JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 25 May 2007
Published in J Environ Qual 36:963-974 (2007)
DOI: 10.2134/jeq2006.0318
© 2007 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Esslemont, G.
Right arrow Articles by Lawrence, I.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Esslemont, G.
Right arrow Articles by Lawrence, I.
Agricola
Right arrow Articles by Esslemont, G.
Right arrow Articles by Lawrence, I.
Related Collections
Right arrow Biogeochemical Processes
Right arrow Nutrient Cycling

TECHNICAL REPORTS

Plant and Environment Interactions

Riparian Plant Material Inputs to the Murray River, Australia

Composition, Reactivity, and Role of Nutrients

Graeme Esslemonta,b, William Mahera,b,*, Phillip Forda,c and Ian Lawrencea

a Cooperative Research Centre for Freshwater Ecology, Univ. of Canberra, ACT 2601 Australia
b Ecochemistry Lab., Institute of Applied Ecology, Univ. of Canberra, ACT 2601 Australia
c CSIRO, Land and Water. GPO Box 1666, Canberra, ACT, 2601

* Corresponding author (Bill.Maher{at}canberra.edu.au)

Received for publication August 15, 2006.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
By changing riparian plants from Eucalypts to pasture and exotic deciduous trees, modern development has altered the type of carbon assimilated by Australian rivers. To investigate influences of plant litter substrates on biochemical oxygen demand, plant materials entering the Murray River were analyzed for their composition and mineralization potential. Plant materials were distinguished compositionally by two principal components, structural carbon and macronutrients, as: (i) Eucalyptus leaves, (ii) Eucalyptus bark and Casuarina cunninghamiana seed cone, (iii) grasses, (iv) macrophytes, (v) aquatic herbs, (vi) non-eucalypt leaf (Salix, Casuarina, Acacia). Ratios of C/P (1879–14524) and C/N (65–267) were relatively high in Eucalyptus bark, while mean N/P (7–60) ratios were similar among plant materials. Terrestrial weathering increased C/P and C/N ratios, while N/P ratios remained similar, due to greater loss of N and P relative to C. Aerobic decay experiments showed that nutrient supplementation accelerated decay of all organic substrates, except for grasses that decayed efficiently without supplementation. Aquatic herbs also had substantial carbon availability, macrophytes and non-eucalypt leaves had intermediate carbon availability, while eucalypt leaf and bark had intermediate to low carbon availabilities. Because biochemical oxygen demand varies with organic substrates sampled from the Murray River, and also with soluble nutrient availability, it is plausible that that modern changes to riverine plant communities and land use have influenced the biogeochemistry of this river toward faster, and more complete, processing of allochthonous carbon.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
FOR the past two decades, ecologists have investigated how carbon input and fluxes apply to flood-plain rivers (Robertson et al., 1999). Conceptual models such as the river continuum concept (Vannote et al., 1980), the serial discontinuity concept (Ward and Stanford, 1983) and the flood pulse concept (Junk et al., 1989) describe sources, quantities, and fluvial transport of organic matter downstream and across the floodplain. In Australian rivers, with vastly different modes of carbon acquisition in headwater streams, floodplain sections, and ephemeral dryland rivers, Robertson et al. (1999, p. 814) noted: "the best general model for large rivers may be one that incorporates the longitudinal perspective of the river continuum concept in upland reaches, and the lateral perspective of the flood pulse concept in unconstrained lowland reaches."

Building on these conceptual models, there is a need to consider how carbon functions as a substrate in river ecosystems, because the microbial mineralization of carbon varies greatly with the source and age of the organic substrate (Waksman and Tenney, 1927b; Westrich and Berner, 1984). Litter bag experiments, using weight loss as a surrogate for microbial mineralization, have shown that organic substrates rich in water-soluble carbohydrates and cellulose mineralize faster than lignin-rich substrates (Hodkinson, 1975; Hunt, 1977; Day, 1982; Melillo et al., 1982). Thus carbon can be viewed as a two-component mixture (reactive and refractory), with rate constants for mineralization differing by at least an order of magnitude (Hunt, 1977; Carpenter, 1982; Weider and Lang, 1982; Boudreau, 1996). Furthermore, nitrogen- and phosphorus-rich substrates degrade faster (Howard and Fisher, 1976; Hunt, 1977; Carpenter and Adams, 1979; Day, 1982), and high dissolved nutrient availabilities in the surrounding water also promote degradation (Anderson, 1978; Carpenter and Adams, 1979; Meyer and Johnson, 1983; Fairchild et al., 1984). So, in the context of the source and transport of organic carbon in Australian rivers, it is also important to understand the assimilation of organic substrates by Australian rivers.

The compositional basis of plant litter entering the Murray River is poorly understood. While investigations of riparian plant litter have been done internationally and in Australia sub-regionally, catchment-scale application of these results to the Murray is hampered by different analytical protocols, limited biogeographical ranges of plants, and inappropriate carbon substrates for Australian rivers. Furthermore, since habitat deterioration is known to result from anoxia manifested by invasion of exotic willows (Lester et al., 1994) and pasture grasses (Pusey and Arthington, 2003), there is a need to quantify decay rates with respect to oxygen drawdown, rather than with respect to weight loss. Weight loss measurements, though common in comparative studies of plant litter decay (e.g., Webster and Benfield, 1986), do not answer questions about oxygen drawdown. These limitations have hindered the development of system-scale process models of carbon and nutrient biochemistry, by which river managers can evaluate the outcome of restorative actions such as riparian revegetation or willow eradication.

