Marine Chemistry 73 Ž2001. 253–271 www.elsevier.nlrlocatermarchem
River or mangrove? Tracing major organic matter sources in tropical Brazilian coastal waters Thorsten Dittmar a,) , Ruben ´ Jose´ Lara b, Gerhard Kattner a a
Alfred-Wegener-Institut fur ¨ Polar- und Meeresforschung, Am Handelshafen 12, 27570 BremerhaÕen, Germany b Zentrum fur Fahrenheitstr. 1, 28359 Bremen, Germany ¨ Marine Tropenokologie, ¨ Received 29 November 1999; received in revised form 26 October 2000; accepted 27 October 2000
Abstract The influence of mangrove-fringed tropical estuaries on coastal carbon budgets has been widely recognised. However, a quantitative differentiation between riverine and mangrove-derived inputs to the dissolved ŽDOM. and microparticulate organic matter ŽPOM. pool of these environments has been hitherto not possible. Based on lignin-derived phenols and stable carbon isotopes a chemical signature for mangrove, terrestrial and marine-derived organic matter was established for a mangrove estuary in North Brazil. A mixing model was applied to calculate the contribution of each of the three sources to the DOM and POM pool in the estuary throughout 18 tidal cycles in the course of one year. Best source assignment for POM was reached with the yield of lignin phenols and d13C as paired indicators, while the origin of DOM was best identified by the yield of lignin phenols and the acid to aldehyde ratio of vanillyl phenols. Although only about 6% of the fluvial catchment area is covered by mangroves, their contribution to the estuarine DOM and POM pool generally exceeded several times the terrigenic input from the hinterland. This outwelling of mangrove-derived organic matter was enhanced during the rainy season. DOM and POM were exported from the mangrove to the estuary in similar proportions. Most mangrove-POM was rapidly removed from the water column, while mangrove-DOM behaved conservatively. In contrast, terrestrial DOM was almost entirely removed in the outer part of the estuary, which was accompanied by a concomitant increase in terrestrial POM. This seems to be the result of a geochemical barrier zone for this type of DOM in the estuary. Generally, a high proportion of mangrove-DOM was present in the outer part of the estuary, even at high tide. This indicates DOM outwelling from mangroves in adjacent bays or estuaries and points to similar driving forces controlling this process on a regional scale. Mangroves probably play a more important role than rivers for marine carbon budgets along the North Brazilian coast south of the Amazon estuary. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Mangrove swamps; Dissolved organic matter; Particulate organic matter; Organic tracers; Lignin phenols; Mixing model; Outwelling
1. Introduction
)
Corresponding author. Tel.: q49-471-4831-1346; fax: q49471-4831-1425. E-mail address:
[email protected] ŽT. Dittmar..
Mangrove forests are highly productive ecosystems, and they fringe about 60–75% of the tropical coasts ŽMacGill, 1958; Clough, 1998.. Plant litter, mainly leaves, represents about one-third of produc-
0304-4203r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 2 0 3 Ž 0 0 . 0 0 1 1 0 - 9
254
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
tion, and up to half this quantity can be exported by mangrove creeks to adjacent waters ŽRobertson et al., 1992.. The export of these large amounts of organic material has a recognisable effect on the food webs in coastal waters Že.g. Odum and Heald, 1975; Alongi et al., 1989; Alongi, 1990.. There is also evidence for a substantial net export of dissolved organic matter ŽDOM., reaching the same order of magnitude as litter export in some mangrove areas ŽTwilley, 1985.. However, no general consensus has hitherto been reached about the role mangroves play for the dissolved and microparticulate carbon budgets. There are very few mangrove areas where quantitative long-term exportrimport balances exist. These are Hinchinbrook Island ŽAustralia. ŽBoto and Bunt, 1981; Boto and Wellington, 1988; Alongi, 1996; Alongi et al., 1998; Ayukai et al., 1998., Rookery Bay ŽFL, USA. ŽTwilley, 1985. and Braganc¸a ŽNorth Brazil. ŽDittmar, 1999; Dittmar and Lara, in press-a.. Many inconsistencies amongst the published data may have resulted from methodological differences and from the difficulties in accurately determining material fluxes in mangroves, which is due in large and apparently random tidal-dependent oscillations ŽBoto and Wellington, 1988.. Furthermore, differences among the studied ecosystems such as tidal-range, topography, sediment chemistry or community structure are possible reasons for inconsistencies amongst the export balances ŽAyukai et al., 1998.. Chemical tracers, such as lignin-derived phenols and stable carbon isotopes, have been applied in coastal environments to identify source and fate of DOM and particulate organic matter ŽPOM.. Lignin is a unique tracer for vascular plant material, even suitable to distinguish vegetation types, e.g. between woody angiosperms, gymnosperms or non-woody vascular plants ŽHedges and Mann, 1979.. Therefore, it has been widely used to trace the fate and transport of terrestrial organic matter ŽOM. in rivers and marine environments Že.g. Hedges and Ertel, 1982; Ertel et al., 1984,1986; Hedges et al., 1986; Moran et al., 1991a; Moran and Hodson, 1994; Kattner et al., 1999.. Benner et al. Ž1990. found that lignin-derived phenols are leached in considerable amount from mangrove leaves Ž Rhizophora mangle . during early diagenesis. A high percentage of lignin might therefore be present in mangrove-derived DOM. Moran et
al. Ž1991b. traced DOM from a mangrove swamp ecosystem at the Berry Islands ŽBahamas. by analysis of dissolved lignin-derived humic substances and naturally fluorescing compounds. Stable carbon isotope measurements were applied to trace mangrove-derived detritus in coastal food webs or to study the dynamics of particulate organic carbon in mangrove environments ŽRodelli et al., 1984; Zieman et al., 1984; Lin et al., 1991; Hemminga et al., 1994; Honculada Primavera, 1996; Marguillier et al., 1997.. Rezende et al. Ž1990. proposed a high contribution of marine-derived OM to total outwelling of POM from mangroves in Sepetiba Bay ŽRio de Janeiro, Brazil. and suggested that outwelling may be much less significant than expected by simple mass balance studies. In all these studies no differentiation was made between mangroves and other terrestrial plants. Therefore, a clear chemical assignment of the OM to these sources was hitherto not possible. The objective of the present investigation was to establish a chemical pattern for mangrove-derived OM, with special emphasis on the differentiation between marine, mangrove and other OM of terrestrial origin. The use of only one source indicator is not appropriate to distinguish between three sources. Therefore, a combination of several indicators, stable carbon isotopes, lignin and its parameters, was used to study origin, fate and flux of OM in a north Brazilian mangrove estuary.
2. Methods 2.1. The study area North Brazilian mangroves are among the most extensive in the world. The research area at the Caete´ River, near Braganc¸a, is located approximately 150 km to the south-east of the Amazon Estuary ŽFig. 1.. These mangroves are characterised by well-developed forests with tree heights reaching 20 m and more. The dominant species are Rhizophora mangle, AÕicennia germinans and Laguncularia racemosa. The catchment area of the Caete´ River comprises about 3000 km2 , from which about 6% Ž186 km2 . is covered by mangroves. Upstream, the
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
255
Fig. 1. Location of the study site, the mangrove creek AFuro do MeioB and the Caete´ Estuary near Braganc¸a, North Brazil.
region is characterised by secondary forest, agriculture and cattle farming ŽSchwendenmann, 1998.. The system can be described as a riverinerfringe mangrove, and is tide dominated with strong bidirectional flux, according to the criteria of Lugo and Snedaker Ž1974. and Woodroffe Ž1992.. The major part of the mangrove is inundated only fortnightly during spring tides. Annual average temperature and rainfall are about 25.58C and 2550 mm, respectively, with 75% of the precipitation in the rainy season between January and May ŽINMET, 1992.. In the dry season, actual evapotranspiration exceeds precipitation, whereas during the rainy season strong rainfalls lead to high terrestrial runoff. The Caete´ Estuary is characterised by macro-tides with tidal amplitudes of about 4 m. The high energy of the tidal current leads to a highly dynamic erosion and sedimentation regime, with fast sandbank migration and a rapidly changing coastline. The Caete´ Estuary is shallow even in the mouth with maximum water depths at high tide of about 10 m. Due to these characteristics, the Caete´ Estuary can be described as a well-mixed estuary ŽBurton and Liss, 1976; Olausson and Cato, 1980..
