994) have examined the systematics of phosphate in the Amazon and several other river systems. These investigators conclude

266

D.J. DeMaster and R. H. Pope

that as much as 50% of the phosphate released to the ocean from the Amazon River is the result of desorption from suspended sediments and that equilibration between dissolved and particulate phases (primarily a ferric phosphate/ferric hydroxide mineral) controls the phosphate concentration in low salinity shelf waters. FIELD P R O G R A M AND ANALYTICAL METHODS Four hydrographic cruises to the Amazon shelf were conducted aboard the R.V. Iselin during the two-year AMASSEDS field program. The cruises occurred during August 1989 (Cruise I-l, falling river discharge), March 1990 (Cruise II-3, rising river discharge), May 1990 (Cruise III-3, high river discharge), and November 1991 (Cruise IV-3, low river discharge). Approximately 55 stations were occupied during each cruise along seven shore-perpendicular transects (Fig. 1). Salinity was measured using a Neil Brown Instrument System model MKIII CTD fish as the primary profiling instrument (Limeburner and Beardsley, 1989). The CTD was equipped with a Sea Tech 5-cm pathlength transmissometer and a Sea Tech fluorometer. Samples for water-column measurements were collected using a General Oceanic rosette sampler with 12 10-1Niskin bottles. Immediately after collection, the water from the Niskin bottles was sampled for gases and stable isotopes and then the remaining sample was transferred to a large carboy. The carboy was shaken prior to dispensing water samples for nutrients, suspended solid concentration, POC and PON. 0.45 micron Nuclepore filters and plastic filter holders were used to separate the dissolved and particulate fractions for all of the cruises. After filtering, a Technicon II Autoanalyzer was used to determine the dissolved silicate, nitrate, nitrite, phosphate, ammonium and urea concentrations (Armstrong et al., 1967; Biggs et al., 1982; Parsons et al., 1984a; Price and Harrison, 1987). Salinity corrections were made to the absorption/concentration relationships for the waters throughout the river/ocean mixing zone. The analytical sensitivity for the nutrients using the Technicon II Autoanalyzer were: silicate---0.5/~mol kg-l; nitrate--0.05 ¢tmol kg-1; phosphate--0.05 #tool kg 1; ammonium---0.1 ktmol kg- 1; nitrite--0.02 ~tmol kg 1; and urea---0.05/~mol kg- 1. Based on duplicate analyses during the field program, the precision of nutrient measurements was on the order of 2-5%. RESULTS AND DISCUSSION Riverine end-member nutrient concentrations

The nutrient concentrations for the riverine end-member are summarized in Table 1 for the four AMASSEDS cruises. During Cruise II, the ship was not allowed to move up the river channel to obtain a good fresh water end-member sample. In this case the low salinity (5-10 psu) concentrations were extrapolated back to 0 psu to establish the riverine endmember. Of the four major nutrients (nitrate, phosphate, ammonium and silicate), the riverine end-member concentrations from the four AMASSEDS cruises varied temporally by as much as 100% (for nitrate) to as little as 5% (for silicate). The relatively small seasonal change in most of these nutrients illustrates how the drainage basin can buffer changes in concentration despite a two-fold variation in the water-discharge rate of the river (Gibbs, 1972). This trend is consistent with the results of Lesack (1993) whose research on the central Amazon Basin indicates the dominance of base-flow discharge

Nutrient dynamics in Amazon shelf waters n'W

111"

R"

O"

52"W

!rl'

SO"

4V"

Jl " W

]1 "

9 "

th "

41"

47"

'16"

45"

47"

~"

~'W

~ "

~1"

N"

267 O"

a"

47"

~7"

Sl'W

SI'

SO"

49"

a"

51" W

Sl "

SO '

49 *

48 "

47 "

~2 " w

51 '

50 '

4~"

a '

47'

~"

6 *

~'No

ll'.

4"°

3'-

2"~

1'-

e'-

1""

z's. 52 • w

$1"

$4"

49"

41"

47"

~ •

46"

Fig. 1. AMASSEDS station locations from the four hydrographic cruises to the Amazon shelf'.

relative to storm flow. During A M A S S E D S , the highest concentrations for nitrate and silicate occurred during the rising river discharge cruise (March 1990), when the river started to flow over the dried banks deposited during the previous high discharge period. The nutrient data presented in Table 1 indicate that nitrate is the dominant form of

268

D.J. DeMaster and R. H. Pope Table 1.

Nutrient concentrations in the Amazon freshwater end-member during A M A S S E D S

Sampling period cruise number river discharge Silicate ~mol kg -a) Phosphate ~ m o l kg -1) Nitrate ~ m o l kg -1) Ammonium ~ m o l kg -1) Nitrite ~ m o l kg -1) Urea (~mol kg -l)

8/89 I Falling

3/90 II Rising

5/90 II1 High

11/91 IV Low

Weighted mean

144 0.7 12 0.4 0.1 0.4

149 0.7 23 0.4 0.2 0.8

141 0.6 13 0.4 0.7 1.2

141 0.8 18 0.5 0.1 0.8

144 0.7 16 0.4 0.3 0.8

inorganic nitrogen (available as a nutrient) at the Amazon River mouth. Riverine nitrate concentrations (12-23/~mol kg -1) exceed the ammonium concentrations (0.3--0.5/~mol kg-1) by a factor of 30-60. Nitrite and urea concentrations in the riverine end-member also are small relative to nitrate, with values typically comparable to those of ammonium. During Cruise III (May 1990), however, the urea concentrations in Amazon River water were as high as 1.1/~mol kg -1. DON measurements were not made as part of this study. Although all six nutrients were not measured in any of the previous estuarine studies, a comparison of some Amazon riverine end-member values still is useful. Milliman and Boyle (1975) observed a silicate concentration of 130/~mol kg -1 for the riverine endmember during a June cruise in 1974. Based on results from their May/June 1976 cruise, Edmond et al. (1981) reported riverine end-member values of 120, 10 and 0.5/~mol kg -t for silicate, nitrate and phosphate, respectively. Measurements from a low-river discharge cruise (December 1982) yielded somewhat higher riverine end-member concentrations for silicate (141/~mol kg -1) and nitrate (19/~mol kg-1), but lower for phosphate (0.03/~mol kg-1; Key et al., 1985; Edmond et al., 1985). During a low-river discharge cruise in October, 1979 as well as during a high discharge cruise in May 1983, the silicate riverine end-member was equal to 135/~mol kg -1 (DeMaster etal., 1983, 1986). With the exception of the low phosphate value in the December 1982 cruise and the low silicate concentration during the May/June 1976 field program, the riverine nutrient data from previous investigators are in reasonable agreement with our results from AMASSEDS. The importance of annual variations in the riverine end-member concentrations could not be addressed systematically, based on the literature and our data set. The most recent studies of nutrient systematics in the river basin are from the CAMREX Program (Richey et al., 1991) and Lesack (1993). Based on 19 Amazon River stations, Richey et al. (1991) reported that nitrate concentrations vary from 5 to 25 ~mol 1-I, ammonium is less than 1 or 2/~mol 1-l, and phosphate values range from 0.1 to 1.2 #mol 1-1. In a detailed study from the central Amazon Basin (near the confluence of the Amazon and Rio Negro), Lesack (1993) observed nitrate, ammonium and phosphate concentrations ranging from 1-20/~mol 1-1, 0-1.3 #mol 1-1 and 0-0.8/~mol 1-1, respectively, between February 1984 and February 1985. Lesack (1993) noted that the highest nitrate concentrations occurred during rising river discharge, whereas Richey et al. (1991) observed the highest nitrate concentrations during periods of low discharge. Given the complexity and size of the Amazon drainage basin, temporal and spatial changes in riverine nutrient concentrations are not unexpected.

269

Nutrient dynamics in Amazon shelf waters 200

A M A S S E D S I-1 A U G U S T , 1989

A

20T ]

A M A S S E D S I-1 AUGUST, 1989

L 100

i

4-' 0

•

•

0

" t',,i

.......... e. ,o?, .~lMib. 10 20 30 - 40 Salinity (PSU)

0

0

A M A S S E D S I-1 AUGUST, 1989

0.8

J

m 0

E ~= 0.6

o

i

•

O

0

'

"'°

10

20

30

1'54

I

A M A S S E D S I-1 AUGUST, 1989

•

"~ 1.0

•O•o

g

o

!ooI

"° ge.~oO

a, o

f. o.2 •

.

,

.

B

.

,

w

10

20

30

•

O0 m ' . . ~ . - : . - . . . . - .