Modern development of the Murray River has altered its riverine plant community, potentially altering the carbon biogeochemistry and contributing to eutrophic conditions. Our research aim was to review and test this concept, and develop data appropriate for process models, using the following hypotheses:

H0: carbon substrates vary with plant litter type, but compositional variations are inconsequential for aerobic processing of plant litter. This would be shown by different organic substrates, yet consistent aerobic mineralization rates.
H1: carbon substrates vary with plant litter type, with potential consequences for aerobic processing of plant litter. This would be shown by different organic substrates and aerobic mineralization rates.

Our primary objectives were: (i) Characterize the composition and biological oxygen demand potential of plant litter assimilated by the Murray. (ii) Refine this concept with a literature review, by comparing our observations with international and Australian investigations. Because {delta}13C and {delta}15N are accepted tracers of organic substrates in aquatic systems, a secondary objective was to characterize the isotopic signatures of plant litter to allow recognition of morphologically unrecognizable particulate organic matter.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Rationale for Representing Plant Material
We selected common plants from the catchment (Table 1) to represent the range of organic substrates entering the Murray River by consulting with experienced botanists and ecologists (see acknowledgments). Fresh materials were green leaf or living bark. Abscission materials were yellow leaves and bark peeling off deciduous trees. Weathered materials were leaves and bark from the littoral soil horizon, or desiccated stands of senescent macrophytes and grasses. The rationale for representing from each plant unweathered (fresh or abscission) and weathered material, when available, was to identify weathering influences on organic substrates at the time of sampling. Sampling occurred during mid-summer (January and February 1999) when water-stressed plants (Eucalyptus, Salix, Lolium) shed senescent leaves and bark into drainage areas (riparian corridors and sometimes dry river beds). Sampling locations were spaced evenly throughout the catchment (Fig. 1 ) and included monitoring sites to facilitate compatibility with other research (Gawne et al., 2000) and continuing work by river managers.


View this table:
[in this window]
[in a new window]

  		Table 1. Types of plant material collected along the Murray and Murrumbidgee rivers.{dagger}

 

Figure 1
View larger version (36K):
[in this window]
[in a new window]

  		Fig. 1. Sampling sites along the Murray and Murrumbidgee rivers.

 
Sampling of Plant Material
Voucher specimens from each plant, per site, were pressed for review by experienced botanists. The sampling protocol used species and plant-portion controls to represent organic substrates commonly assimilated by the river. Because the project represented catchment scale variation, not site-, tree-, or leaf-scale variation, we systematically collected composite samples that were replicated at the site scale.

We collected a composite sample of each plant component, from each site if available. For riparian trees, this was 20 leaves (or bark portions) from 20 trees at each site (400 leaves). Weathered materials were collected below trees before a limb was shaken to shed fresh and abscission leaves onto a tarpaulin. Emergent stems of macrophytes and Myriophyllum were cropped from several plants at the waterline, with the stems being pooled to provide a composite sample of consistent volume per site. Composites of floating leaves were used for Potamogeton, Brasenia, and Ottelia, and submerged leaves for Vallisnera. A soft scourer and scalpel were used in the field to remove invertebrate larvae and biofilms from aquatic leaf surfaces. All samples were promptly chilled on ice, and later frozen. Cleaning of emergent macrophytes and riparian tree material was done later in the laboratory by the same process.

Chemical Analysis of Plant Material
Analysis of plant materials conformed to minimum dataset requirements recommended by Palm and Rowland (1997). Samples were thawed to room temperature (20 ± 1°C), then dried in convection ovens (1 to 5 d) at the mean daily maximum air temperature over summer for most of the Murray River (33 ± 1°C; Gentilli, 1971). Samples were split, with one portion being blended (Moulinex D72/Q50 turbo blender), milled (Siebtechnik T250), then sieved (200 µm nylon mesh) to achieve homogeneous subsamples. Total carbon and nitrogen concentrations were analyzed by a Europa ANCA-SL 2020 Mass Spectrometer, or LECO total carbon and nitrogen analyzer. The reference material (White Wings Flour) was within 3% recovery for C and 6% recovery for N (reference values were 43.24% C and 1.58% N). Stable isotopes measured by the mass spectrometer were between ± 0.08{per thousand} {delta}13C and ± 0.47{per thousand} {delta}15N of the White Wings Flour reference (reference values for {delta}13C = –24.4{per thousand} and {delta}15N = +2.4{per thousand}). Total phosphorus was measured by inductively coupled plasma-mass spectroscopy (ICP–MS) (PerkinElmer SCIEX ELAN 6000) after microwave-digestion (0.1-g subsamples, nitric acid, 110°C), with calibration by matrix matched external standards (EM Science multi element standard; PerkinElmer Pure Plus multi element standard) as described by Esslemont et al. (2000). Two certified reference materials were analyzed per batch (National Bureau of Standards 1572 citrus leaves, National Institute of Environmental Studies No 7 tea leaves), with average recoveries ±10% of reference values (0.13 µg g–1 = 4.20 µM kg–1 as P for NBS 1572; 0.37 µg g–1 = 11.9 µM kg–1 as P for NIES 7). To facilitate comparison between elements with different atomic weights, elements were expressed in molar quantities (M kg–1 or M L–1), by dividing concentrations (mg kg–1 as a solid or mg L–1 as a liquid with 1 L representing 1000 g) by atomic weight.