The investigated tidal creek AFuro do MeioB ŽFig. 1. has a length of about 4 km and is near the outer part of the estuary. The sampling site is situated in the central part of the mangrove forest, where the creek has a width of ca. 35 m and a maximum depth of ca. 5 m at high tide and 1 m at low tide. In the forest, surface sediments contain about 70% silt–clay Ždry weight. ŽSchwendenmann, 1998.. The presence of abundant crab holes suggests a high degree of sediment bioturbation. The principle pathway for nutrient and DOM transport from the mangrove to the estuary is porewater flow from the upper sediment horizon to the tidal creek and subsequently to the estuary. Porewater flow and material transport are directly coupled with tidal regime ŽLara and Dittmar, 1999; Dittmar and Lara, in press-b.. 2.2. Sampling and chemical analyses Between July 1996 and August 1997, surface water samples were taken in the creek every three weeks, alternating between neap and spring tides Ž18 campaigns.. Within the estuary, sampling was carried out at three fixed locations every nine weeks Žseven campaigns.. Samples were taken Ž3 l. during
256
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
ebb short before low tide and during flood short before high tide at daytime. Immediately after sampling 250 ml were filtered through precombusted Ž4 h, 4508C. Whatman GFrC filters Žnominal pore size f 1 mm. for the analysis of particulate lignin and dissolved organic carbon ŽDOC.. The filtrate was acidified with HCl to pH s 2 and kept frozen in sealed precombusted glass ampoules Žy208C. prior to analysis of DOC. For the determination of stable carbon isotopes and particulate organic carbon ŽPOC., 250 ml were filtered immediately after sampling through precombusted Ž4 h, 4508C. quartz microfibre filters ŽMunktell MK 360; nominal pore size f 0.5 mm.. All filters were freeze-dried and kept frozen Žy208C. in glass vials until analysis. The remaining 2.5 l of sample were decanted from the settled particulate matter and filtered through a precombusted GFrC filter the day after sampling. These filtered samples Ž2 l. were used for the extraction and analysis of dissolved lignin. DOC was measured by high temperature catalytic oxidation with a Dohrman DOC-190 instrument equipped with a standard catalyst of Al 2 O 3 containing 0.5% Pt ŽStatham and Williams, 1983; Skoog et al., 1997.. The acidified samples ŽpH s 2. were sparged for 3 min with N2 and 100 ml of the sample were injected on the top of the catalyst. Oxygen was used as carrier gas. The evolving CO 2 was purified, dried and quantified by a non-dispersive linearised IR gas analyser. A solution of potassium hydrogen phthalate was used as calibration standard. For accuracy, DOC samples were analysed fivefold including a control standard and a blank every 10 samples. The relative standard deviation was less than 3.5% Ž p s 0.05. for each run, according to the method of Funk Ž1985.. The detection limit was 25 mM C. Lignin-derived polymeric material was isolated from the aquatic matrix by solid phase extraction. The filtered water samples were acidified with HCl to pH s 2 and 2 l were passed through a serial combination of macroporous styrene–divinylbenzene ŽXAD-2 and XAD-4. and acrylic ester ŽXAD-7. resins ŽLara and Thomas, 1994.. The pH-adjustment reduces possible matrix-depending fractionations besides increasing the affinity of carboxyl- and hydroxyl groups to the adsorber material. Adsorbed DOM was eluted with 100 ml of 0.2 N NaOH Žhydrophobic acid fraction, HbA. followed by 100
ml of methanol Žhydrophobic neutral fraction, HbN.. The HbA-fraction was neutralised and buffered at pH s 7 with hydrochloric and boric acid and poisoned with HgCl 2 . The HbA fraction was kept in polyethylene bottles and the HbN-fraction in sealed precombusted ampoules and subsequently stored frozen Žy208C. until analysis. This method exhibits extraction efficiencies for phytoplankton-derived DOC and lignin from natural aquatic matrices of about 65% ŽLara and Thomas, 1994; Lobbes, unpublished data.. For the chemical characterisation of lignin, its macromolecular structure was disrupted by the CuO oxidation procedure established by Hedges and Ertel Ž1982., slightly modified by Lobbes et al. Ž1999.. This oxidation method produces 11 phenolic monomers derived entirely or in part from the lignin polymer. It was performed on duplicates of each individual sample ŽHbA, HbN, GFrC-filters.. The HbN-fraction Žmethanol. was evaporated to 8 ml before oxidation, the other fractions processes directly without preceding steps. Immediately after the CuO oxidation, samples were acidified and extracted three-fold with diethylether, evaporated with N2 to dryness and redissolved in 0.5 ml Milli-Q water for analysis. Chromatography and detection of lignin phenols were carried out after Lobbes et al. Ž1999., whose work is based on the method of Steinberg et al. Ž1984.. Briefly, this method was performed on a Merck-Hitachi HPLC system with diode array detector ŽDAD. using a reversed phase column ŽLichrosphere 100 RP 18, 5 mm particle diameter, 250 mm length, 4 mm inner diameter. and a multi-step solvent gradient system for separation. The phenols were identified by their retention times and UV-absorption spectra between 230 and 370 nm, recorded continuously by the DAD detector. The extinction at 280 nm was generally used for quantification. External standards of the 11 phenols ŽFluka, Switzerland; Aldrich, USA; Sigma, USA. were used for calibration. The detection limits for the individual phenols ranged between 15 and 40 pmol Ž p s 0.05; Funk, 1985.. Considering the different procedural steps, detection limits for the aquatic environmental samples were for the HbA-fraction 0.5 to 1.1 nM, for the HbN-fraction 0.03 to 0.09 nM and for POM 0.1 to 0.4 nM.
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
The relative standard deviation of double determinations was on average 15% reaching in maximum 40%, which is in good agreement with data from the literature ŽHedges and Mann, 1979; Goni ˜ et al., 1993; Goni ˜ and Hedges, 1995; Lobbes et al., 1999.. The standard deviation of the HPLC-quantification was less than 3%. No phenols or other substances with similar retention times were found in HbA, HbN and POM blanks. To check the validity of the method for the different sample types and to quantify losses of the phenols during the different steps of sample preparation, a standard composed of the 11 phenols was added to a subset of samples at different stages of sample preparation. The losses during ether extraction, evaporation of ether and solution in water were low; the recovery of the phenols ranged between 95% and 100% and was independent of the initial sample matrix. No significant effect of the sample matrix on the quality of HPLC separation and detection was found. However, phenol monomers added to samples before the CuO oxidation were lost in different amounts. The recovery efficiency including this step of sample preparation ranged between 25% and 100% for the individual phenols, as also reported by Eggers Ž1994.. According to Standley and Kaplan Ž1998. the best compromise between monomer release from the polymer and monomer destruction is reached after 3 h of oxidation as applied in this study. To check the linearity of the relationship between sample amount and release of phenol monomers, different aliquots of a POM sample Žbetween 2 and 200 mmol C. were subjected to CuO oxidation. Reproducible results within the general average deviation were obtained at sample amounts exceeding 10 mmol OC Ž90 nmol phenolic C.. In our study, the amount of organic carbon per sample always exceeded 50 mmol OC Ž100 nmol phenolic C..