40

0

10

Salinity (PSU) 1.2~

0.8

°•%%

,= .~ .,=,~.: =_~a)

Z~ 20

__ .,--~+'-~_ 30 40

Salinity (PSU)

A M A S S E D S I-1 AUGUST, 1989

~I.O

•

l

2"

A M A S S E D S I-1 AUGUST, 1989

I

o

•

..

E= 0 . 6 1 .

•

0.4 `•-T.~,.~I •

0

o •

•

•

o

•

< 0.2

"

•

,

o

1 "

. oo

@~ •

"

10

:3

Salinity (PSU)

30

40

O" 0

o

• •

•;,o •

~'0" °~

20

o •

°

0o oO° • °

0.0 0

°oo#,

°

•

_Q . . . . . . 0

40

•

•

~ o.4

0.0

•

Salinity (PSU)

1.0 A

ID•

0

• 10

o , o

• O

~

°~

t

I

oe 20

30

40

Salinity (PSU)

Fig. 2. Nutrient vs salinity relationships for the Amazon River/ocean mixing zone during Cruise 1 (August 1989; falling river discharge). The filled circles indicate surface water samples, whereas the empty circles represent subsurface samples. The solid lines represent the ideal mixing line between riverine and oceanic end-members. Points falling below the line indicate nutrient removal, whereas points falling above the line require nutrient addition. In the phosphate salinity diagram, there is initial phosphate removal from surface waters (0-6 psu) and then desorption increases the dissolved phosphate concentration to a maximum at - 17 psu.

Nutrient systematics on the shelf Silicate and nitrate. T h e n u t r i e n t / s a l i n i t y r e l a t i o n s h i p s f r o m A M A S S E D S C r u i s e 1 a r e s h o w n in Fig. 2 f o r s i l i c a t e , n i t r a t e , p h o s p h a t e , n i t r i t e , a m m o n i u m a n d u r e a . O f t h e f o u r m a j o r n u t r i e n t s m e a s u r e d , silicate is p r o b a b l y t h e n u t r i e n t w h o s e p r o f i l e is l e a s t a m b i g u o u s

270

D.J. DeMasterand R. H. Pope

in assessing biological uptake and nutrient recycling. Silicate is present at high concentrations over much of the shelf, does not undergo any oxidation/reduction reactions (like the nitrogen species), and is not involved to a large extent in adsorption/desorption reactions nor in inorganic precipitation reactions (like phosphate). Complementing the silicate/salinity data from Cruise I, Fig. 3 shows the silicate (and nitrate)/salinity relationships for AMASSEDS Cruises II, III and IV. The silicate data from all four cruises indicate year-round nutrient uptake in surface waters and dissolution of biogenic silica at depth. The salinity of initial silicate removal varied from 10 to 20 psu during the AMASSEDS cruises, and the envelope of silicate data suggests that phytoplankton production in surface waters utilizes from 30 to 60% of the riverine silicate flux (see Officer, 1979, for description of method). During our four Amazon shelf cruises, the greatest depletion of silicate in the water column (relative to the ideal mixing line) occurred during Cruise I (August 1989) and Cruise II (March 1990). Removal of silicate in low salinity waters is clearly inhibited by the high concentrations of lithogenic sediment (Milliman and Boyle, 1975). The percent deviations of the observed silicate and nitrate concentrations from the ideal mixing line (based on the riverine and oceanic end members) are shown in Fig. 4 for surface waters from Cruise II (the period of most intense nutrient depletion). Both nutrients behaved as conservative chemical tracers in surface waters with suspended solid concentrations greater than 5 or 10 mg 1-~. The observation that silicate removal was undetectable in the high silicate, high turbidity waters near the river mouth suggests that little, if any, inorganic precipitation of authigenic silicate minerals occurred within the river/ocean mixing zone (Mackenzie and Garrels, 1966). Control of nutrient uptake by light limitation has been described previously for the Amazon shelf (Milliman and Boyle, 1975; Edmond et al., 1981; DeMaster et al., 1986) as well as for other major estuarine systems (Cloern, 1987; Lohrenz et al., 1990; Turner et al., 1990). Recent research in San Francisco Bay has documented the effect of spring/neap tidal stirring intensity on phytoplankton bloom dynamics (Cloern, 1991). These physical processes may have similar relevance for the Amazon shelf. To examine the areal distributions of the high and low nutrient uptake regions on the Amazon shelf, the deviations of the observed silicate and nitrate concentrations from the ideal mixing line (in/~mol kg -1) were calculated for AMASSEDS Cruise II surface waters (Fig. 5). The data show clearly a large region on the outer shelf where over 30/~mol kg-~ of silicate have been removed. This same region also was characterized by large amounts of particulate biogenic silica in the surface waters (DeMaster et al., 1992). The only area that exhibited a positive silicate anomaly was on the inner shelf just north of the River Mouth Transect. These high values probably were the result of fluid mud resuspension by tidal currents (Kineke et al., 1996), which released porewaters, enhanced in nutrients, directly to the water column. Enhanced levels of nutrients also may be generated by resuspension of reactive particles followed by dissolution or by desorption from resuspended particles, which were equilibrated with the higher nutrient concentrations in porewaters (Fanning et al., 1982). The stations located on the inner Amazon shelf that exhibit the positive silicate and nitrate anomalies (Fig. 5) are the same stations that reveal the highest surface water concentrations of Ra (Moore et al., 1996) on the shelf• This correlation supports resuspension as an important mechanism for nutrient enrichment because 224Ra (3.6-d half life) primarily is produced in sediment porewaters from the decay of its parent, 22STh. Based on flux core measurements, the release of silicate and nitrate from the seabed via predominantly diffusive processes is relatively slow and •

224

Nutrient dynamics in Amazon shelf waters

!-

200 I AMASSEDS 11-3 ~CH, 1990

0

0

200,

, L

,o0 0

0

10

20

30

Salinity (PSU)

AMASSEDS Ill-3 MAY, 1990

I [

I

30

oMARCH' ~ 190

-

2o-1"-~ "°

~

lO

O

0

20

E

~ 10

20

30

=

tP'q~,

0

40

0

o0=

i

40

Salinity (PSU)

11 9 9AMASSEDS MAY' 0 ~ 111-3

Salinity (PSU) 200

°o

25 j

, [

°°

20

AMASSEDS 11-3

30e1

i I

10

40t °

40

°Oo•

271

o

'

o

10

20

I I

30

40

Salinity (PSU) 40 I

AMASSEDS IV-3 NOVEMBER, 1991

] 1 30

100'

",

AMASSEDSIV-3 NOVEMBER, 1991oo o

20-

•

m o m

o

0

10

20

30

Salinity (PSU)

40

0

10

20

30 Salinity (PSU)

Fig. 3. Silicate and nitrate concentrations as a function of salinity for A M A S S E D S Cruises 1I (March 1990; rising river discharge), III (May 1990; high river discharge), and IV (November 1991; low river discharge). The filled circles indicate surface water samples, and the empty circles represent subsurface samples. These data (plus the data shown for Cruise I in Fig 2) indicate that nutrient uptake on the Amazon shelf occurs year round. During the A M A S S E D field program, the most extensive removal of nutrients (i.e. the greatest deviation from the ideal mixing line) occurred during Cruises I and II.

40

272

D.J. DeMasterand R. H. Pope A m a s S e d s C r u i s e II, March 1990

Leg3

100 A

l-

~'_ '10

60

*"

.0

"0

40

o~

o

80

[] ,--

20

r~

0

gi

m

- 2 0 -,

~ 4oZ

E -6o-

~: -80:

mm

mm ~ [] []

[] [] •B

M

B

[]

B

m

B

B

m id~

lira

- 1 0 0 - -e"~'~ ~ . . m . . . ,

,

........ 10

100

Suspended Sediment

AmasSeds

C r u i s e II, Leg 3 March 1990

100too v ¢O

m

so:

°m

60'

~

40"

20-

>

0 -20, 0

1000

(mg/I)

o

~-40.

•--_~

.SO, m ~ -SO,

[ ] •m

,/&-.

rl

[] B

[]

[]

IJ

[] m

-1O0

I~1 la

--

B

il . . . . . . . .

i 10

.......

i

. . . . .