Proximate carbon fractions were measured in 1-g subsamples using the standard forest products analyses of the Technical Association of the Pulp and Paper Industry (Ryan et al., 1990). This involved gravimetric measurements after sequentially removing non-polar compounds (fats, oils, waxes, and some pigments) (TAPPI, 1988a), water-extractable polar compounds (simple sugars, starches, and polyphenols) (TAPPI, 1988b), and sulfuric acid-extractable compounds (cellulose and hemicellulose) from klason lignin (Effland, 1977). Filtered extracts (Whatman GF/C filters) were analyzed by ultraviolet-visible spectroscopy (Shimadzu UV-2401 PC), for simple sugars and polyphenols in the polar extract (Dubois et al., 1956; Allen et al., 1974, respectively), and hydrolyzed carbohydrates (cellulose and hemicellulose) in the acid extract (Dubois et al., 1956). Consistent differences between gravimetrically measured totals and ultraviolet-visible measured components represented unresolved proximate fractions: a hot water polar extract (unresolved A) and an acid-soluble extract (unresolved B). Proximate fractions were expressed in weight %.

Analysis of Aerobic Mineralization Rates
Manometers (Oxytop WTW) were used to measure aerobic heterotrophic respiration (mineralization) under paired nutrient-enriched and nutrient-limited conditions. Procedurally, we could use only a few milligrams of milled plant subsample to maintain the aerobic assumptions of the experiment. Pilot work showed that leaf-scale heterogeneity caused variation in measurements at this subsample scale. Respiration measurements for Typha and Lolium showed that cut leaf portions had two to three times the variation of milled leaf. Furthermore, rate measurements for cut leaf were two times different from milled leaf, and Typha decay rates were influenced by whether the leaf tip or base was subsampled. Therefore, to establish homogenous substrates for analysis, milled subsamples representing a composite of several leaves of a plant species from a site were used, in preference to cut leaf portions. Oxygen depletion for each composite sample was measured in duplicate, with between one and three plant species being represented per principal component group.

Incubation was at 20 ± 1°C. One manometer represented nutrient-enriched conditions (0.065 mM L–1 KH2PO4, 0.125 mM L–1 K2HPO4, 0.125 mM L–1 Na2HPO4.7H20, 0.032 mM L–1 NH4Cl, 0.091 mM L–1 MgSO4.7H2O, 0.248 mM L–1 CaCl2, 0.001 mM L–1 FeCl3.6H2O) using the standard biological oxygen demand (BOD) procedure (APHA, 1992; Esslemont et al., 2001) with plant material as the carbon source. A dextrose-glutamic acid standard showed normal BOD5 test performance. These dissolved phosphorus and nitrogen concentrations were respectively 73 and 10 times the average in the lower Murray River, exceeding or equivalent to the maximum reported concentrations (58 µM L–1 and 30 µM L–1 respectively) (Crabb, 1997). The other paired manometer measured mineralization under nutrient limitation in deionized water (Millipore Milli Q). Because different plant materials had different total carbon concentrations, we normalized the progressive BOD relative to total carbon in the raw plant litter. Thus mineralization of bioavailable carbon to carbon dioxide was represented as a molar proportion of the total carbon in the manometer:

Formula 1[1]
where NBODt = normalized BOD at time t, (O2)t = moles of oxygen consumed at time t, and (C)0 = moles of carbon at the start of the experiment. Pilot work showed that 100 to 130 mg plant material in 250 mL solution optimized the manometer's capacity to measure all plant materials over 80 d while maintaining aerobic conditions (>81.5% O2 saturation throughout the experiment). These loads gave molar ratios of total carbon in plant material to total dissolved nitrogen and phosphorus in the nutrient-enriched condition of about 30:3:1 (not including the nitrogen and phosphorus already in the plant biomass). Eucalyptus total carbon loads were increased by 30% mass/volume because these substrates were relatively unreactive. Begon et al. (1990) suggests a C/N/P ratio of about 100:10:1 for sustaining microbial populations. Hence, the stoichiometry of bioavailable nitrogen and phosphorus presented to microbes to biochemically decay the total carbon load was greater than the Redfield ratio of 106:16:1, was nitrogen-rich relative to bacterial and fungal biomass (N/P of 17 to 8) reported in natural substrates (Swift et al., 1979), and was the N/P ratio reported in animal biomass (Ågren and Bosatta, 1998). Bacteria augmentation was not used because pilot work showed adequate self-inoculation by the microbes already in the plant materials.