257
The analysis of POC and stable carbon isotopes was carried out with a Europa Scientific ANCA 20–20 stable isotope analyser. GFrC filters containing POM were acidified with 0.1 N HCl to remove inorganic carbon, dried at 608C Ž12 h. and put into Sn-vials which were completely oxidised by flash combustion at temperatures ) 10008C under pure O 2 . The isotope composition was analysed by mass spectrometry. Results were normalised to the Pee Dee Belemnite ŽPDB. standard and expressed as d13 C ŽFry and Sherr, 1984.. Samples were analysed in duplicate including a control standard and a blank every four samples. Urea was used as external standard substance. The amount of 12 C and 13 C for each sample was within the analytical linearity range. POC was calculated as the sum of all carbon isotopes. The relative standard deviation of d13 C and POC between the duplicates never exceeded 3%. 2.3. A three-source mixing model for estuarine organic matter Based on the pattern of chemical indicators, a three-source mixing model was established in this study for marine, mangrove and purely terrestrially derived OM in the estuary. Other OM sources, in particular autochthonous aquatic primary production, probably play a subordinate role. Phytoplankton contributed only ca. 10% to POC on annual average. Chlorophyll a concentrations were generally low Ž2.5 mg P ly1 on annual average., and light penetration was only a few cm in the turbid estuarine water ŽSchories et al., in press.. Therefore, aquatic primary production was not considered as an additional end member in the model. Different lignin compositional parameters as defined in Table 1 were used. Mass ratios, e.g. the
Table 1 Definitions of lignin parameters Parameter
Definition
SrV CrV ŽAdrAl.v X lignin L8
molar ratio of total syringyl phenols to vanillyl phenols molar ratio of total cinnamyl phenols to vanillyl phenols molar ratio of vanillic acid to vanillin sum of vanillyl, syringyl and cinnamyl phenols in mmol carbon per mol organic carbon ŽDOC, POC. in sample Ž‰. sum of vanillyl, syringyl and cinnamyl phenols in mg phenol per 100 mg organic carbon ŽDOC, POC. in sample Ž%.
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
258
commonly used parameter L8 , are less appropriate for model calculations. Therefore, the molar ratio X lignin is introduced in this study. The lignin parameters and stable carbon isotope composition of OM derived from the three sources were compared and characteristic chemical patterns identified. Then, using a combination of two indicators, the proportion of each source in sample-OM was calculated, based on the geometrical approach shown in Fig. 2. Samples presenting indicator values within the grey-shaded area can be considered as a mixture of the three OM sources, and the contribution of each source to the sample OM can be estimated. Samples with values on the lines connecting the source reference points can be considered as a mixture of only two types of OM. Indicator values outside the greyshaded area indicate the presence of other sources or processes beyond the scope of this model. Taking into account random errors, introduced by sampling and quantification of lignin and stable carbon isotope composition, a tolerance interval Žt . was established as:
ts
aX a
= 100% s
bX b
= 100% s
cX c
= 100%
t-tolerance interval Ž%.; a, b, c variables characterising the strict validity of the model ŽFig. 2.; aX , bX , cX variables characterising the expanded validity of the model ŽFig. 2..
The variables aX , bX and cX Žsee Fig. 2. depend directly on the degree by which the validity of the mixing model is expanded. The size of the tolerance interval Žt . can be defined before each calculation. Samples with indicator values outside the strict validity of the model Žgrey-shaded area., but within the tolerance Žwhite rim around grey area., were handled in the same way as samples on the connecting lines between two sources Žmixture of OM from only two sources.. Samples with compositional pattern within the small trapezia at the reference points of pure source material, but outside the grey-shaded area, were attributed completely to the source at the trapezium. OM derived entirely from marine primary producers does not contain lignin, and therefore lignin parameters ŽTable 1. are not defined. For this case, the mixing model was slightly modified. The proportion between the sources B and C in the sample was first determined. Then, the proportion between A and the sum of B and C was calculated. Based on these two proportions, the contribution of each of the three sources to the sample was calculated. The tolerance interval Žt . was applied as described above. These calculations were implemented in a VBA module for MS Excel 97.
3. Results and discussion 3.1. Source indicators
Fig. 2. Scheme of a three-source mixing model for OM.
The combined phenolic CuO oxidation products accounted for less than 1% of the DOC and POC. Highest phenol yields were found in OM at Station 1 in the Caete´ Estuary. OM samples taken at the stations in the mangrove, in the middle and the mouth of the estuary exhibited generally less than half of these yields ŽTable 2.. The HbA-fraction contained about 75% of dissolved phenols. The HbN-fraction provided more than half of syringaldehyde, less than one third of vanillyl phenols and only ca. 10% of p-hydroxyl phenols ŽTable 2.. For the identification of OM sources, the concentrations of both DOM fractions were summed. The following criteria were applied to establish a set of organic
484 " 192 507 " 329 86 " 12
167 " 45 164 " 49 785 " 134 221 " 157 127 " 28 678 " 217 14 " 2 13 " 5 30 " 3 Identical to ATidal Creek LTB
POM DOM HbA DOM HbN POM, DOM
542 " 103 417 " 122 64 " 14
59 " 15 120 " 41 449 " 72 379 " 144 108 " 91 141 " 104 64 " 51 121 " 74 72 " 23 71 " 25 193 " 68 203 " 114 128 " 93 42 " 17 42 " 18 49 " 17 17 " 7 17 " 11 43 " 15 40 " 14 49 " 22 21 " 9 17 " 10 15 " 11
9
VON
196 " 86 187 " 54 757 " 76 622 " 311 164 " 128 188 " 88 123 " 74 177 " 104 77 " 26 86 " 31 360 " 130 363 " 192 346 " 329 86 " 44 107 " 57 71 " 32 22 " 8 25 " 10 50 " 21 48 " 27 48 " 34 14 " 4 16 " 7 12 " 5 1079 " 348 748 " 162 96 " 25
3
VAD
1045 " 291 197 " 146 81 " 8
294 " 88 519 " 192 1005 " 261 995 " 637 440 " 453 660 " 541 243 " 185 445 " 490 68 " 26 37 " 13 47 " 38 73 " 81 72 " 96 35 " 36 42 " 35 29 " 22 80 " 33 45 " 16 48 " 19 51 " 17 96 " 45 40 " 15 35 " 22 30 " 17
8
SAL
943 " 21 - dl 36 " 5
180 " 54 255 " 105 665 " 253 552 " 326 175 " 138 252 " 199 141 " 78 216 " 227 - dl - dl - dl - dl - dl - dl - dl - dl - dl - dl 15 " 15 16 " 14 26 " 28 14 " 12 5 " 10 - dl
11
SON
908 " 4 - dl 44 " 2
140 " 39 176 " 72 690 " 269 565 " 311 163 " 147 209 " 173 83 " 50 101 " 160 35 " 21 1"2 38 " 45 9 " 13 44 " 41 - dl 3"6 3"6 17 " 16 5"4 24 " 12 25 " 10 34 " 17 12 " 4 11 " 7 7"3
4
SAD
400 " 38 100 " 39 5"1
73 " 16 90 " 25 335 " 101 249 " 107 79 " 67 96 " 68 53 " 26 89 " 47 80 " 18 83 " 22 93 " 33 101 " 31 87 " 37 53 " 25 65 " 33 67 " 36 6"2 7"3 5"2 4"2 6"2 6"1 2"1 3"2
7
CAD
266 " 5 - dl 3"0
36 " 13 48 " 24 216 " 123 152 " 112 44 " 37 52 " 53 23 " 15 17 " 24 - dl - dl - dl - dl - dl - dl - dl - dl - dl - dl 1"1 1"1 2"2 1"1 0"0 - dl
10
FAD
)
433 " 228 891 " 54
)
)
)
)
)
)
)
)
267 " 53 159 " 37 500 " 130 336 " 31 302 " 56 238 " 47 235 " 55 125 " 31 478 " 56 236 " 25 387 " 116 436 " 118 313 " 72 264 " 65 350 " 120 239 " 27
y27.4 " 0.8 – –
y28.1 " 1.5 y24.0 " 1.1 y27.0 " 0.6 y26.5 " 0.6 y25.3 " 1.1 y23.6 " 1.6 y23.1 " 2.2 y22.7 " 1.2 – – – – – – – – – – – – – – – –
DOC ) , POC d 13 C
4 4 4
15 16 5 5 5 4 6 4 17 17 5 5 5 5 6 6 17 17 5 5 5 5 6 6
n
Abbreviations: PAL, p-hydroxybenzaldehyde; PON, p-hydroxyacetophenone; PAD, p-hydroxybenzoic acid; VAL, vanilin; VON, acetovanillone; VAD, vanilic acid; SAL, syringaldehyde; SON, acetosyringone; SAD, syringic acid; CAD, p-coumaric acid; FAD, ferulic acid; LT, low tide; HT, hight tide; dl, detection limit; –, not determined; ) , DOC values refer to bulk DOM.