100

000

Suspended Sediment (mg/I) Fig. 4. Percent deviation from the ideal mixing line for silicate and nitrate in surface waters plotted as a function of suspended solid concentration for AMASSEDS Cruise II (March 1990; rising river discharge)• Both nutrients show conservative behavior in waters with suspended solid concentrations greater than 5-10 nag 1- l . Seaward of the turbid plume, light limitation decreases and both silicate and nitrate show extensive biological uptake. Stations with salinities >34 psu have not been included in this plot because the observed minus predicted nutrient concentrations are very low causing relatively large errors in the percent nutrient deviation.

ineffective (see later discussion). Figure 6 shows the silicate and nitrate anomalies as a function of depth for the Open Shelf Transect (Stas 38-47) during Cruise II. The release of silicate from innershelf sediments is evident from these data, although the station closest to the coast (Sta. 47) shows a very weak silicate regeneration signal. On the outer shelf, the uptake of silicate in the upper 15 m is readily apparent with positive silicate anomalies at depth, indicating regeneration and/or upwelling. The nitrate/salinity distribution also is indicative of nutrient uptake and regeneration, but generally the patterns are less obvious than those for silicate because of the transformations among nitrate, nitrite, urea and ammonium. Nitrate concentrations

273

Nutrient dynamics in Amazon shelf waters

5z.v~

~1-

lOOm

50

20

10

49"

aS"

47"

46"

March 1990

5Z"W Sl " s'N I . I \, 100m 8 ~

'4

4 •=

~

S'~'k '

50" 49" 48" I 1 I s AMASSEDS C R U I S E 11, Leg #3 M a r c h 1990

4".

-4

i

[

/

=

l

.$ •

"-

f 4

i"

S2 ~

st

so

4~

4~"

47

1~

Z's I = 5z" ~

"~" I 6"N

Iv (umol/k~l [" 5

Nitrate A ....

-8

47" I

i

| 51

=

I 50"

'

| ~9"

'

I 4~"

~

j 47

.

| -'s i.

Fig. 5. Distribution of the silicate and nitrate anom~:ly (observed minus value predicted from the ideal mixing line between riverine and oceanic end-members) for surface waters of the Amazon shelf during Cruise II. Both tracers indicate that the outer shelf is the zone of greatest nutrient uptake. In the coastal waters north of the river mouth, the positive anomalies for silicate and nitrate indicate that regenerated nutrients are being mixed upward into the surface waters (probably as a result of tidal resuspension).

during the four AMASSEDS cruises went to zero at salinities ranging from 12 to 17 psu. During most cruises (I, II and IV), the high nitrate concentration of the riverine endmember (12-23/~mol kg -1) was diluted conservatively in the high turbidity, low salinity waters of the inner shelf and then went to near zero values after the suspended solid concentrations dropped below 5 or 10 mg 1-1 (Fig. 4). During Cruise III (May 1990), however, the nitrate concentrations remained nearly constant between 0 and 6 psu, indicating an additional source of nitrate to the system (such as nitrification of other nitrogenous species). A similar trend in nitrate was noted by Edmond et al. (1981) during their May/June 1976, high river discharge cruise. Additional research is needed to understand the source of the nitrate in the low salinity mixing zone. During Cruise II, the nitrate anomaly on the Amazon shelf (Fig. 5) was the most negative ( - 4 to -12/zmoi kg -1) in essentially the same area as the negative deviation in silicate (i.e. the outer shelf). Seaward of the 100 m isobath, the salinity was sufficiently high that there was no longer 4 or more/~mol kg ~ of nitrate left in the surface waters and therefore, the anomaly decreased. Like silicate, positive nitrate anomalies occurred on the inner shelf north of the river mouth (Fig. 6), and most likely resulted from resuspension of nitrogen-rich porewaters. Although ammonium is released from porewaters during resuspension events and via desorption (Fanning et al., 1982), nitrifying bacteria quickly oxidize the ammonium to nitrate in these oxygen containing Amazon shelf waters (R. C. Aller, personal communication). Unlike silicate, however, the seabed source of nitrogen extended all the way to the coast during AMASSEDS Cruise II.

274

D . J . DeMaster and R. H. Pope

AmasSeds

o.

!

:

"

4'7

4'5

44

11-3 Silicate M a r c h , 1990

:

Anomaly

(pmoilk9)

~-25-:-1'

"

1020. 30A 40vE J:: 50. E3 60. 70. 80'

90 100

4'3 4'2 Station # ( o f f s h o r e

AmasSeds

0. 10.

!

4'1 --->)

11-3 Nitrate A n o m a l y M a r c h , 1990

4'0

3'9

3'8

(pmol/kg)

..:2

20. 30 E

40

.c 50,,~ 60 ° 7080'

90 100

4'7

4'5

44

4'3 42 4'1 Station # (offshore --->)

4'0

3~9

3'8

Fig. 6. Silicate and nitrate anomaly plotted as a function of depth along the Open Shelf Transect (Stas 38-47) during Cruise II. The enhancement of the positive anomaly with depth on the inner shelf suggests that the source of the regenerated nutrients is probably the seabed. Tidal resuspension may be the dominant mechanism for releasing the nutrient-rich porewater to the overlying water column.

Nutrient dynamicsin Amazon shelf waters AMAZON

SHELF

SURFACE

May

200

"~ ,.,¢

WATERS

1990

y = 19.423+ 6.8739X ,~= 0.7~,/

150 ,

0

E :z

275

BgB m

100

B

0 m

~

50"

[]

m

[]

[]

v0

10

2'0

Nitrate

(/.Imol/kg)

0.8 II

D

m

[]

0.6"

[]

=

[]

n m m

v

0.4"' [] ~

/

~

[] B

~

[] =m

0.2 * y = {~14+ 0~37x R^2 = 0.730 00F 0

-

,

•

10 Nitrate

,

20 (/.Imol/kg)

Fig. 7. Phosphate and silicate concentrations in Amazon shelf surface waters plotted as a function of nitrate concentration for AMASSEDS Cruise III (May 1990;high river discharge). In all four of the cruisesthe phosphate and the silicateintercepts(i.e. zero nitrate concentration)were positive and significantlydifferent from zero indicating that the algae on the shelf are primarily limited by nitrate and not phosphate or silicate. Regressions of the various nutrient concentrations have been used to evaluate the limiting nutrient in estuarine systems (Dortch and Whitledge, 1992). The regression of silicate vs nitrate from A M A S S E D S Cruise III (Fig. 7) is controlled primarily by dilution rather than by biological uptake. The slopes of the silicate/nitrate regressions for the four Amazon shelf cruises varied from four to nine, with the silicate intercept (at zero nitrate) ranging from 10 to 25/~mol silicate kg -1 (with the 95% confidence limit varying from 2-7 /~mol kg-1). These data suggest that phytoplankton production on the Amazon shelf rarely, if ever, is limited by silicate.

Phosphate. The systematics of estuarine phosphate behavior have been discussed extensively during the past 10 years (e.g. Fox et al., 1986; Froelich, 1988; Fox, 1989; Berner and Rao, 1994). A kinetic model based on adsorption/desorption reactions is

276

D.J. DeMaster and R. H. Pope

favored by Froelich (1988), whereas a thermodynamic approach, based on a ferric phosphate/ferric hydroxide mineral, is supported by Fox (1989). The supporting measurements necessary to examine phosphate control mechanisms on the Amazon shelf (e.g. adsorption and desorption rate constants, and dissolved iron concentrations) were not made during the field program, therefore, only a descriptive approach will be used in this paper. During all four cruises there is an increase in the phosphate concentration at intermediate salinities (5-20 psu) within the mixing zone, probably as a result of desorption reactions. The maximum phosphate concentration (-0.8/~mol kg -l) occurred during the four field programs at salinities ranging between 15 and 20 psu. Seaward of the maximum, phosphate concentrations in surface waters dropped off quickly during some cruises (e.g. Cruise II) and relatively slowly during others (Cruise IV). The phosphate depletion at these high salinities corresponds to the removal of nitrate and silicate, and consequently, is attributed to biological uptake. Fox et al. (1986) described significant removal of dissolved phosphate at salinities between 0 and 4 psu, which he ascribed to coprecipitation with iron and humic acid compounds. These low salinity processes are apparent in the phosphate/salinity relationships during Cruise I (Fig. 2) as well as during Cruises Ill and IV. During Cruise II our inability to obtain a series of low salinity samples probably prevented us from observing this phenomenon. Chase and Sayles (1980) were among the first to recognize the release of soluble phosphorus from Amazon River sediment during river/ocean mixing. They suggested that the particulate phosphorus, released into solution during estuarine mixing, may be as much as 15 times greater than the dissolved phosphate in the riverine end-member. Confirming these ideas, Richey et al. (1991) report a mean dissolved phosphate concentration of 0.79 Bmol 1-1 in Amazon River water and a particulate phosphate value of 7.8 ~tmol 1 i (most of which is sorbed to inorganic suspended particles). Fox et al. (1986) modeled the partitioning of dissolved and particulate phosphate on the Amazon shelf and concluded that more than half of the phosphate reaching the open ocean is released from the particulate phase during river/ocean mixing. Berner and Rao (1994) recently described measurements of the phosphorus contents of Amazon basin and Amazon shelf sediments including partitioning among organic, iron-associated, and detrital phases. They report that nearly two-thirds of the riverine phosphate released to the open oceans comes from solubilized sediment phosphorus. Regressions of phosphate vs nitrate concentrations (Fig. 7) for the four Amazon shelf cruises yield slopes varying from 0.036 to 0.051 and phosphate intercepts (zero nitrate) ranging from 0.09 to 0.22~mol kg- i (with the 95% confidence limit varying from 0.02~3.03 /~mol kg-l). These values are nearly identical to those of Edmond et al. (1981) who reported P:N slopes of 0.04 to 0.05 and a zero nitrate intercept of 0.1/~mol kg- 1 phosphate. Although these phosphate intercepts are quite low, the phosphate intercept in all four cruises was positive and significantly greater than zero, suggesting that the system is not strongly phosphate limited. A m m o n i u m , nitrite and urea. Ammonium concentrations generally exhibited significant scatter when plotted against salinity (Fig. 2), most likely as a result of the complexity of the uptake, regeneration and nitrification reactions occurring on the shelf. The concentrations of this nutrient typically were below 1/~mol kg -~, with the exception of Cruise III when values in surface and subsurface waters reached as high as 1.7 ~mol kg -1. Subsurface waters, especially those at high salinities, tended to be enriched in ammonium relative to