Data Analysis
Plant material was ordinated by molecular composition using correlation-based principal component analyses (PATN) on 151 samples (Table 1). Classification was by hierarchical agglomerative cluster analysis (PC-ORD; Sorensen distance measure; flexible UPGMA; ß = –0.1) using median values for 31 plant materials.

Mineralization data under nutrient-rich and nutrient-limited conditions were compared using repeated measures ANOVA (SPSS). For each plant material, the daily average of the duplicate oxygen depletion measurements was used as test data. Thirty-seven time values (i.e., the repeated measurements done on each sample) were used for the within subjects comparison. For the between subjects comparison 12 plant material types and two nutrient conditions were used. The Greenhouse Geissner and Huynh Feldt epsilon adjusted values were both the same as unadjusted F values, demonstrating that the sphericity assumption was valid. Post hoc comparisons of mineralization data under paired nutrient-enriched and limited conditions, were done for each plant material type by repeated measures t tests (SPSS) using the 37 time values.

To represent only the carbon fraction involved with biochemical processing rather than the total carbon load (including the non-degraded fraction), mineralization data were represented as the normalized BOD over the 80-d experiment (NBOD80 – NBODt), This specifically shows the remaining potential for carbon mineralization at each point in time, from which mineralization rates were derived. Mineralization rates (molar proportions of oxygen per carbon unit, converted to carbon dioxide per day) were calculated as a double exponential model (Hunt, 1977; Weider and Lang, 1982) using Sigmaplot. This is the most ecologically relevant and best fitting model for leaf decay (Carpenter, 1982; Brock et al., 1985b).

Formula 2[2]

Formula 3[3]
C1 and C2 respectively are the proportions (in molar units) of more reactive and less reactive carbon in the plant samples. Mineralization rate constants (d–1) for these carbon fractions are respectively k1 and k2. However, as a single exponential model Ct = Ce–kt is generally reported (Webster and Benfield, 1986), we also report first-order rates (d–1) to facilitate comparisons with published values. Correlation analyses were used to compare the total mineralized (i.e., bio-available) carbon (C1 and C2), mineralization rates (k1 and k2), proximate carbon fractions, and macronutrient concentrations (SPSS).

To review published litter trap data from the Murray River as biochemically available carbon loads, we combined our rate data with litter trap data to describe "rapidly biodegradable carbon acquisitions" entering the lower Murray. We calculated rapidly biodegradable carbon by multiplying biologically accessible carbon in plant material (C1) with their mineralization rate constant (k1). For example, Salix sp. leaf had a moderate proportion of reactive carbon (0.11 M) available in the first reaction phase, which had a mineralization rate constant of 0.07 d–1. Hence, the rapidly biodegradable carbon fraction is 0.11 x 0.07 = 0.0077 M d–1. From this we calculated rapidly biodegradable carbon acquisitions using published litter trap data (Gawne et al., 2000). Hence, a section of the Murray River acquiring 15 M m–2 of carbon annually as Salix sp. leaf, has annual loads of 15 x 0.0077 = 0.1 M m–2 d–1 as rapidly bioavailable carbon (i.e., over long periods). Monthly or seasonal acquisitions may be more appropriate for describing short-term events. This approach describes plant material types as bioavailable carbon (food) sources to aerobic heterotrophs, and thus their potential to influence BOD in rivers.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Proximate Carbon Fractions
Proximate carbon analysis (Fig. 2 ) showed potentially reactive and refractory organic substrates in riparian plant material. Ordination arranged these plant materials by their structural organic carbon and major nutrient composition (Fig. 3 ): (i) Eucalyptus leaves were rich in non-polar compounds (fats, oils, waxes and some pigments). (ii) Eucalyptus bark and C. cunninghamiana cones contained the most klason lignin, with substantial cellulose and less non-polar material. (iii) Grasses were rich in cellulose and water-soluble compounds, with the least klason lignin. (iv) Macrophytes were also rich in cellulose and water-soluble compounds, with little klason lignin. (v) Aquatic herbs had intermediate proportions of water-soluble compounds, substantial cellulose, and little klason lignin. Non-polar compounds in some aquatic herbs (Brasania sp., P. tricarinatus) were abundant. (vi) Non-eucalypt leaf (Salix, Casuarina, Acacia), had intermediate proportions of water-soluble compounds, cellulose, and klason lignin. In most cases the ordination showed negative shifts, on both factors, of weathered relative to fresh or abscission plant material (Fig. 3). This means less total carbon and increased lignin content in response to weathering. For example, weathered Typha and P.australis had less water-soluble carbon (factor 1) and more lignin (factor 2).


Figure 2
View larger version (47K):
[in this window]
[in a new window]

  		Fig. 2. Proximate organic components (% mass) in fresh material (dark shading), abscission material (light shading), and weathered material. Data are means ±2 standard errors.