Terrestrial Terrestrial Terrestrial Mangrove
135 " 46 330 " 116 631 " 164 607 " 413 228 " 219 369 " 295 168 " 155 355 " 234 118 " 37 127 " 38 248 " 92 289 " 168 258 " 230 87 " 23 120 " 59 107 " 51 57 " 22 55 " 18 69 " 27 69 " 20 88 " 31 44 " 11 41 " 19 41 " 23
318 " 172 306 " 101 637 " 121 633 " 234 198 " 166 275 " 118 194 " 83 366 " 117 283 " 71 329 " 76 549 " 132 581 " 122 491 " 241 246 " 44 287 " 86 239 " 53 13 " 7 26 " 11 48 " 24 42 " 11 60 " 16 56 " 25 35 " 14 49 " 18
27 " 7 40 " 10 154 " 22 133 " 39 32 " 25 41 " 22 30 " 12 43 " 25 71 " 17 56 " 12 86 " 22 101 " 29 66 " 34 36 " 6 40 " 14 30 " 9 6"2 5"1 11 " 4 12 " 5 14 " 3 8"2 6"3 5"2
127 " 41 229 " 131 241 " 92 267 " 131 95 " 91 180 " 95 92 " 63 213 " 87 232 " 53 215 " 41 293 " 149 302 " 88 326 " 82 216 " 108 219 " 92 197 " 69 28 " 7 37 " 11 20 " 9 19 " 8 28 " 5 22 " 9 17 " 7 17 " 5
Tidal Creek
6
VAL
LT HT Estuary St. 1 LT HT Estuary St. 2 LT HT Estuary St. 3 LT HT DOM HbA Tidal Creek LT HT Estuary St. 1 LT HT Estuary St. 2 LT HT Estuary St. 3 LT HT DOM HbN Tidal Creek LT HT Estuary St. 1 LT HT Estuary St. 2 LT HT Estuary St. 3 LT HT
POM
1
PAD
5
PON
2
HPLC-order
PAL
Table 2 Concentrations of phenolic CuO oxidation products in ppm of total organic carbon Žm mol phenolic C per mol DOC or POC, respectively .. DOC and POC concentrations Žm M . and d 13 C values Ž‰ .. Average values, confidence intervals Ž p s 0.05 . and number of samples Ž n . for 1 year at each sampling station at low ŽLT . and high tide ŽHT .
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271 259
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
260
tracers for mangrove, terrestrial and marine origin for the three-source mixing model. 3.1.1. MangroÕe-deriÕed OM The compositional pattern of OM in the tidal creek Furo do Meio at low tide was used to trace mangrove-OM in the estuary. Since the compositions of lignin phenols and stable carbon isotopes in fresh mangrove leaf litter markedly change during decomposition they are not suited to trace OM flux ŽDittmar and Lara, in press-c.. An AidealB organic tracer must be chemically stable on the temporal and spatial scale of interest Že.g. Reeves and Preston, 1989.. During ebb, nutrient and OM-rich porewater flows from the mangrove to the tidal creek leading to strongly enhanced concentrations in its water column. DOC reached values exceeding 600 mM ŽLara and Dittmar, 1999.. This aquatic OM is primarily derived from mangrove leaves in a highly degraded, recalcitrant state ŽDittmar, 1999.. No evidence was found for seasonal or other temporal variations of its compositional pattern except for some days with strong rainfall, when elevated X lignin-values were measured in POM. This was probably caused by enhanced erosion of surface sediments containing less decomposed OM. The annual averages of the source indicators at low tide in Furo do Meio were used as reference points for mangrove-OM in the three-source mixing model ŽTable 3.. The outliers for the POM composition during the rainy season were excluded. 3.1.2. Terrestrial OM Samples representing maximum terrestrial influence Žsalinitys 0. were taken in the innermost part
of the Caete´ Estuary ŽFig. 1, Station 1. at low tide. Only few kilometres seawards from this station, the mangrove-fringed part of the estuary begins and even at low tide a considerable influence of this mangrove area might be present. In March 1997, heaviest rainfalls were accompanied by lowest estuary salinities and strong terrestrial runoff characterised by high OM content ŽDOC ; 900 mM.. The water flow at Station 1 was then directed seawards even during flood, until about 1 h after low tide. This indicates a much higher water flow during ebb than during flood, and therefore the influence of the adjacent mangrove area was probably negligible at low tide. We assume that the compositional pattern of riverine OM did not change considerably throughout the year, as suggested by Ertel et al. Ž1986. and Hedges et al. Ž1986. for the Amazon River. The values of the OM tracers from March 1997 Žlow tide, Station 1. were considered to be characteristic for aquatic OM derived from terrestrial plants Žother than mangroves.. These values were used as reference points for pure terrestrial OM in the estuary. Terrestrial DOM was characterised by a higher proportion of lignin-derived phenols Ž X lignin . than mangrove-derived DOM ŽTable 3. probably due to the relatively low content of lignin in mangrove leaves ŽBenner et al., 1990; Dittmar, 1999.. Raw HPLC-DAD chromatographic data are shown in Fig. 3 to illustrate the compositional differences in CuO oxidation products from mangrove- and terrestrially derived organic matter. In particular vanillic acid Žpeak no. 3 in Fig. 3. exhibited characteristic high yields from both fractions of terrigenous DOM ŽHbA and HbN.. The high acid to aldehyde ratios within
Table 3 Chemical source indicators Žreference points. for marine, terrestrial and mangrove-derived DOM and POM in the Caete´ Estuary. Average values, confidence intervals Ž p s 0.05. and number of samples Ž n.. Other lignin parameters effected by concentrations frequently near or below the detection limit were not taken into account. The ratios were calculated from the summed concentrations of the HbA and HbN fractions for each sample and then averaged
DOM
POM
Source
SrV
CrV
ŽAdrAl.v
X lignin Ž‰.
L8 Ž%.
d13 C Ž‰.
n
Marine Terrestrial Mangrove Marine Terrestrial Mangrove
– 0.17 " 0.00 0.46 " 0.09 – 1.24 " 0.25 1.51 " 0.29
– 0.05 " 0.00 0.25 " 0.05 – 0.29 " 0.11 0.28 " 0.08
– 1.64 " 0.16 0.60 " 0.12 – 0.89 " 0.25 1.30 " 0.48
0 2.38 " 0.83 0.65 " 0.20 0 5.67 " 0.91 1.11 " 0.30
0 0.40 " 0.14 0.11 " 0.03 0 0.97 " 0.16 0.19 " 0.05
– – – y20.8 " 1.1 y27.4 " 0.8 y28.1 " 1.5
4 4 17 4 4 15
–, not defined or not measured.