Nutrient dynamicsin Amazon shelfwaters

277

surface values of comparable salinity. Because nitrate concentrations in the riverine end-member exceeded ammonium concentrations by a factor of 30 to 60 in our study, nitrate was the dominant nutrient form of inorganic nitrogen in the low salinity waters of the shelf. As described above, nitrate values in surface waters commonly went to near zero values at salinities between 12 and 17 psu. Seaward of this nitrate depletion zone, ammonium and nitrate occurred at comparable levels in surface waters. Nitrite and urea concentrations in the Amazon River/ocean mixing zone (Fig. 2) generally were less then 1/~mol kg-1, with little systematic pattern as a function of salinity. The highest nitrite values occurred in high salinity, subsurface waters, probably as a result of local regeneration or upwelling. Urea concentrations were as high as 2-3/~mol kg -1 during Cruise III (the same cruise that exhibited the high ammonium values). As with ammonium, nitrite and urea concentrations in high salinity surface waters often were comparable in magnitude to the nitrate concentrations. At high salinities, subsurface nitrate values, however, commonly reached 4 or 5/~mol kg -~, which was significantly greater than any of the other nitrogen species measured (ammonium, nitrite or urea).

Nutrients fluxes from Amazon shelf sediments In order to examine the rate of nutrient exchange between Amazon shelf sediments and the water column, flux core measurements were made during Cruise IV at eight site (Stas 4308, 15, 19, 33, 44, 45, 47 and 49) representing various depositional regimes in the study area. Immediately after collecting a box core, an 8.25-cm I.D. subcore was inserted through the overlying water into the sediment. Typically, about 5 cm of overlying water were allowed to remain over the sediment core. The nutrient concentrations in the overlying water were sampled initially and then two or three more times during the next 12 days. The subcore was stored in the dark at room temperature (comparable to bottom water temperatures on the shelf) and the overlying water was stirred gently prior to sampling. The slope of the nutrient concentration vs time plot was used to calculate the chemical flux from the seabed. A more detailed description of the flux core technique can be found in Aller et al. (1985). Flux core data from two of the Amazon shelf cores are shown in Fig. 8. Silicate concentrations increased with time in all eight of the cores examined, whereas the slopes of the nitrate, ammonium, and phosphate plots showed mixed (positive and negative) values. A summary of the seabed nutrient fluxes from Cruise IV is given in Table 2. A flux for nitrate plus ammonium was calculated because much of the ammonium diffusing out of the seabed may have undergone nitrification reactions in the oxic overlying water. Silicate has the largest flux of all the nutrients measured. The mean phosphate flux is negative (i.e. into the seabed in six of the eight cores), but the variation in the fluxes makes the standard deviation greater than the mean flux. The mean nitrate and ammonium fluxes are near zero or slightly positive, however, the standard deviation again exceeds the mean value. These fluxes are in reasonable agreement with the results of Aller et al. (1992, 1996), who has examined porewater fluxes during all four AMASSEDS cruises. The flux core measurements were made at sea on a moving/rolling ship, and consequently the fluxes may be slightly greater than fluxes generated solely by diffusion. If the mean silicate flux from Cruise IV is considered to be typical of the Amazon mud wedge ( - 7 × 1010 m 2) and representative of the long-term flux, the amount of silicate released from the seabed is equal to 2.9 × ]0 7 m o l d -1, which is only 1.3% of the riverine

278

D.J. DeMaster and R. H. Pope 30

AMASSEDS IV-3 Flux Core Data Be 4 3 0 8

0

E

~ ' ~

/

20

10

•

4308-NH4

O •

4308-P 4308.NO3

+ A

4308-Si 4308-N+N

c

o

J_ v

z

200

',00

' 600

" 800

" 10'00

1200

Time (MIn)

80A

AMASSEDS IV-3 Flux Core Data

0

E v

P c o O

40"

• O •

4319-NH4 4319-P 4319-NO3

-I-

4319-Si 4319-N+N

2O

I=

6

o

Z

/ +

60

m

0 ' 0

O, 1000

2000

Time (MIn)

Fig. 8. Flux core measurements from two box cores (Stas 8 and 19) collected during A M A S S E D S Cruise IV. Silicate increases steadily in the waters overlying the sediment core, indicating a release of the nutrient from the seabed. Nitrate and ammonium fluxes from shelf sediments were positive in some cases (e.g. BC 4308), but zero in others (e.g. BC4319). The flux of phosphate to the overlying waters was zero or slightly negative (i.e. into the seabed) for the eight cores examined during the cruise.

supply (2.2 × 109 mol d-X). This silicate flux represents a minimum estimate of the seabed/water column exchange rate because some silicate may be released to the water column via tidal resuspension (see earlier discussion). Quantifying the flux of silicate released via tidal resuspension is very difficult because of the stochastic nature of the process. However, even if it were twice as great as the seabed flux measured using flux cores, the total silicate flux from the mud wedge to the overlying waters would be small relative to the riverine supply. A similar calculation for phosphate indicates that mud wedge sediments take up approximately 2% (0.4 × 106 m o l d -1) of the phosphate released from riverine sources in the Amazon River/ocean mixing zone (2.2 × 107 mol d - l ; Fox et al., 1986). Resuspension also may play an important role in the case of phosphate regeneration because not only would phosphate-rich porewaters be released to the water

Nutrient dynamicsin Amazon shelf waters

279

Table 2. Nutrient fluxes from Amazon shelf sediments

Nutrient Silicate Phosphate Nitrate Ammonium Ammonium + Nitrate

Range in flux (mmol m-2 d -1)

Mean flux (mmolm-2 d-1)

Std. dev. * (mmolm-2 d -l)

0.13--1.25 -0.015-0.0004 -0.15-0.12 - 0.14-0.24 -0.25-0.36

0.42 -0.0055 0.00 0.030 0.043

0.36 0.0059 0. ] 0 0.16 0.22

*N = 8, for all of the nutrient flux determinations.

column, but the seabed particles themselves would be exposed to lower phosphate concentrations in the water column, which would promote desorption. Mackin et al. (1988) modeled the time necessary to develop porewater ammonium and iodine profiles observed at numerous sites on the shelf. Values ranged from 2-8 months, suggesting that frequent resuspension to substantial depths in the seabed is unlikely. Our data indicate clearly that very little, if any, biologically available inorganic nitrogen ( < 1 % of riverine flux) is released diffusively from the seabed to the overlying water column. Therefore, the diffusive fluxes of silicate, biologically available inorganic nitrogen, and phosphate across the sediment/water interface probably play a relatively minor role in the nutrient budgets of the shelf. For comparative purposes, Aller et al. (1985) reported seabed nutrient fluxes of 0.1-13.2 (mean = 2.1) mmol m -2 d - 1 for silicate, -0.37-0.63 (mean = 0.13) mmol m -2 d -1 for nitrate, and 0.13-0.87 (mean = 0.54) mmol m -2 d -1 for ammonium from East China Sea sediments near the mouth of the Changjiang River. The deltaic sediments of the Changjiang generally release more nutrients than their counterparts on the Amazon shelf. Advection o f nutrients f r o m offshore waters

The lateral extent of the algal blooms on the Amazon shelf may vary temporally, but the basic feature is present on the outershelf during all seasons (Curtin and Legeckis, 1986; Muller-Karger et al., 1988; DeMaster et al., 1992). As described previously, the external sources of nutrients to these blooms are not well documented. The nutrient flux from the river and seabed can be quantified in a reasonably straightforward fashion (see previous sections), however, the nutrients supplied from the open ocean (as a result of the estuarine-like circulation) are more difficult to quantify. To estimate the nutrient flux from offshore waters, the depth of origin for the waters as well as the magnitude of the flow must be known. The depth of origin for the advecting waters can be estimated from the density profiles for the shelf and surrounding waters. Figure 9 shows the distribution of sigma-t values for the Open Shelf Transect (Stas 4t--50) during Cruise III. The isopycnal contours indicate that the water moving onto the shelf originated from a depth of 60 to 100 m. The magnitude of the shoreward flow can be approximated using the naturally occurring radionuclide, 21°Pb (half life 22 years). The inventory of this radionuclide in Amazon deltaic sediments (2.4 x 1017 dpm) greatly exceeds the amount that can be supported from atmospheric and riverine sources (0.6 x 1017 dpm), leaving 1.8 × 1017 dpm to be supplied from advecting offshore waters

280

D.J. DeMasterand R. H. Pope STA. 0

20

~

50

49

I

I

I

I

I

60

45

44

43

42

41

I

I

I

I

I

o.