 

Figure 3
View larger version (29K):
[in this window]
[in a new window]

  		Fig. 3. Principal components of riparian plant materials based on molecular proximate components, plus total carbon, nitrogen, and phosphorus. Codes for plant materials are given in Table 1. Zla represents the superposition of plant materials Ala, Bla, and Cla. Classification of plant materials (flexible unweighted pair-group method using arithmetic averages [UPGMA], ß = –0.1) shows the grouping of leaf litter compositions. Visually recognizable plant material circled in the ordination ([1] Eucalyptus leaf; [2] Eucalyptus bark and C. cunninghamiana seed cone; [3] Grasses; [4] Macrophytes; [5] Aquatic herbs; [6] Non-eucalypt leaves), are symbolized in the classification. TN, total nitrogen; TP, total phosphorus.

 
Total Carbon, Nitrogen, and Phosphorus Concentrations
Aquatic herbs were rich in phosphorus and nitrogen, while Eucalyptus bark was very low in phosphorus and nitrogen (Table 2). Weathered material had slightly lower carbon concentrations and much lower phosphorus and nitrogen concentrations than fresh and abscission material (Table 2). For example, weathered tree leaves collectively lost on average 5% total carbon, compared with 32% total nitrogen and 19% total phosphorus (Table 2). Similarly weathered grasses collectively lost on average 1% total carbon, 30% total phosphorus, and 30% total nitrogen (Table 2). These changes are reflected in the C/N/P ratios of plant materials.


View this table:
[in this window]
[in a new window]

  		Table 2. Carbon, nitrogen, and phosphorus concentrations in plant materials from the Murray and Murrumbidgee rivers. Data are means ± standard deviation.

 
The mean C/P ratios for fresh and abscission material were bark > tree leaves > macrophytes > grasses > aquatic herbs (Table 2). The mean C/N ratios for fresh and abscission material were bark ~ tree leaves ~ macrophytes ~ grasses > aquatic herbs (Table 2). For weathered material, mean N/P ratios were similar to unweathered ratios (Table 2), in contrast to C/N and C/P ratios that were higher in weathered samples.

Carbon and Nitrogen Isotopes
The common riparian plants of the River Murray represented here (Fig. 4 ), were C3 (–20 to –30{per thousand} {delta}13C) except for the grasses P. distichum and C. dactylon with C4 signatures (–10 to –16{per thousand} {delta}13C). Nitrogen isotopes in riparian plant litter represented nitrogen derived from the atmosphere or synthetic fertilizer (15N close to 0{per thousand}), or from the dissolved nitrate pool ({delta}15N close to 7{per thousand}). Several riparian plants (e.g., Typha) had high intra-specific variation on {delta}15N across the catchment (Fig. 4).


Figure 4
View larger version (32K):
[in this window]
[in a new window]

  		Fig. 4. (a–f) Isotopic ratios of {delta}13C and {delta}15N in riparian plant materials.

 
Aerobic Mineralization of Plant Material, and Influences of Nutrient Availability on Mineralization Rates
Carbon mineralization featured an initial rapid phase followed by a secondary slower phase (Fig. 5 and 6 ). Biological oxygen consumption differed significantly with plant materials and nutrient availability (Table 3). The high treatment effects ({eta}2 is a measure of variation associated with each experimental treatment, as detailed by Sheskin, 1997) showed that most variation of oxygen consumption was associated with the mineralization of plant materials (0.98) and nutrient availability (0.67). Within the manometer effects of time (0.98), material type interacting with time (0.93) and nutrients interacting with time (0.37) were all highly significant (p < 0.001). The abiotic oxidation of components other than carbon (e.g., oxidation of ammonium, which is a confound for BOD studies noted by APHA, 1992) was of only minor consequence because the treatment effect of nutrients interacting with time was low (0.37).


Figure 5
View larger version (32K):
[in this window]
[in a new window]

  		Fig. 5. Aerobic mineralization of plant materials, represented as cumulative biological oxygen demand (BOD) (M O2 consumed per M C in the original biomass) over time (d). Average measurement errors (standard error) were 0.005 to 0.015 mole/mole O2.C–1 for most plant materials, 0.025 mole/mole O2.C–1 for Typha sp., and 0.033 mole/mole O2.C–1 for Lolium sp.

 

Figure 6
View larger version (39K):
[in this window]
[in a new window]

  		Fig. 6. Aerobic mineralization of plant material: (a, d) Rate constants (k1 and k2), fractions of (b) reactive carbon (C1) and (e) refractory carbon (C2), and the reactive product (C.k) that represents the biodegradable carbon fractions during the (c) first and (f) second decay stages of decay.

 

View this table:
[in this window]
[in a new window]

  		Table 3. Aerobic mineralization of riparian plant materials: Statistical differences between nutrient-enriched and nutrient-limited mineralization.

 
Post hoc comparison showed higher rates of oxygen drawdown for all plant materials with nutrient supplementation, except for Lolium sp. (df = 36, t = 1.60, p = 0.12). Although aerobic mineralization with nutrient supplementation was faster for C. dactylon (t = 3.06, p = 0.004) and C. eragrostus (t = 2.99, p = 0.005), absolute differences were marginal, and nutrient supplements did not consistently enhance mineralization (Fig. 5).