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
the lignin phenol families in terrestrial DOM ŽŽAdrAl.v f 1.5. indicate a high degree of aerobic
261
degradation, whereas mangrove-derived DOM was evidently influenced by anaerobic degradation ŽDittmar and Lara, in press-c., which led to low ŽAdrAl.v values of ca. 0.6. These different degradation processes may also influence the other lignin parameters, SrV and CrV, which are considerably lower in terrestrial DOM than in mangrove-derived DOM. Due to these clear differences, the lignin parameters can be used as source indicators for DOM. Similarly to DOM, terrestrial POM was characterised by much higher X lignin -values than mangrove-derived POM. The other lignin parameters and d13 C, however, were not statistically different from the mangrove-POM. Thus, they were not suitable as source indicators to distinguish between mangrove and terrestrial-derived POM. 3.1.3. Marine OM For an estimate of the compositional pattern of ApureB marine OM, high-tide samples Žsalinity of 30–36. from the mouth of the estuary during the dry season and at the beginning of the rainy season were selected. Then, linear regressions were calculated between salinity and X lignin and d13 C, respectively Ž r ) 0.995, p - 0.001, n s 4.. According to the regression equations, a AstandardB seawater salinity of 35 resulted in an X lignin of zero and a d13 C ratio of POM of y20.8‰. This is in good agreement with values reported in the literature as typical for OM of solely marine production Že.g. Fry and Sherr, 1984; Meyers-Schulte and Hedges, 1986.. The three-source mixing model was applied to calculate the contribution of each primary source to the OM pool in the estuary based on the data shown in Table 3. The mixing model for POM was calibrated with X lignin and d13 C ŽFig. 4.. About 80% of all POM samples can be explained as a mixture of mangrove, terrestrial and marine origin using a tolerance of t s 25%. Accepting t s 40%, the validity of
Fig. 3. HPLC chromatograms Žabsorbance at 280 nm. of CuO oxidation products of mangrove and terrestrial-derived DOM ŽHbA and HbN fractions.. Example of a DAD chromatogram Žabsorbance spectra, 230–340 nm.. Peaks were identified by the DAD spectra. For the quantification of the individual phenols, different wavelengths were used. The corresponding names of the phenols are given in Table 2.
262
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
Fig. 4. Three-source mixing model for POM using X lignin and d13 C as tracers for mangrove, marine and terrestrial-derived POC. Reference points for each source and average values for the samples taken in the Caete´ Estuary and the tidal creek Furo do Meio at low and high tide, with confidence intervals Ž p s 0.05..
the model is extended to 91% of the POM samples. Taking into account the probably high variability mainly introduced by POM sampling, t s 40% might be a more realistic value to describe methodological random errors. The best results for DOM source identification were achieved using X lignin and ŽAdrAl.v as source tracers: 94% and 100% of the samples presented compositions, which could be explained as mixtures of the three DOM types, accepting t s 25% and t s 30%, respectively. The combination of X lignin and SrV allowed the composition of 90% and 96% of all DOM samples to be explained. This percentage was reduced to only about 80% using the combination of X lignin and CrV. An additional, non-lignin source for p-coumaric acid as part of a labile DOM pool might be the reason for this relatively weak validity for the model using CrV as source tracer ŽHartley, 1973; Kirk et al., 1980; Dittmar, 1999.. In Fig. 5, the two models for DOM providing the best
results are presented. The phenol ratios used as end members in the model calculations do not mix linearly with regard to bulk OM. Therefore, the contribution of the different OM sources to the lignin pool was calculated as a first step, and secondly these values were transformed in OC proportions by the use of X lignin values. The non-linear character of this mixing is also reflected in Fig. 5. 3.2. Contribution of the different sources: annual aÕerage 3.2.1. Particulate organic carbon The use of chemical source indicators and the application of a three-source mixing model revealed a clear picture of the origin of OC in the Caete´ Estuary. The concentration of mangrove-derived POC reached a maximum in the middle of the estuary at Station 2, exceeding the amount of terrestrial POC
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
263
Fig. 5. Two alternative three-source mixing models for DOM, using X lignin , ŽAdrAl.v and SrV as tracers for mangrove, marine and terrestrial-derived DOC. Reference points for each source and average values for the samples taken in the Caete´ Estuary and the tidal creek Furo do Meio at low and high tide, with confidence intervals Ž p s 0.05.. The shown values were calculated from summed concentrations of the HbA and HbN fractions.
ŽFig. 6.. The concentration of mangrove-POC increased at the mangrove-fringed part of the estuary ŽStations 2 and 3. at low tide, whereas at the innermost part of the estuary ŽStation 1., a slight increase of the mangrove-POC fraction occurred at high tide.
Fig. 6. Origin of POC in the tidal creek Furo do Meio and the Caete´ Estuary. Average concentrations of the whole sampling period at low ŽLT. and high tide ŽHT. split into three OM sources. d13 C and X lignin were used as source indicators.
These findings strongly indicate a POC outwelling from the whole mangrove area to the estuary. During flood, part of this mangrove-POC, introduced into the estuary during ebb, reached the estuary’s head. A considerable part of the POC pool at Station 1 therefore consisted of mangrove-POC, mainly at high tide ŽTable 4.. In the estuary’s mouth ŽStation 3. the influence of marine-derived POC increased. A remarkable feature is that in the middle and the mouth of the estuary terrestrial POC showed higher concentrations at high tide than at low tide, contrary to the general decrease of POC concentration at high tide. An inflow of terrestrial OC into this mangrovefringed part of the estuary during flood is very unlikely. A better explanation is the transformation of terrestrial DOM into POM and the existence of a geochemical barrier zone in the estuary for terrestrial DOM. The strong decrease of the terrestrial DOC concentration in this part of the estuary from low to high tide supports this hypothesis ŽFig. 7.. Lapin et al. Ž1990. observed in an estuary in Japan that most of the terrigenous DOM was deposited or altered in a zone with a salinity from 1.3 to 24.0, similar to the conditions in the Caete´ Estuary. Such a geochemical barrier zone for terrigenous DOM has been found in
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
264
Table 4 The three OM sources given as percentage of the OM pool in the tidal creek Furo do Meio and the Caete´ Estuary at low ŽLT. and high tide ŽHT.. Average values of the whole sampling period, confidence intervals Ž p s 0.05. and number of samples Ž n.. Different combinations of lignin parameters as source indicators were alternatively used for DOM Source Ž%.
Creek
Estuary Station 1
POM Ž X lignin and d13 C. DOM Ž X lignin and ŽAdrAl.v. DOM Ž X lignin and SrV. DOM Ž X lignin and CrV. n
Marine Terrestrial Mangrove Marine Terrestrial Mangrove Marine Terrestrial Mangrove Marine Terrestrial Mangrove
Station 2
Station 3
LT
HT
LT
HT
LT
HT
LT
HT
13 " 10 16 " 12 72 " 18 29 " 13 1"1 70 " 13 31 " 14 4"2 66 " 15 32 " 11 7"7 61 " 11 17
47 " 11 25 " 9 28 " 15 29 " 10 1"1 70 " 10 65 " 8 19 " 7 15 " 7 38 " 10 7"5 54 " 11 17
7"8 76 " 16 17 " 15 10 " 14 46 " 37 43 " 34 21 " 21 70 " 29 9"9 23 " 18 40 " 21 38 " 30 6 ŽPOM 5.
13 " 6 58 " 33 30 " 32 10 " 9 43 " 34 47 " 29 6 " 11 89 " 20 5"9 20 " 20 38 " 26 42 " 29 6 ŽPOM 5.
32 " 12 19 " 24 50 " 33 14 " 20 25 " 38 61 " 34 27 " 34 24 " 20 49 " 32 22 " 26 35 " 32 43 " 25 5
44 " 19 27 " 20 29 " 32 43 " 17 5"7 52 " 20 53 " 25 9"4 38 " 27 49 " 21 7"5 43 " 23 5
59 " 16 11 " 11 30 " 16 35 " 29 5"5 61 " 30 44 " 33 24 " 30 32 " 32 47 " 25 9"8 44 " 26 6
63 " 15 24 " 16 13 " 9 36 " 27 1"2 63 " 28 70 " 15 14 " 8 16 " 17 46 " 25 5"4 49 " 26 6
several estuaries Že.g. Burton and Liss, 1976; Sholkovitz, 1976; Olausson and Cato, 1980; Libes, 1992.. Terrestrial DOM yielded less phenols than terrestrial POM ŽTable 3.. The suggested transformation would therefore imply a change in X lignin of the
terrestrial end member in the POM mixing model. Therefore, the proportions of terrestrial POM are probably underestimated at Stations 2 and 3 and should be interpreted as a lower limit. A selective aggregation of DOM fractions due to physicochemical alterations in the estuary would also lead to
Fig. 7. Origin of DOC in the tidal creek Furo do Meio and the Caete´ Estuary. Average concentrations of the whole sampling period at low ŽLT. and high tide ŽHT. split into three OM sources. Different combinations of lignin parameters as source indicators were alternatively used.