X~.

40

8O

48 47 46

22.0 ~ 2 3 . 0 ~

'\ . ~. ~ o

111-30ST SIGMA T

~100

120 -

~ • 5 ,,~.~,--

//--26.o

•

Fig. 9. Sigma-t contours along the Open Shelf Transect (Stas 41-50) during Cruise III (May 1990; high river discharge). Based on the slope of these isopycnal lines, the nutrients moving up onto the shelf (as a result of the estuarine-like circulation) originate from a depth of -60-100 m.

(DeMaster and Nittrouer, 1987; Smoak et al., 1992; Smoak, 1994). Lead is a particlereactive species in seawater, and consequently, is scavenged from solution by particulate matter. In open ocean waters the particulate flux is low and 21°pb can reach secular equilibrium with its effective parent, 226Ra (0.1 dpm l-a; Chan et al., 1976). As the open ocean waters move shoreward, however, they encounter the high turbidity of the Amazon shelf, which scavenges the dissolved 21°pb from the advecting waters, depositing the radionuclide in the deltaic sediments below (DeMaster et al., 1986). To sustain the . missing, , 2 1 °Pb inventory of 1.8 × 1017 dpm requires a steady-state flux of 5.6 × 1015 dpm y-1. If the advecting waters carry 0.1 dpm 1-1 210pb (McKee, 1986) and all of it ends up in the deltaic sediments, then the particles annually must be scavenging 2~°pb from approximately 5.6 × 10161 of ocean water. The scavenging process is believed to occur up on the Amazon shelf, where suspended solid concentrations are one to two orders of magnitude greater than in open-ocean waters. This supply of offshore water (5.6 z 1016 1y-~) to the Amazon shelf is approximately ten times the flow of the Amazon River (Oltman, 1968), but only 9% of the North Brazilian Current (Philander and Pacanowski, 1986). These calculations are dependent on accurate estimates of atmospheric 21°Pb supply as well as the 21°pb riverine flux (Smoak, 1994). However, even if the final onshore flux of water were in error by a factor of two (an upper limit for these parameters), the shoreward flow of water would be at least five times the river discharge from the Amazon. A unique feature of using 21°pb as a water-mass tracer is that it integrates transport processes over a 100-year time scale (and is not very sensitive to seasonal fluctuations in

Nutrient dynamics in Amazon shelf waters Table 3.

281

Sources o f nutrients to outer shelf Amazon blooms

Riverine/offshore flow

Nutrient Silicate Nitrate Phosphate Ammonium

Riverine end-member (ttmol kg-l)

Offshore end-member* ~mol kg-1)

144 16 1.4t 0.4

3.0 1.0 0.15 0.16

1:5 1:10 % nutrient % nutrient river/offshore river/offshore 91/9 76/24 65/35 33/67

83/17 62/38 48/52 20/80

*These values represent the mean of the silicate, nitrate, phosphate and ammonium values measured at -80 m depth in offshore stations sampled during Cruises I, II, III and IV. tThis riverine end member value includes the 0.7 /~mol kg 1 of dissolved phosphate in the river plus 0.7 /2mol kg-1 of phosphate that desorbs during river/ocean mixing (Fox et al., 1986).

discharge). Presuming that the shoreward m o v e m e n t of 2t°pb-rich waters also carries nutrients characteristic of the origination depth (60-100 m), the shoreward flux of silicate, nitrate, a m m o n i u m and phosphate can be determined (Table 3). Once the relative discharges of the river and shoreward flow are designated, the percentage of the external nutrient flux reaching the outer shelf that originates from the river becomes a direct function of the relative concentrations in the riverine e n d - m e m b e r and the offshore subsurface e n d - m e m b e r . The ratio of the riverine to offshore nutrient values is largest for silicate (48), and consequently, most (83-91%) of the silicate externally supplied to blooms in outershelf waters comes from the river. Nitrate has a ratio of 16 for the riverine and offshore e n d - m e m b e r s , and therefore, the riverine proportion of this nutrient supplied to outershelf waters (62-76%) is less than silicate. Phosphate and a m m o n i u m have riverine/offshore e n d - m e m b e r ratios of 9 and 2.5, respectively. The percentage of these nutrients externally supplied from the river is 48--65% for phosphate and 20-33% for a m m o n i u m . Therefore, the dominant external source of silicate and nitrate to the outershelf algal blooms is from rivers; approximately equal amounts of phosphate are supplied to the blooms from rivers and shoreward advection; and the primary source of a m m o n i u m to the outershelf blooms is from the shoreward advection of offshore, subsurface waters. The conclusions of Cadee (1975) and Van B e n n e k o m and Tijssen (1976) that offshore upwelling sustains much of the productivity off the east coast of Brazil and G u y a n a appears to be corroborated most clearly by the A M A S S E D S a m m o n i u m and phosphate data. To obtain a m o r e complete view of nutrient cycling on the A m a z o n shelf, nutrient recycling (an internal regenerative flux) should be considered in addition to the external supply from rivers and upwelling. Based on 14C primary productivity estimates made during A M A S S E D S (Smith and DeMaster, 1996), the mean productivity for the shelf is on the order of 2.2 g C m -2 d -1, which corresponds to a production rate for the shelf (0.89 x 10 H m 2) of - 2 . 0 x 10 it g C d-1. Assuming a Redfield C/N (106/16) and a Redfield C/P (106/1) mole ratio for phytoplankton uptake, the carbon production on the shelf requires 25 × 108 moles d -1 of nitrogen and 1.5 x l0 s moles d -1 of phosphate. A

282

D.J. DeMaster and R. H. Pope

similar calculation based on a Si/C mole ratio of 0.15-0.4 (typical of coastal plankton, Parsons et al., 1984b; Brzezinski, 1985) yields a silicate uptake rate of 24-65 × 108 moles d -1 . Subtracting the riverine and offshore advection fluxes (4.3 x 10s moles d -1 for nitrate and ammonium; 0.46 × 108 moles d -I for phosphate, and 27 × 108 moles d -1 for silicate) from the total nutrient requirements for primary production yields the nutrient supply rate from regeneration on the shelf. These data indicate that - 8 3 % of the nitrate and ammonium, - 6 9 % of the phosphate, and 0-59% of the silicate used in primary production on the shelf must come from regeneration. The percentage of the primary production based on regenerated nitrogen for the Amazon shelf is comparable to the value measured off the coast of Rhode Island ( - 7 0 % , Harrison, 1992). Therefore, despite the large influx of nutrients from the Amazon River and from upwelling of offshore waters, the largest supply of nutrients to the algae on the outer shelf comes from internally regenerated biogenic material. Nutrient budgets on the A m a z o n shelf