The rapidly biodegradable carbon fraction (Fig. 6c and f) clearly showed that grasses contained the most biodegradable carbon, while eucalypt bark contained the least. Correlation analysis among mineralization rates, carbon fractions, and macronutrient concentrations revealed a significant negative correlation only between total mineralized carbon (C1 and C2) and klason lignin concentration (r2 = –0.30, p = 0.05).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Chemical Compositions of Plant Materials and Weathering Influences on Carbon and Nutrient Concentrations
The principal component analysis (Fig. 3) supports studies showing that grasses, macrophytes, and aquatic herbs are rich in cellulose and water-soluble compounds (De la Cruz and Gabriel, 1974), while tree leaves, bark, and needles are lignin-rich (Rege, 1927). Soluble carbohydrates and lipids contribute to first-phase mineralization, while cellulose and hemicelluloses contribute to second-phase mineralization (Waksman and Tenney, 1927a, 1927b; Suberkropp et al., 1976; Esteves and Barbieri, 1983; Brock et al., 1985a, 1985b). As lignin is refractory, tree leaves and bark decay more slowly than grasses, macrophytes, and aquatic herbs, as demonstrated by the reactive carbon fractions in the rate models (Table 4, Fig. 6) and the negative correlation between the total mineralized carbon fraction and klason lignin concentration (r2 = –0.30, p = 0.05). This supports the understanding that lignin-rich plant material resists decay (Rege, 1927; Daubemire and Prusso, 1963; Day, 1982) to form the basis of soil humus (Jensen, 1929).


View this table:
[in this window]
[in a new window]

  		Table 4. Aerobic mineralization of plant material: Single exponential rate model, mean rate constants, and mean fraction of reactive carbon. Leaves unless indicated.

 
The high phosphorus and nitrogen concentrations in aquatic herbs, grasses, and macrophytes, with their relatively low C/P and C/N ratios (Fig. 7 , Table 2), should favor their mineralization potential over bark and tree leaves (Rege, 1927; Melin, 1930; Weider and Lang, 1982; Day, 1982; Meyer and Johnson, 1983). Our results support this prediction, particularly for decay-resistant Eucalyptus bark and leaf (Fig. 5). Furthermore, weathered material had higher C/P and C/N ratios, resulting from slightly lower carbon concentrations in relation to much lower phosphorus and nitrogen concentrations, when compared with fresh or abscission material (Table 2). However, as only grasses did not require nutrient supplementation for maximum mineralization (Fig. 5) despite the mean N/P ratio among plant materials being relatively linear at about 27:1 (Fig. 8 ), some plant materials rich in total nitrogen and phosphorus (Table 2) may contain these elements in less bioavailable forms.


Figure 7
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 7. (a–d) Ratios of C/N relative to {delta}13C in plant materials.

 

Figure 8
View larger version (18K):
[in this window]
[in a new window]

  		Fig. 8. (a, b) Total N and P in plant material. Means ± maximum and minimum values.

 
Kinetics of Mineralization in Relation to Bioavailable Carbon and Nutrients
Our kinetic data (Fig. 5) showed the classic reactive and refractory phases (Anderson, 1978; Hill, 1979; Carpenter, 1982; Westrich and Berner, 1984; Brock et al., 1985b). First phase rate constants were 1 to 0.03 d–1, with second phase constants of 0.006 to 0.0002 d–1 (Fig. 6). Our results did not exactly match the order of plant litter decay identified by a review of the international literature by Webster and Benfield (1986): aquatic herbs ≥ woody plants ≥ emergent macrophytes, grasses, and herbs. This was because we measured rate constants for BOD relative to biodegradable carbon, instead of mass loss rates for leaf packs. Furthermore, unlike some other studies, we did not find significant correlations between mineralization rates and lignin/nitrogen ratio (Melillo et al., 1982) or nitrogen concentrations (Day, 1982). Discrepancies may result from factors that include decomposer efficiencies, or strategies used by plants to retain biochemically useful elements. Aerts (1997) observed that chemical predictors of litter decomposition, such as the lignin/nitrogen ratio, were absent in temperate regions (the Murray River is temperate), by contrast with tropical and Mediterranean regions. To explain this, Aerts suggested that the absence of white-rot fungi in temperate regions limited the decay of lignin, so the lack of an effective decomposer of lignin in the plant-litter microfauna might have influenced our rate measurements. Alternatively, Cornelissen (1996) did manage to correlate decomposition rates with the lignin/nitrogen ratio of green leaves, in a temperate region with fertile soils and reliable rainfall. This was done using plant species that adopted an ecological strategy of rapid growth rather than resource retention. By contrast, plants native to the Murray are adapted to water limitation and infertile soils, so would follow the resource retention strategy. Eucalypts have the classic traits of resource retainers; they reduce evapotranspiration using waxy cuticles, avoid predation using chemical defenses, and retain nutrients during leaf abscission. So, a combination of region-specific biological factors might have influenced the decay rates of lignin-rich material from the Murray catchment.