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
shifts of the terrestrial end member in the DOM mixing-model. However, a detailed examination of the composition of HbA and HbN-DOM indicates that selective aggregation of terrestrial DOM fractions did not take place in the estuary. Both DOM fractions, HbA and HbN, showed the same characteristic patterns for terrigenous and mangrove-derived organic matter, e.g. the high yield of vanillic acid Žpeak no. 3 in Fig. 3. in terrestrial DOM. During mixing in the estuary, these patterns changed concurrently in HbA and HbN accordingly to the calculated proportion of the two DOM sources. It is very unlikely that these fractions, with their different physicochemical characteristics, experienced the same fractionations of their phenol composition. Fractionations between HbA and HbN would more likely occur, but also no evidence was found for selective removal of HbA or HbN. The contribution of each of the two fractions to total dissolved lignin phenols did not change significantly in the course of the estuary. Mangrove-derived OM, on the other hand, was formed under physicochemical conditions in the mangrove sediment, which are close to that of the estuarine water column. Alterations of the mangrove end member are therefore not to be expected. Microbial OM degradation, which would lead to increased ŽAdrAl.v values ŽCrawford, 1981; Ertel and Hedges, 1984., was not observed and might therefore not influence the end members on the estuarine scale. At the tidal creek Furo do Meio, the influence of mangrove-derived POC dominated at low tide, whereas at high tide the three fractions were present in similar proportions ŽFig. 6.. On annual average, mangrove-POC did not reach 100% of total POC at low tide, as would be expected according to the definition of mangrove-derived OM given above. The reason for this is that the percentages of the different sources Žmaximum 100%. were calculated individually for each sampling day and then these percentages were averaged over the whole sampling period.
3.2.2. DissolÕed organic carbon All three combinations of source indicators consistently point to an outwelling of DOC from the mangrove. As shown in Fig. 7 and Table 4, a
265
considerable part of DOC in the estuary was derived from the mangrove. In contrast to POC, in the mouth of the estuary ŽStation 3. a relatively high proportion of mangrove-DOC was present even at high tide. Generally, the differences between low and high tides were much lower than those for POC. These inferences, which were independent of the combination of source indicators, show that the mangroves contributed significantly not only to the estuarine, but also to the near-shore DOC pool. This was most probably the reason for the high proportion of mangrove-derived DOC in the water flowing at high tide into the estuary. Thus, a substantial fraction of the estuarine DOM pool probably originates from adjacent mangrove areas, leading to the similar proportion of mangrove-DOC at both low and high tide in the creek ŽTable 4.. Detailed examination of the different combinations of source indicators revealed some different patterns: Calculations with X lignin and ŽAdrAl.v as source indicators resulted in almost identical percentages of mangrove-derived DOC at low and high tide in Furo do Meio ŽTable 4.. Nevertheless, the use of X lignin and SrV suggested a higher contribution of mangrove-DOC at low tide, as expected from the higher porewater input. In the estuary, the proportion of mangrove-DOC was generally similar at low and high tide, independent of the combination of source indicators. We infer that the clear decrease of terrestrial DOC along the estuary and from low to high tide is due to the transformation into POC, and not only due to increasing dilution with marine water. Mixing diagrams Žsalinity vs. concentration plots. for the Caete´ estuary also suggest non-conservative characteristics of terrigenous DOC ŽDittmar and Lara, in press-a., supporting the speculations based on the present data set. In contrast to the non-conservative terrestrial DOC, mangrove-DOC was transported longer distances, indicating more refractory and conservative behaviour. Therefore, we deduced that the mangroves in this region export more DOC to near-shore areas than the hinterland, despite the fact that only ca. 6% of the catchment area of the Caete´ River is covered by mangroves. This outwelling of mangrove-derived DOC is on annual average about 10 mmol C per m2 mangrove and day, as shown by Dittmar Ž1999. and Dittmar and Lara Žin press-a. in a long-term balance study at the Caete´ Estuary.
266
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
In order to assess whether OM was preferentially exported from the mangrove in particulate or dissolved form, mangrove-derived OC concentrations at different stations and tide situations were compared. At low tide, mangrove-POC was similar to DOC in the estuary and mangrove, suggesting an export of POC and DOC on a similar scale. However, in the
mouth of the estuary at high tide, mangrove-POC was much lower than DOC, indicating that POC was rapidly removed from the water column and only DOC was exported from the estuary to the coastal waters in significant amounts. Thus, the export of mangrove-POM might have a rather local impact, e.g. on the planktonic or benthic community or the
Fig. 8. Origin of POC in the tidal creek Furo do Meio and the Caete´ Estuary throughout 12 months of sampling. Concentrations at low ŽLT. and high tide ŽHT. related to three OM sources. For clarity, the concentration axis is inverted at low tide, representing outflow. d13 C and X lignin were used as source indicators.
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
sedimentary composition, whereas the export of mangrove-DOM might influence the OM pool in near and even offshore regions. 3.3. Contribution of the different sources: annual trends 3.3.1. Particulate organic carbon During the rainy season an enhanced outwelling of mangrove-derived POC was evident from its in-
267
creased concentration at low tide in the mangrovefringed part of the estuary ŽFig. 8.. In the middle of the estuary ŽStation 2. mangrove-POC was then elevated even at high tide. This increased export might be due to the erosive force of the rain itself and the lower particle cohesion at the sediment surface softened by the high water content during the rainy season. Despite this outwelling, mangrove-POC was consistently negligible at high tide in the mouth of the estuary. This indicates that the mangroves did not
Fig. 9. Origin of DOC in the tidal creek Furo do Meio and the Caete´ Estuary throughout 12 months of sampling. Concentrations at low ŽLT. and high tide ŽHT. related to three OM sources. For clarity, the concentration axis is inverted at low tide, representing outflow. X lignin and ŽAdrAl.v were used as source indicators.
268
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
contribute considerably to the near-shore POM pool, even in the rainy season. At the estuary’s head ŽStation 1. a considerable influence of mangrovePOC was only observed at the end of the rainy season, probably due to still enhanced mangrovePOC in the estuary and an already decreased terrestrial runoff. At Furo do Meio the calculated concentrations of mangrove-derived POC at low tide decreased to a minimum in the rainy season, although an increase proportional to total POC would be expected ŽFig. 8.. During rainfall, enhanced erosion within the mangrove caused remarkable maxima of total POC concentration Žup to 59 mM. in the creek ŽDittmar and Lara, in press-b.. This might lead to a mobilisation of relatively fresh sedimentary OM, which was accumulated during the dry season, such as less degraded leaf litter and decomposition products of wood with higher lignin contents than the bulk of sedimentary OM. This could cause the high X lignin values and lead occasionally to modifications of the mangrove end member in the POM mixing model at Furo do Meio. During rainy events, the model probably overestimates the terrestrial contribution to POM at low tide in the tidal creek. This seems to be only a local phenomenon at Furo do Meio, since it was not observed in the estuary. 3.3.2. DissolÕed organic carbon The annual trend of the mangrove-derived DOC concentration at low tide in the mangrove-fringed part of the estuary ŽStations 2 and 3. shows an increased outwelling of DOC from the mangrove during the rainy season, similar to POC ŽFig. 9.. Maxima of mangrove-DOC occurred simultaneously in the estuary and in the tidal creek at low tide. Besides further supporting the existence of a significant outwelling of mangrove-DOC, this shows that the processes involved were relatively stable and did not change within several days. A main driving force for the seasonality of outwelling might be enhanced water exchange between porewater and tidal creek during the rainy season. In contrast to POM, the pattern of mangrove-derived DOM did not change considerably due to strong rainfalls, which indicates that DOM was continuously released from the same sources within the mangrove and introduced into the tidal creek in a similar stage of degradation. At
Station 1 the increased terrestrial runoff during the rainy season led to a predominance of terrestrial DOC, whereas during the dry season, mangrove-DOC contributed a major part to the DOC pool even at the estuary’s head. At high and low tide, mangrove-DOC showed the same annual trend along the whole estuary, even at the outermost Station 3. This further points to a conservative behaviour of mangrove-derived DOC and suggests that adjacent mangrove areas undergo a similar annual trend of DOC outwelling as in our study region.