A fundamental question concerning nutrient dynamics in estuarine systems is: "how much of the riverine nutrient flux ultimately makes it to the open ocean?". Twenty to thirty years ago, the concept that estuaries function as nutrient traps was common in the literature (see Norwicki and Oviatt, 1990, for a review). More recent research, however, indicates that very little retention of nutrients (i.e. burial of nutrients as biogenic matter) occurs within many river/ocean mixing zones. For example, Nixon (1987) studying the Chesapeake estuary determined that approximately 95% of the inorganic nitrogen, 86% of the phosphate, and about 40% of the silicate coming down the Chesapeake River ultimately makes it to the open ocean. Based on biogenic silica measurements in Amazon shelf sediments, DeMaster et al. (1983) concluded that despite 30-50% of the riverine silicate being consumed by phytoplankton within the mixing zone, less than 4% of the riverine silicate flux was buried in Amazon shelf sediments. Previous discussion of the seabed silicate flux indicated that very little silicate ( - 1.5% of the riverine flux) is transported diffusively back into the overlying waters, therefore, most of the biogenic silica dissolution probably occurs in the water column. This is consistent with the results of Nelson (1992) who measured silica uptake and dissolution rates in Amazon shelf waters and observed little if any net silica production outside of the "green waters" that occur just seaward of the turbid plume. To characterize the retention of nitrogen and phosphorus, the amount of marine organic matter buried on the shelf must be known. Based on stable carbon isotope signatures in riverine and planktonic end-members as well as on carbon isotopic data from shelf sediments, Showers and Angle (1986) used 21°pb sediment accumulation rates to calculate the rate of marine carbon burial for the Amazon shelf (1.4 x 1012 g org. C y-l). Although this number is only a rough estimate, it can be used to approximate the estuarine trapping efficiency for nitrogen and phosphorus. Based on measured C/N values of Amazon shelf plankton samples, the C/N weight ratio of marine organic matter in the field area is approximately six (in good agreement with the Redfield ratio). This value coupled with the marine carbon burial rate yields a nitrogen burial rate of 2.3 x 1011 g y-1 (or 0.45 x l0 s moles d-l). Adding all of the nutrient fluxes of dissolved inorganic nitrogen (nitrate, nitrite and ammonium) in the river and advecting offshore waters together, yields a total flux of 4.3 x 10s moles of N d -1. Therefore, of all the externally supplied, nutrient

Nutrient dynamicsin Amazon shelfwaters

283

inorganic nitrogen available to plankton blooms on the Amazon shelf, only 10% is buried in shelf sediments and 90% is released to the open ocean. A similar calculation for phosphate (assuming a P/C mole ratio 0.4/106; Froelich et al., 1982) yields a phosphorus burial rate of 1.2 × 10 6 moles d -1, which corresponds to <3% of the total phosphate available to plankton during Amazon River/ocean mixing (riverine plus offshore advection; 0.46 × 108 moles of P d - 1). Even if the phosphorus is buried with a Redfield P/C mole ratio of 1:106, less than 7% of the externally supplied phosphate accumulates as organic matter in Amazon shelf sediments. Therefore, none of the major nutrients (silicate, nitrogen or phosphate) are trapped very efficiently in Amazon shelf sediments. Comparison o f nutrient dynamics on the A m a z o n shelf to other estuarine systems

Nutrient dynamics in the Para River, located just south of the Amazon River mouth, offer an interesting comparison with the processes in the Amazon River because the Para River is a low-land river with relatively low suspended sediment concentrations (typically <20 mg 1-1). Figure 10 shows the nutrient/salinity relationships for the Para River/ocean mixing zone during November 1991 (Cruise IV). The silicate, nitrate, and phosphate concentrations of the riverine end-member are somewhat lower than the Amazon, whereas the ammonium, nitrite and urea concentrations are comparable to those reported in Table 1. Because of the relatively low suspended sediment concentrations, the uptake of silicate, nitrate and phosphate occurs at rather low salinities (3-10 psu) in the Para estuary. Nitrate remains the dominant form of inorganic nitrogen up to a salinity of about 25 psu. Phosphate concentrations in the Para estuary increase between 0 and 4 psu (probably as a result of desorption), but then decrease between 5 and 10 psu as a result of biological uptake (coincident with silicate and nitrate removal). In the high salinity portion of the mixing zone, nutrient levels remain significantly above detection limits, with the exception of nitrate, which goes to zero at about 29 psu. A comparison of Amazon and Zaire riverine end-members (Van Bennekom et al., 1978) shows that the Amazon is enhanced in nitrate (16 vs 6.5 #mol kg-1), but has comparable amounts of silicate, phosphate and ammonium. The interesting feature of the Zaire, a major tropical dispersal system, is that it has only 30 mg 1-1 of suspended sediment (as compared to the Amazon which has - 2 0 0 mg 1-1). There is, however, very little indication of nutrient uptake during Zaire River/ocean mixing (Van Bennekom et al., 1978), and there is a significantly lower amount of phytoplankton biomass in the Zaire shelf water column than on the Amazon shelf (Cadee, 1978; Smith and DeMaster, 1996). Van Bennekom et al. (1978) state that phytoplankton production on the Zaire shelf may be light limited in certain areas (there is a large dissolved organic component). In addition, the authors state that short residence times for surface waters in the mixing zone may not give the phytoplankton ample time for growth and bloom formation. Primary production in the Amazon River/ocean mixing zone is limited by light near the river mouth, and by nutrients (primarily nitrate) further seaward. The development of algal blooms on the Amazon shelf also appears to be controlled by surface water residence time (DeMaster et al., 1992; DeMaster et al., 1996), which commonly is controlled by the wind speed and direction. Nutrient systematics have been examined for two other tropical South American rivers, the Orinoco and the Magdalena. Fanning and Maynard (1978) observed little evidence of silicate removal in the Magdalena estuary, whereas in the Orinoco, Edmond et al. (1985) observed essentially conservative behavior for silicate, but modest uptake of

284

D . J . DeMaster and R. H. Pope

100

14

Para River November, 1991

A

=_. eo-

0

E 6o-

E 0-

3

+ 40-

O

1++

r~

o

6-

t_

a

4-

+ + +

+

+ 1't.

Z

20

;o l's 2'o 2's 3'o 3s

o

+

1-

1,2"

Para River November, 1991

E

=_.

50.8 O

; 0.4. £

+ + ++ ++

+

E

.l 0.4.

÷

+ .+ +

+ 1-

I

5

I

I

± +4

]

'I

10 15 20 25 30 35

salinity (PSU)

Para River November, 1991 + +

1-

+ +1- + 1-

+ + ++

Ii

+

~ 0.2. £

0.8

~= o.0.

÷ +

+

I

0

sa,n~ (PSU)

-~0 0.8.

0.

0 1-

= 20-

..., gl

Para River .~1-+November,1991

Z 0.2.

0

1'0 I'5 210 2'5 3'0 35

1'0 I'5 20 215 3'0 35

Salinity (PSU)

Para River

J¢ 0.8.

a 0.8.

November, 1991

E ~= 0.6. E = 0.4 0

Salinity (PSU)

÷

m

0 O.6-

+

E

++ 3 + ~ 0.4. +++

+ ÷ +

E 0.2 E

Para River November, 1991

+

+

t-

0 +

~

+ t-

+

~

|

+ ÷

-~ 0.2 + -I"

o ~ lo is 2.0 2's 35 35 salinity (PSU)

0

+

+

+

+ +

+

~ 1'o 1'5 2'0 15 3.0 35 Salinity (PSU)

Fig. 10. Nutrient vs salinity relationships for surface waters from the Para River/ocean mixing zone during Cruise IV (November 1991; low river discharge). Because of the lack of turbidity in this low-land river, the biological uptake of nutrients begins at much lower salinities than in the Amazon River/ocean mixing zone.

Nutrient dynamicsin Amazonshelfwaters

285

nitrate and phosphate. In a more recent study of the Orinoco, Bonilla et al. (1993) reports uptake of silicate and nitrate in the intermediate salinities of the mixing zone. The cause of this variability is unknown at present. There are interesting differences and similarities comparing nutrient systematics in the Changjiang (Yangtze) River to those of the Amazon. In addition to the obvious difference in latitude and drainage basin, the Changjiang appears to be significantly impacted by human activity with nitrate concentrations in the riverine end-member of 50 or 60/~mol kg -1 (Edmond et al., 1985). Because of this enhanced supply, nutrient uptake seaward of the turbid plume generally is limited by the phosphate concentration. Another difference between these two large dispersal systems is that the Changjiang shows significant seasonal variation in nutrient uptake (e.g. 20% silicate removal from surface waters during June and conservative behavior for silicate in November; DeMaster and Nittrouer, 1983; Edmond et al., 1985), whereas in the Amazon mixing zone, the nutrient uptake from surface waters generally occurs year round. Based on biogenic silica measurements in Changjiang shelf sediments, DeMaster and Nittrouer (1983) concluded that nearly all of the silicate coming down the fiver ultimately is released to the ocean (similar to the situation for the Amazon dispersal system). Like the Changjiang, phytoplankton growth in the Huanghe mixing zone appears to be limited by light (as a result of high turbidities) as well as by low phosphate concentrations (Turner et al., 1990). In contrast to the Chinese rivers, silicate and nitrate concentrations appear to be important factors in the formation of diatom blooms in the Mississippi dispersal system (Dortch and Whitledge, 1992). SUMMARY Nutrient data from four Amazon shelf cruises have led to the following conclusions: (1) The dominant form of inorganic nitrogen in the Amazon River is nitrate, which exceeds the ammonium and nitrite concentrations by a factor of 30 to 60. The dominance of nitrate continues throughout the river/ocean mixing zone out to the region (salinities between 12 and 17 psu) where biological activity reduces the nitrate levels to values comparable to those of ammonium, nitrite and urea (<1 ~mol kg-1). (2) During all four hydrographic cruises to the Amazon shelf, biological uptake of nutrients was apparent. This year-round productivity is in contrast to the Changjiang estuary, which exhibits a seasonal pattern in plankton production. On the Amazon shelf, light limits nutrient uptake within the turbid plume of the inner shelf, whereas on the outer shelf, nitrate and the surface water residence time appear to control the production to a greater extent than either phosphate or silicate. (3) Although some silicate is released from Amazon shelf sediments to the overlying water column, the diffusive nutrient flux from the seabed, in general, plays a very minor role in the nutrient dynamics of the study area. (4) The dominant external source of nutrients to the algal blooms on the outershelf depends on the particular nutrient specified. Nearly all of the silicate (83%) and most of the nitrate (62%) supplied to the outer shelf come from the river, whereas only half of the phosphate and only a fifth of the ammonium have a riverine source. The shoreward advection of subsurface waters is the dominant source of ammonium and an important source of phosphate to the algal blooms on the outer shelf, and this flow transports approximately 5-10 times the annual flow of the Amazon River.