In leaf litter from the Murray region, with the few exceptions, we found that nutrient supplementation influenced mineralization rates (Fig. 5 and Table 4), suggesting clear differences in carbon, nitrogen, and phosphorus concentrations and bioavailabilities from within different plant materials. Other studies have also shown that high nutrient availabilities in the surrounding water column promotes mineralization of some plant litter types (Anderson, 1978; Fairchild et al., 1984; Brock et al., 1985a, 1985b) through stimulated microbial activity (Melin, 1930), while other plant materials have sufficiently bioavailable nutrients for complete mineralization (Day, 1982; Brock et al., 1985a). We observed an order of plant litter decay: grasses ≥ macrophyte and aquatic herbs ≥ Eucalyptus leaf and bark (Fig. 6c). Grasses were rapidly and nearly completely mineralized without nutrient supplementation, while Eucalyptus materials mineralized slowly under aerobic conditions with utilization of added nutrients (Fig. 5). In our study allopathic compounds probably inhibited the aerobic mineralization of Eucalyptus leaf, because the small percentage of bioavailable carbon (Fig. 6b) did not correspond with the relatively abundant soluble carbohydrate in the proximate analysis (Fig. 2). This is consistent with the resource-retention strategy. Conway (2005) has reported inhibition of aerobic decomposition by soluble compounds derived from Eucalyptus.

Using Carbon and Nitrogen Isotopes to Identify Sources of Plant Litter along the Murray River
Stable isotopes have been used with C/N ratios to source organic matter in river sediments (Bird and Pousai, 1997) and in trophic structures (Bunn and Boon, 1993; Thorp et al., 1998). We observed that because isotopic ratios in plant materials varied substantially across the catchment (Fig. 4), distinguishing specific plants using these tracers is not possible at the catchment scale. Interpretations are constrained by natural variations, since isotopic carbon is heterogeneous across local to global scales (Bird and Pousai, 1997). Nitrogen isotopes are also subject to enrichment or depletion by soil and biotic influences (Högberg, 1997). However, our results suggest that careful application of isotopes and C/N ratios might crudely identify plant material acquisitions in some altered vs. unaltered rivers in East Australia. For example, an urban subcatchment dominated by Salix sp. and C. dactylon would have lower C/N and {delta}13C values relative to an unaltered river section dominated by Eucalyptus and Phragmites (Fig. 7). This approach might be a useful way of identifying contributions from urbanized subcatchments that link with undeveloped subcatchments along the Murray river.

Regional Scale Acquisitions of Allochthonous Carbon and Nutrients along the Murray River
The Murray River crosses the vegetation zones of the Eastern Highlands and Murray Lowlands (Fig. 1), which influence the quantity and composition of carbon entering this river (Robertson et al., 1999). Furthermore, carbon acquisitions change downstream because allochthonous contributions are potentially higher in forested, low-order streams of the Eastern Highlands where narrow channels are shaded. Here we review our results in the context of known carbon acquisitions entering this river.

Eucalypts provide 36 to 66% mass of plant material (12 to 23% mass as bark, and 12 to 54% mass as leaf) entering forested mountain streams, with peak accession in summer (Attiwill et al., 1978; Blackburn and Petr, 1979). In lowland sections E. camaldulensis dominates the riparian zone (bank covers are between 50 and 100% bank area)(Gawne et al., 2000), with annual contributions containing 12 to 26% mass as bark, and 21 to 60% mass as leaf (Briggs and Maher, 1983; Gawne et al., 2000). Our data show that E. viminales, common in the Eastern Highlands, contains most of its organic matter as waxes and lignin (Fig. 2). Organic constituents in E. camaldulensis resemble E. viminales. Hence the dominant riparian tree, Eucalyptus, contributes substantial refractory material that requires supplementary nutrients to be fully mineralized.

In places Salix sp. is an established weed (MDBC, 1990), covering up to 17% of the bank (Gawne et al., 2000). Salix sp. has at least equivalent concentrations of mineralizable carbon (Table 4 and Fig. 6), and higher phosphorus and nitrogen concentrations compared with native trees (Table 2). However, acquisition of plant materials by rivers needs to be considered. Table 5 summarizes C/N/P concentrations acquired via leaf and bark in litter trap data (Gawne et al., 2000). High carbon and nutrient loads, particularly of phosphorus, occur via acquisition of Salix sp. leaves relative to E. camaldulensis leaves. Furthermore, macrophytes shed localized loads that are very rich in bioavailable organic carbon, total nitrogen, and total phosphorus. Note that we have not accounted for translocation of carbohydrates (Klimes et al., 1999) and nutrients (Denny, 1987) into rhizomes during macrophyte senescence, and so we may have overestimated bioavailable carbon and nutrient contributions in macrophyte material.


View this table:
[in this window]
[in a new window]

  		Table 5. Riparian materials entering the Murray and Murrumbidgee rivers: Estimated carbon, nitrogen, and phosphorus acquisitions per year at three sites shown in Fig. 1.

 
These data show that bioavailable carbon and nutrients in plant material entering the Murray River vary greatly, and that acquisitions will vary with the timing and manner by which plant materials enter the river. The introduction of non-native trees (e.g., Salix sp., Lolium sp.) in place of Eucalypts, may substantially increase nitrogen, phosphorus, and biologically reactive carbon entering Australian rivers from riparian sources.