4. Conclusions Based on lignin-derived phenols and stable carbon isotopes a chemical signature for mangrove, terrestrial and marine-derived OM was established. Outwelling of DOM and POM from the mangrove was evident from chemical tracers and generally exceeded the terrigenous input from the river’s catchment area by several times. DOM and POM were exported from the mangrove to the estuary in similar proportions. A considerable amount of mangrovePOM was rapidly removed from the water column, while mangrove-DOM could be transported over long distances, indicating conservative and recalcitrant properties. In contrast, terrestrial DOM was almost entirely removed from the water column, which seems to be result of a geochemical barrier zone for this type of DOM in the estuary. It is likely that mangroves in adjacent bays or estuaries exhibit export characteristics similar to the mangroves at the Caete´ River. Since similar geomorphological features caused the development of analogous ecosystems along the entire coastline Southeast of the Amazon estuary, it can be assumed that mangroves play a key role for the aquatic carbon budgets of coastal and marine ecosystems in the whole region. This comprises a mangrove belt of about 6700 km2 ŽHerz, 1991.. Based on these geographical features and on the present work, flux values for the Braganc¸a mangroves ŽDittmar, 1999. were extrapolated to this area. This resulted in an estimated outwelling of mangrove-DOC of about 2.7 = 10 10 mol C per year for the coastline of the Brazilian states Para´ and Maranhao. ˜
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
Mangroves may considerably influence marine carbon cycles and budgets on a scale far beyond the direct vicinity of the mangroves itself. On a global scale, mangroves may contribute about 4% Ž7 = 10 11 mol C per year. of the globally averaged land-derived DOC flux, if similar export characteristics for mangroves in general are assumed ŽMeybeck, 1982; Spalding et al., 1997; Dittmar, 1999.. The phenomenon of coastal outwelling may therefore considerably effect marine DOC cycles even on a global scale and must therefore be considered in models of global biogeochemical cycles. The chemical tracers and data analysis established in this study appear to be a suitable tool for the assessment of the fate of mangrove-OM even in offshore marine environments.
Acknowledgements We thank J. Lobbes for the introduction to the lignin analysis. We appreciate the excellent technical assistance of K.-U. Richter and M. Birkicht. We are also grateful to T. Eglinton and three anonymous reviewers for detailed and constructive comments on the manuscript. This study was carried out as part of the Brazilian–German co-operation project MADAM and was supported by the Brazilian National Research Council ŽCNPq. and the German Ministry for Education and Research ŽBMBF. under project code 03F0154A. This is MADAM-Contribution No. 25.
References Alongi, D.M., 1990. Abundances of benthic microfauna in relation to outwelling of mangrove detritus in a tropical coastal region. Mar. Ecol. Prog. Ser. 63, 53–63. Alongi, D.M., 1996. The dynamics of benthic nutrient pools and fluxes in tropical mangrove forests. J. Mar. Res. 54, 123–148. Alongi, D.M., Boto, K.G., Tirendi, F., 1989. Effect of exported mangrove litter on bacterial productivity and dissolved organic carbon fluxes in adjacent tropical nearshore sediments. Mar. Ecol. Prog. Ser. 56, 133–144. Alongi, D.M., Ayukai, T., Brunskill, G.J., Clough, B.F., Wolanski, E., 1998. Sources, sinks, and export of organic carbon through a tropical, semi-enclosed delta ŽHinchinbrook Channel, Australia.. Mangroves Salt Marshes 2, 237–242. Ayukai, T., Miller, D., Wolanksi, E., Spagnol, S., 1998. Fluxes of nutrients and dissolved and particulate organic matter in two
269
mangrove creeks in northeastern Australia. Mangroves Salt Marshes 2, 223–230. Benner, R., Weliky, K., Hedges, J.I., 1990. Early diagenesis of mangroves leaves in a tropical estuary: molecular-level analyses of neutral sugars and lignin-derived phenols. Geochim. Cosmochim. Acta 54, 1991–2001. Boto, K.G., Bunt, J.S., 1981. Tidal export of particulate organic matter from a northern Australian mangrove system. Estuar. Coast. Shelf Sci. 13, 247–255. Boto, K.G., Wellington, J.T., 1988. Seasonal variations in concentrations and fluxes of dissolved organic and inorganic materials in a tropical, tidally-dominated, mangrove waterway. Mar. Ecol. Prog. Ser. 50, 151–160. Burton, J.D., Liss, P.S., 1976. Estuarine Chemistry. Academic Press, London, 229 pp. Clough, B., 1998. Mangrove forest productivity and biomass accumulation in Hinchinbrook Channel, Australia. Mangroves Salt Marshes 2, 191–198. Crawford, R.L., 1981. Lignin biodegradation and transformation. Wiley, New York, 154 pp. Dittmar, T., 1999. Outwelling of organic matter and nutrients from a mangrove in north Brazil: evidence from organic tracers and flux measurements, ZMT-Contribution, 5. Center for Tropical Marine Ecology, Bremen. 229 pp. Dittmar, T., Lara, R.J., in press-a. Do mangroves rather than rivers provide nutrients to coastal environments south of the Amazon River? Evidence from long-term flux measurements. Mar. Ecol. Prog. Ser. Dittmar, T., Lara, R.J., in press-b. Dynamics of inorganic and organic nutrients in a mangrove tidal creek in North Brazil: influence of meteorological, biological and hydrological factors. Estuar. Coast. Shelf Sci. Dittmar, T., Lara, R.J., in press-c. Molecular evidence for lignin degradation in sulfate reducing mangrove sediments ŽAmazonia, Brazil.. Geochim. Cosmochim. Acta. ˆ Eggers, J., 1994. Charakterisierung des Eintrags von organischen Substanzen anhand von Markersubstanzen in marinen ¨ Okosystemen. MSc. Thesis, University of Bremen, 68 pp. Ertel, J.R., Hedges, J.I., 1984. The lignin component of humic substances: distribution among soil and sedimentary humic, fulvic, and base-insoluble fractions. Geochim. Cosmochim. Acta 48, 2065–2074. Ertel, J.R., Hedges, J.I., Perdue, E.M., 1984. Lignin signature of aquatic humic substances. Science 223, 485–487. Ertel, J.R., Hedges, J.I., Devol, A.H., Richey, J.E., 1986. Dissolved humic substances of the Amazon River system. Limnol. Oceanogr. 31, 739–754. Fry, B., Sherr, E.B., 1984. d13 C measurements as indicators of carbon flow in marine and freshwater ecosystems. Bull. Mar. Sci.rContrib. Mar. Sci. 27, 13–47. Funk, W., 1985. Statistische Methoden in der Wasseranalytik. VCH, Weinheim, 225 pp. Goni, ˜ M.A., Hedges, J.I., 1995. Sources and reactivities of marine-derived organic matter in coastal sediments as determined by alkaline CuO oxidation. Geochim. Cosmochim. Acta 59, 2965–2981. Goni, ˜ M.A., Nelson, B., Blanchette, R.A., Hedges, J.I., 1993.
270
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271
Fungal degradation of wood lignins: geochemical perspectives from CuO-derived phenolic dimers and monomers. Geochim. Cosmochim. Acta 57, 3985–4002. Hartley, R.D., 1973. Carbohydrate esters of ferulic acid as components of cell-wall of Lolioum multiflorum. Phytochemistry 12, 661–665. Hedges, J.I., Ertel, J.R., 1982. Characterization of lignin by gas capillary chromatography of cupric oxide oxidation products. Anal. Chem. 54, 174–178. Hedges, J.I., Mann, D.C., 1979. The characterization of plant tissues by their lignin oxidation products. Geochim. Cosmochim. Acta 43, 1803–1807. Hedges, J.I., Clark, W.A., Quay, P.D., Richey, J.E., Devol, A.H., Santos, U.M., 1986. Composition and fluxes of particulate organic matter in the Amazon River. Limnol. Oceanogr. 31, 717–738. Hemminga, M.A., Slim, F.J., Kazungu, J., Ganssen, G.M., Nieuwenhuize, J., Kruyt, N.M., 1994. Carbon outwelling from a mangrove forest with adjacent seagrass beds and coral reefs ŽGazi Bay, Kenya.. Mar. Ecol. Prog. Ser. 106, 291–301. Herz, R., 1991. Manguezais do Brasil. Instituto Oceanografico, ´ Universidade de Sao ˜ Paulo, SP, Brazil, 227 pp. Honculada Primavera, J., 1996. Stable carbon and nitrogen isotope ratios of penaeid juveniles and primary producers in a riverine mangrove in Guimaras, Philippines. Bull. Mar. Sci. 58, 675– 683. INMET, 1992. Normas Climatologicas. Instituto Nacional de Me´ teorologia, Brasilia DF. Kattner, G., Lobbes, J.M., Fitznar, H.P., Engbrodt, R., Nothig, ¨ E.M., Lara, R.J., 1999. Tracing dissolved organic substances and nutrients from the Lena River through Laptev Sea ŽArctic.. Mar. Chem. 65, 25–39. Kirk, T.K., Higuchi, T., Chang, H., 1980. Lignin Biodegradation: Microbiology, Chemistry, and Potential Applications. CRC Press, Boca Raton, FL. Lapin, I.A., Anikiyev, V.V., Vinnikov, Y.Y., Tambiyeva, N.S., Shumilin, Y.N., 1990. Biogeochemical aspects of the behavior of dissolved organic matter in the Razdolnaya river estuary ŽAmor Gulf, Sea of Japan.. Oceanology 30, 170–174. Lara, R.J., Dittmar, T., 1999. Nutrient dynamics in a mangrove creek ŽNorth Brazil. during the dry season. Mangroves Salt Marshes 3, 185–195. Lara, R., Thomas, D.N., 1994. Isolation of marine dissolved organic matter: evaluation of sequential combinations of XAD resins 2, 4 and 7. Anal. Chem. 66, 2417–2419. Libes, S.M., 1992. An Introduction to Marine Biogeochemistry. Wiley, New York, 734 pp. Lin, G., Banks, T., Sternberg, L.S.L.O., 1991. Variation in d 13 C values for the seagrass Thalassia testudinum and its relations to mangrove carbon. Aquat. Bot. 40, 333–341. Lobbes, J.M., Fitznar, H.P., Kattner, G., 1999. High-performance liquid chromatography of lignin-derived phenols in environmental samples with diode array detection. Anal. Chem. 71, 3008–3012. Lugo, A.E., Snedaker, S.C., 1974. The ecology of mangroves. Annu. Rev. Ecol. Syst. 5, 39–64.
MacGill, J.T., 1958. Map of coastal landforms of the world. Geogr. Rev. 48, 402–405. Marguillier, S., Van der Velde, G., Dehairs, F., Hemminga, M.A., Rajagopal, S., 1997. Trophic relationships in an interlinked mangrove–seagrass ecosystem as traced by d 13 C and d 15 N. Mar. Ecol. Prog. Ser. 151, 115–121. Meybeck, M., 1982. Carbon, nitrogen, and phosphorous transport by world rivers. Am. J. Sci. 282, 401–450. Meyers-Schulte, K.J., Hedges, J.I., 1986. Molecular evidence for a terrestrial component of organic matter dissolved in ocean water. Nature 321, 61–63. Moran, M.A., Hodson, R.E., 1994. Dissolved humic substances of vascular plant origin in a coastal marine environment. Limnol. Oceanogr. 39, 762–771. Moran, M.A., Pomeroy, L.P., Sheppard, E.S., Atkinson, L.P., Hodson, R.E., 1991a. Distribution of terrestrially derived dissolved organic matter on the southeastern US continental shelf. Limnol. Oceanogr. 36, 1134–1149. Moran, M.A., Wicks, R.J., Hodson, R.E., 1991b. Export of dissolved organic matter from a mangrove swamp ecosystem: evidence from natural fluorescence, dissolved lignin phenols, and bacterial secondary production. Mar. Ecol. Prog. Ser. 76, 175–184. Odum, E.P., Heald, E.J., 1975. The detritus bases food web of an estuarine mangrove community. In: Cronin, L.E. ŽEd.., Estuarine Research. Academic Press, New York, pp. 265–286. Olausson, E., Cato, I., 1980. Chemistry and Biogeochemistry of Estuaries. Wiley, New York, 452 pp. Reeves, A.D., Preston, M.R., 1989. The composition of lignin in estuarine suspended particulates and the distribution of particulate lignin in estuaries as determined by capillary gas chromatography of cupric oxide oxidation products. Estuar. Coast. Shelf Sci. 29, 583–599. Rezende, C.E., Lacerda, L.D., Ovalle, A.R.C., Silva, C.A.R., Martinelli, L.A., 1990. Nature of POC transport in a mangrove ecosystem: a carbon stable isotopic study. Estuar. Coast. Shelf Sci. 30, 641–645. Robertson, A.I., Alongi, D.M., Boto, K.G., 1992. Food chains and carbon fluxes. In: Robertson, A., Alongi, I., Alongi, D.M. ŽEds.., Tropical Mangrove Ecosystems-Coastal and Estuarine Series 41. American Geophysical Union, Washington, pp. 293–326. Rodelli, M.R., Gearing, J.N., Gearing, P.J., Marshall, N., Sasekumar, A., 1984. Stable isotope ratio as a tracer of mangrove carbon in Malaysian ecosystems. Oecologia 61, 326–333. Schories, D., Kadler, S., Meeßen, I., Mehlig, U., in press. Phytoplankton, chlorophyll a and seston variabilities in a small mangrove tidal creek near Braganc¸a ŽPA, North Brazil.. Wetlands Ecol. Manage. Schwendenmann, L., 1998. Tidal and seasonal variations of soil and water properties in a Brazilian mangrove ecosystem. MSc. Thesis, University of Karlsruhe, 101 pp. Sholkovitz, E.R., 1976. Flocculation of dissolved organic and inorganic matter during the mixing of river water and sea water. Geochim. Cosmochim. Acta 40, 367–375. Skoog, A., Thomas, D., Lara, R., Richter, K.-U., 1997. Method-
T. Dittmar et al.r Marine Chemistry 73 (2001) 253–271 ological investigations on DOC determinations by the HTCO method. Mar. Chem. 56, 39–44. Spalding, M., Blasco, F., Field, C., 1997. World Mangrove Atlas. International Society for Mangrove Ecosystems, Okinawa, Japan. Standley, L.J., Kaplan, L.A., 1998. Isolation and analysis of lignin-derived phenols in aquatic humic substances: improvements on the procedures. Org. Geochem. 28, 689–697. Statham, P.J., Williams, P.J.L., 1983. Determination of total CO 2 . In: Grasshoff, K., Erhardt, M., Kremling, K. ŽEds.., Methods of Seawater Analysis. Verlag Chemie, Weinheim, pp. 378– 383. Steinberg, S.M., Venkatesan, M.I., Kaplan, I.R., 1984. Analysis of the products of the oxidation of lignin by CuO in biological
271
and geological samples by reversed-phase HPLC. J. Chromatogr. 298, 427–434. Twilley, R.R., 1985. The exchange of organic carbon in basin mangrove forest in a southwestern Florida estuary. Estuar. Coast. Shelf Sci. 20, 543–557. Woodroffe, C., 1992. Mangrove sediments and geomorphology. In: Robertson, A.I., Alongi, D.M. ŽEds.., Tropical Mangrove Ecosystems, Coastal and Estuarine Studies 41. American Geophysical Union, Washington, DC, pp. 7–41. Zieman, J.C., Macko, S.A., Mills, A.L., 1984. Role of seagrass and mangroves in estuarine food webs: temporal and spatial changes in stable isotope composition and amino acid content during decomposition. Bull. Mar. Sc. 35, 380–392.