286

D.J. DeMaster and R. H. Pope

(5) T h e A m a z o n s h e l f is n o t a n e f f i c i e n t n u t r i e n t t r a p . L e s s t h a n 4 % o f t h e silicate t r a n s p o r t e d b y t h e A m a z o n R i v e r is b u r i e d in s h e l f s e d i m e n t s . - 1 0 % o f t h e b i o l o g i c a l l y a v a i l a b l e i n o r g a n i c n i t r o g e n s u p p l i e d to t h e s h e l f is b u r i e d in t h e s e d i m e n t s b e l o w , w h e r e a s less t h a n 3 % o f t h e a v a i l a b l e p h o s p h a t e is r e m o v e d to t h e s e a b e d .

Acknowledgements--We would like to thank the crew of the R.V. Columbus Iselin and our Brazilian colleagues for their assistance in sample collection. We are grateful for the long hours put in by all of the Principal Investigators and by NCSU graduate students and sea-going staff, who included: Stephen Harden, Kevin Craig, Carrie Exton and Heinz Seltmann. Nutrients were analyzed by D. Guffy of the Texas A&M Nutrient Research Group (D. Biggs, Director). This AMASSEDS research was supported by grants from the National Science Foundation, Chemical Oceanography Program. REFERENCES Aller R. C., N. E. Blair, Q. Xia and P. D. Rude (1996) Remineralization ratio, recycling and storage of carbon in Amazon shelf sediment. Continental Shelf Research (in press). Aller R. C., J. E. Mackin, W. J. Ullman, C. H. Wang, S. M. Tsai, J. C. Jin, Y. N. Sui and J. Z. Hong (1985) Early chemical diagenesis, sediment-water solute exchange, and storage of reactive organic matter near the mouth of the Changjiang, East China Sea. Continental Shelf Research, 4, 227-251. Aller R. C., Q. Xia and P. D. Rude (1992) Early diagenesis and benthic carbon remineralization in Amazon shelf muds. EOS (Abstract), 73, 268-269. Armstrong F. A. J., C. R. Stearns and J. D. H. Strickland (1967) The measurement of upwelling and subsequent biological processes by means of the Technicon Autoanalyzer and associated equipment. Deep-Sea Research, 14,381-389. Berner R. A. and J. L. Rao (1994) Phosphorus in sediments of the Amazon River and estuary: Implications for the global flux of phosphorus to the sea. Geochimica Cosmochimica Acta, 58, 2333-2340. Biggs D. C., M. A. Johnson, R. R. Bidigare, J. C. Guffy and O. Holm-Hansen (1982) Shipboard Autoanalyzer Studies of Nutrient Chemistry, 0-200 m, in the Eastern Scotia Sea During FIBEX (January-March 1981). Technical Report 82-1 l-T, Dept. of Oceanography, Texas A&M University, College Station, TX, 98 pp. Bonilla J., W. Senior, J. Bugden, O. Zafirious and R. Jones (1993) Seasonal distribution of nutrients and primary productivity on the eastern continental shelf of Venezuela as influenced by the Orinoco River. Journal of Geophysical Research, 98, 2245-2257. Brzezinski M. A. (1985) The Si:C:N ratio of marine diatoms: interspecific variability and the effect of some environmental variables. Journal of Phycology, 21,347-357. Cadee, G. C. (1975) Primary production off the Guyana Coast. Netherlands Journal of Sea Research, 9, 128-143. Cadee G. C. (1978) Primary production and chlorophyll in the Zaire River, estuary and plume. Netherlands Journal of Sea Research, 12, 368-381. Chan L. H., J. M. Edmond, R. F. Stallard, W. S. Broecker, Y. C. Chang, R. F. Weiss and T. L. Ku (1976) Radium and barium at GEOSECS stations in the Atlantic and Pacific. Earth and Planetary Science Letters, 32,258-267. Chase E. M. and F. L, Sayles (1980) Phosphorus in suspended sediments of the Amazon River. Estuarine and Coastal Marine Science, 11,383-391. Cloern J. E. (1987) Turbidity as a control on phytoplankton biomass and productivity in estuaries. Continental Shelf Research, 7, 1367-1381. Cloern J. E. (1991) Tidal stirring and phytoplankton bloom dynamics in an estuary. Journal of Marine Research, 49,203-221. Curtin T. B. (1986a) Physical observations in the plume region of the Amazon River during peak discharge--ll. Water masses. Continental Shelf Research, 6, 53-71. Curtin T. B. (1986b) Physical observations in the plume region of the Amazon River during peak discharge--III. Currents. Continental Shelf Research, 6, 73-86. Curtin T. B. and R. V. Legeckis (1986) Physical observations in the plume region of the Amazon River during peak discharge--l. Surface variability. Continental Shelf Research, 6, 31-51. DeMaster D. J., G. B. Knapp and C. A. Nittrouer (1983) Biological uptake and accumulation of silica on the Amazon continental shelf. Geochimica et Cosmochimica Acta, 47, 1713-1723. DeMaster D. J., S. A. Kuehl and C. A. Nittrouer (1986) Effects of suspended sediments on geochemical

Nutrient dynamics in Amazon shelf waters

287

processes near the mouth of the Amazon River: examination of biological silica uptake and the fate of particle-reactive elements. Continental Shelf Research, 6, 107-125. DeMaster D. J. and C. A. Nittrouer (1983) Uptake, dissolution, and accumulation of silica near the mouth of the Changjiang River. In: Sedimentation on the Continental Shelf, With Special Reference to the East China Sea, Vol. I, China Ocean Press, 235-240. DeMaster D. J. and C. A. Nittrouer (1987) Geochemical processes occurring at the river/ocean boundary and their effects on chemical fluxes to the open ocean. In: Ocean Margins in GOFS, U.S. GOFS Planning Report No. 6, Woods Hole Oceanographic Institution, Woods Hole, MA, pp. 105-114. DeMaster D. J., R. H. Pope and J. M. Smoak (1992) Biogeochemical processes occurring in shelf waters near the mouth of the Amazon River. EOS (Abstract), 73,269. DeMaster D. J., W. O. Smith Jr., D. M. Nelson and J. Aller (1996) Biogeochemical processes in Amazon shelf waters: Chemical distributions and uptake rates of silicon, carbon and nitrogen. Continental Shelf Research (in press). Dortch Q. and T. E. Whitledge (1992) Does nitrogen or silicon limit phytoplankton production in the Mississippi River plume and nearby regions. Continental Shelf Research, 12, 1293-1309. Edmond J. M., E. A. Boyle, B. Grant and R. F. Stallard (1981) The chemical mass balance in the Amazon plume-l. The nutrients. Deep-Sea Research, 28, 1339-1374. Edmond J. M., A. Spivack, B. C. Grant, M. H. Hu, A. Chen, S. Chen and X. Zeng (1985) Chemical dynamics of the Changjiang estuary. Continental Shelf Research, 4, 17-36. Fanning K., K. L. Carder and P. R. Betzer (1982) Sediment resuspension by coastal waters: a potential mechanism for recycling on the ocean's margins. Deep-Sea Research, 29, 953-965. Fanning K. A. and V. I. Maynard (1978) Dissolved boron and nutrients in the mixing plumes of major tropical rivers. Netherlands Journal of Sea Research, 12,345-354. Fox L. E. (1989) A model for inorganic control of phosphate concentrations in river waters. Geochimica et Cosmochimica Acta, 53, 417-428. Fox L. E., S. L. Sager and S. C. Wofsy (1986) The chemical control of soluble phosphorus in the Amazon estuary. Geochimica et Cosmochimica Acta, 50,783-794. Froelich P. N. (1988) Kinetic control of dissolved phosphate in natural rivers and estuaries: A primer on the phosphate buffer mechanism. Limnology and Oceanography, 33, 649--667. Froelich P. N., M. L. Bender, N. A. Luedtke, G. R. Heath and T. DeVries (1982) The marine phosphorus cycle. American Journal of Science, 282, 474-511. Gibbs R. J. (1972) Water chemistry of the Amazon River. Geochimica et Cosmochimica Acta, 36, 1061-1066. Gibbs R. J. (1982) Currents on the shelf of Northeastern South America. Estuarine Coastal Shelf Science, 14. 283-299. Harrison W. G. (1992) Regeneration of nutrients. In: Primary Production and Biogeochemical Cycles in the Sea, P. G. Falkowski and A. D. Woodhead, editors, Environmental Science Research Series, Vol. 43, pp. 385407. Hulburt E. M. and N. Corwin (1969) Influence of the Amazon River outflow on the ecology of the western tropical Atlantic--Ill. The planktonic flora between the Amazon River and the Windward Islands. Journal of Marine Research, 27, 55-72. Kennish M. J. (1990) Ecology of Estuaries. Volume II: Biological Aspects, CRC Press, Boston, MA, 391 pp. Key R. M., R. F. Stallard, W. S. Moore and J. L. Sarmiento (1985) Distribution of Ra-226 and Ra-228 in the Amazon River estuary. Journal of Geophysical Research, 90, 6995-7005. Kineke G. C., R. W. Sternberg, J. H. Trowbridge and W. R. Geyer (1996) Fluid mud processes on the Amazon continental shelf. Continental Shelf Research (in press). Kuehl S. A., D. J. DeMaster and C. A, Nittrouer (1986) Nature of sediment accumulation on the Amazon continental shelf. Continental Shelf Research, 6,209-225. Lcsack L. F. W. (1993) Export of nutrients and major ionic solutes from a rain forest catchment in the central Amazon Basin. Water Resources Research, 29, 743-758. Limeburner R. and R. C. Beardsley (1989) CTD Observations on the North Brazil Shelf during AMASSEDS, August 1989, Technical Report, Woods Hole Oceanographic Institution, 137 pp. Lohrenz S. E., M. J. Dagg and T. E. Whitledge (1990) Enhanced primary production at the plume/oceanic interface of the Mississippi River. Continental Shelf Research, 10, 639-664. Mackenzie F. T. and R. M. Garrels (1966) Chemical mass balance between rivers and oceans. American Journal of Science, 264, 507-525.

288

D.J. DeMaster and R. H. Pope

Mackin J. E., R. C. Aller and W. J. Ullman (1988) The effects of iron reduction and nonsteady state diagenesis on iodine, ammonium, and boron distributions in sediments from the Amazon continental shelf. Continental Shelf Research, 8,363-386. McKee B. A. (1986) The fate of particle-reactive radionuclides on the Amazon and Yangtze continental shelves. Ph.D. Dissertation, North Carolina State University, Raleigh, NC, 239 pp. Meade R. H., T. Dunne, J. E. Richey, U. De M. Santos and E. Salati (1985) Storage and remobilization of suspended sediment in the lower Amazon River of Brazil. Nature, 278, 161-163. Milliman J. D. and E. A. Boyle (1975) Biological uptake of dissolved silica in the Amazon River estuary. Science, 189, 995-997. Moore W. S., D. J. DeMaster, J. M. Smoak, B. A. McKee and P. W. Swarzenski (1996) Radionuclide tracers of sediment-water interactions on the Amazon shelf. Continental Shelf Research (in press). Muller-Karger F. E., C. R. McClain and P. L. Richardson (1988) The dispersal of the Amazon's water. Nature, 333, 56-59. Nelson D. M. (1992) Biological removal of dissolved silica in the plume of the Amazon River: Spatial constraints set by nutrient limitation and turbidity. EOS (Abstract), 73,277. Nittrouer C. A., D. J. DeMaster, A. G. Figueiredo and J. M. Rine (1991) AMASSEDS: An interdisciplinary investigation of a complex coastal environment. Oceanography, 4, 3-7. Nittrouer C. A., S. A. Kuehl, A. G. Figueiredo, M. A. Allison, C. K. Sommerfield, D. Dukat, J. M. Rine, L. E. C. Faria and O. M. Silveira (1996) The geological record preserved by Amazon shelf sedimentation. Continental Shelf Research (in press). Nixon S. (1987) Chesapeake Bay nutrient budgets: a reassessment. Biogeochemistry, 4, 77-90. Nowicki B. L. and C. A. Oviatt (1990) Are estuaries traps for anthropogenic nutrients? Evidence from estuarine mesocosms. Marine Ecology Progress Series, 66,131-146. Officer C. B. (1979) Discussion of nonconservative dissolved constituents in estuaries. Estuarine and Coastal Marine Science, 9, 95-99. Oltman R. E. (1968) Reconnaissance investigations of the discharge and water quality of the Amazon River. U.S. Geological Survey. Circular 552, Washington D.C., 16 pp. Parsons T. R., Y. Maita and C. M. Lalli (1984a) A Manual of Chemical and Biological Methods for Seawater Analyses, Pergamon Press, 167 pp. Parsons T. R., M. Takahashi and B. Hargrove (1984b) Biological Oceanographic Processes, 3rd edition, Pergamon Press, 330 pp. Philander S. G. H. and R. C. Pacanowski (1986) The mass and heat budget in a model of the tropical Atlantic Ocean. Journal of Geophysical Research, 92, 3691-3708. Picaut J., J. Servain, P. Lecompte, M. Seva, S. Lukas and G. Rougier (1985) Climatic Atlas of the Tropical Atlantic, Wind Stress and Sea Surface Temperature, 1964-1979, Universite de Bretagne Occidentale, 467 pP. Price N. M. and P. J. Harrison (1987) Comparison of methods for the analysis of dissolved urea in seawater. Marine Biology, 94, 307-317. Richey J. E., R. L. Victoria, E. Salati and B. R. Forsberg (1991) The biogeochemistry of a major river system: The Amazon case study. In: Biogeochemistry of Major World Rivers, SCOPE 42, E. T. Degens, S. Kempe and J. E. Richey, editors, John Wiley & Sons, New York, pp. 57-74. Ryther J. H., D. W. Menzel and N. Corwin (1967) Influence of the Amazon River outflow on the ecology of the western tropical Atlantic, I: Hydrology and nutrient chemistry. Journal of Marine Research, 25, 69-82. Showers W. J. and D. G. Angle (1986) Stable isotopic characterization of organic carbon accumulation on the Amazon continental shelf. Continental Shelf Research, 6, 227-244. Smith W. O. Jr. and D. J. DeMaster (1996) Phytoplankton biomass and productivity in the Amazon River plume: correlation with seasonal rioter discharge. Continental Shelf Research, 16,291-319. Smoak J. M. (1994) The behavior of particle-reactive tracers in a high turbidity estuarine environment: 234Th, 228Th and 21°Pb on the Amazon shelf. Master's Thesis, North Carolina State University, Raleigh, NC. 53 pp. Smoak J. M., D. J. DeMaster, R. H. Pope and B. A. McKee (1992) Particle scavenging and seabed dynamics on the Amazon shelf as determined from measurements of 234Th and 21°pb. EOS (Abstract), 73,277. Stallard R. F. and J. M. Edmond (1983) Geochemistry of the Amazon, 2. The influence of geology and weathering environment on the dissolved load at the time of peak discharge. Journal of Geophysical Research, 88, 9671-9688.

Nutrient dynamics in Amazon shelf waters

289

Turner R. E., N. N. Rabalais and Z. N. Zhang (1990) Phytoplankton biomass, production and growth limitation on the Huanghe (Yellow River) continental shelf. Continental Shelf Research, 10,545-571. Van Bennekom A. J., G. W. Berger, W. Helder and R. T. P. De Vries (1978) Nutrient distributions in the Zaire estuary and river plume. Netherlands Journal of Sea Research, 12, 296-323. Van Bennekom A. J. and S. B. Tijssen (1976) Nutrients on and off the Guyana shelf related to upwelling and Amazon outflow. Symposium on Progress in Marine Research in the Caribbean and Adjacent Regions, FA O of the United Nations, 233-253. Williams P. M. (1968) Organic and inorganic constituents of the Amazon River. Nature, 218, 937-938.