The Relevance of Riparian Plants to Managing Water Quality in the Murray River
The principal components of plant litter composition (Fig. 3) are consistent with principal components of specialized ecological traits that occur at the global scale (Díaz et al., 2004), which describe traits by which plants capture nutrients and water resources in productive landscapes (negative loading on axis 1) vs. less productive landscapes (positive loading on axis 1). Acquisitive-type plants with negative loading have traits that allow for the rapid acquisition of resources, which may be advantageous in landscape patches reliably supplied with resources. Retentive-type plants with positive loadings conserve resources within well-protected tissues, which may be of benefit when resources are poor and irregularly supplied. These traits allow specialist plants to inhabit parts of the riparian landscape with varying harshness (Díaz et al., 2004). Of relevance, the replacement of Eucalypts with willows and grasses, together with stabilizing water supply using weirs and irrigation, means that retentive-type plants that efficiently acquire and protect their photosynthetic store from herbivory have been replaced by acquisitive-type plants.

It has been suggested that development of Australian catchments has speeded up the cycling of elements (Harris, 2002). Clearly, plant material has the greatest potential for mineralization immediately after entering a fluvial system, whether acquisition is directly into the main river or a disconnected channel (e.g., billabong, flood chute). Aerobic mineralization is supported by flowing water that distributes oxygen within the water column, and the period following stormwater runoff events is critical for the initial decay of allochthonous plant material. Partly mineralized plant material can concentrate in low energy fluvial environments, such as lakes and dams at the macro-scale, stormwater impoundments and macrophyte beds at the meso-scale, and on the lee side of ripples and pebbles on river bottoms at the micro-scale, by hydrodynamic sorting due to its low specific gravity. In these environments, slow, late-stage mineralization can influence the dissolved oxygen concentrations of bottom waters and in aquatic sediments, with consequences for phosphorus and nitrogen processes. Fish kills are sometimes observed with anoxic "black-water events" associated with high flows, for example at site 9 (observed by the landholder), and also in Broken Creek, Victoria (http://www.envict.org.au/inform.php?menu=7&submenu=221&item=472 [verified 29 Mar. 2007]).

One of the consequences of river regulation is flow alteration, and summer releases supply water downstream during dry periods. During dry weather, pools of water in shaded anabranches and billabongs become isolated, stagnant, and in many cases anoxic as a result of carbon accumulations (McGinness et al., 2002). By reconnecting drainage lines, releases displace the anoxic "black-water" pools that then flow into the main channel. This arguably occurs naturally in ephemeral streams that flow through forests of deciduous Eucalyptus, and may be an important part of river ecology by stabilizing selection as suggested by the intermediate disturbance hypothesis (Connell, 1978). However, regulatory changes to flow (e.g., from winter and spring, to summer when litter fall in Eucalyptus forests is at it peak) and landscape connectivity with respect to nutrients and carbon entering the river system are landscape pressures that warrant integrated catchment management by both governments and landholders.

We have shown that carbon bioavailability and mineralization rates differ with plant materials and nutrient availability, and that these factors influence dissolved oxygen concentrations. Therefore, our study suggests that changing riparian plants from Eucalyptus to grasses will increase loads of bioavailable organic carbon if these plant materials enter aquatic systems. This has been reported in Queensland streams infested with para grass (Urochloa mutica), with detrimental environmental consequences (Pusey and Arthington, 2003). Furthermore, increased phosphorus and nitrogen loads entering aquatic systems due to increased fertilizer application will promote the mineralization of plant materials that lack bioavailable nutrients. This is significant to catchment management, because land use influences the supply and mineralization of plant materials that underpin river ecosystems. Many natural riparian corridors contain extensive stands of mature bank area interspersed with sparse stands of early successional plants in naturally disturbed areas. However, developed landscapes usually reverse this pattern by featuring extensive grasslands and stands of exotic trees, with mature bank area growing in the less spatially dominant patches. Implicit with these changes are alterations to the Murray River's capacity to assimilate carbon, resulting from mismatches between the types and timing of acquired plant materials. This will plausibly contribute to altered biological oxygen demand in waterways.


   ACKNOWLEDGMENTS
 
The authors are grateful to the CRC for Freshwater Ecology for support. We thank Zia Hussain and Nick Hill (University of Canberra) for their assistance in the field and laboratory, and Richard Phillips (CSIRO) for the analysis of isotopes. The following persons suggested plant types suited to this investigation: Jane Roberts (CSIRO, Canberra), Lisa Evans, David Williams (University of Canberra), Judy Frankenberg (Murray Darling Freshwater Research Centre), and Ben Gawne (Lower Murray Laboratory). David Williams and Judy Frankenberg verified our field identification of the plant types. Katerina Mikac (University of Canberra) helped with the statistics.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Esslemont, G.
Right arrow Articles by Lawrence, I.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Esslemont, G.
Right arrow Articles by Lawrence, I.
Agricola
Right arrow Articles by Esslemont, G.
Right arrow Articles by Lawrence, I.
Related Collections
Right arrow Biogeochemical Processes
Right arrow Nutrient Cycling

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome