Abstract. The strength and generality of the relationship between decomposition rates and detritus carbon, nitro- gen, and phosphorus concentrations was ...
Oecologia (1993) 94:457M71
Oecologia 9 Springer-Verlag 1993
Review article
Patterns in decomposition rates among photosynthetic organisms: the importance of detritus C :N :P content S. Enriquez ~, C.M. Duarte ~, K. Sand-Jensen 2 t Centro de Estudios Avanzados de Blanes, (CSIC), Cami de Santa BArbara, 17300 Blanes, Girona, Spain 2 Freshwater Biological Laboratory, University of Copenhagen, 51 Helsingorsgade, 3400 Hillerod, Denmark Received: 30 January 1993 / Accepted: 4 April 1993
Abstract. The strength and generality of the relationship between decomposition rates and detritus carbon, nitrogen, and phosphorus concentrations was assessed by comparing published reports of decomposition rates of detritus of photosynthetic organisms, from unicellular algae to trees. The results obtained demonstrated the existence of a general positive, linear relationship between plant decomposition rates and nitrogen and phosphorus concentrations. Differences in the carbon, nitrogen, and phosphorus concentrations of plant detritus accounted for 89% of the variance in plant decomposition rates of detritus originating from photosynthetic organisms ranging from unicellular microalgae to trees. The results also demonstrate that moist plant material decomposes substantially faster than dry material with similar nutrient concentrations. Consideration of lignin, instead of carbon, concentrations did not improve the relationships obtained. These results reflect the coupling of phosphorus and nitrogen in the basic biochemical processes of both plants and their microbial decomposers, and stress the importance of this coupling for carbon and nutrient flow in ecosystems. Key words: Decomposition - Plant kingdom - Nutrients
Carbon fixed by photosynthetic organisms is made available to other ecosystem components via herbivores or detritivores. The detrital path is a major determinant of the flow of carbon fixed by plants in ecosystems were herbivores consume a modest fraction of primary production, as is often the case (Swift et al. 1979). Decomposition of plant detritus is largely conducted by bacteria and fungi (e.g. Persson et al. 1980), and the rate of this process depends, therefore, on all factors influencing their activity. These may be separated, following Swift et al. (1979), into abiotic factors, the physicoThis work was funded through a grant of CICYT (MAR91~503) to C.M.D. Correspondence to: S. Enriquez
chemical conditions under which the decomposition occurs, and substrate quality (e.g. biochemical composition of plant litter), which constrains its suitability for microbial growth. Photosynthetic organisms can directly influence decomposition rates through their biochemical composition. For instance, plants may accumulate defence chemicals in their tissues which, besides decreasing their palatability to grazers (e.g. Coley et al. 1985), also reduce their quality as a substrate for decomposer microorganisms (Swift et al. 1979). Similarly, nutrient reabsorption before abscission of plant tissues may, in addition to improving the internal nutrient economy of the plant (Chapin 1980), affect their suitability as substrate for microbial decomposers. Decomposer organisms tend to have very high nitrogen and phosphorus contents (Findlay 1934; Thayer 1974; Swift et al. 1979; Goldman et al. 1987; Vadstein and Olsen 1989) indicative of high requirements for these nutrients. For instance balanced bacterial growth requires substrates with carbon, nitrogen, and phosphorus in an (atomic) ratio of 106:12:1 (Goldman et al. 1987), although bacteria have some capacity to vary these requirements (e.g. Tezuka 1990). These high nutrient contents are only encountered in fast-growing phytoplankton cells (Goldman et al. 1979; Duarte 1992), and microbial decomposers are often supplied with plant detritus depleted in nitrogen and phosphorus relative to their requirements. Recent research has demonstrated that bacterial growth efficiency (i.e. the fraction of the carbon used allocated to growth) decreases about 100-fold with increasing C/N and C/P ratios in their substrate (Goldman et al. 1987). Thus, detritus with high nitrogen and phosphorus content should decompose fast because of the associated fast growth of the microbial populations, whereas excess carbon in the plant litter should lead to nutrient-controlled carbon remineralization (cf. Goldman et al. 1987; Vadstein and Olsen 1989). These arguments provide an explanation for the increase in decomposition rate with increasing nutrient concentration, or decreasing carbon/nutrient ratios, demonstrated six decades ago (Tenny and Waksman 1929), and confirmed since for different aquatic (e.g.
458
Valiela et al. 1984; Twilley et al. 1986; Harrison 1989; Reddy and DeBusk 1991) and terrestrial (e.g. Gosz et al. 1973; Swift et al. 1979; Berg et al. 1982; Taylor et al. 1989; Upadhyay et al. 1989) systems. In addition to reflecting direct nutrient effects, these relationships also appear to have an indirect component, derived from a tendency towards reduced carbon quality and increasing amounts of secondary metabolites in plant litter as nutrient availability decreases (Coley et al. 1985, Chapin et al. 1987). Hence, some ratios incorporating a descriptor of carbon quality (e.g. lignin/N ratios) have also been shown to be related to decay rates of plant litter (e.g. Melillo et al. 1982; Aber et al. 1990). However, lignin/N ratios appear to outperform C/nutrient ratios as a predictor of decay rates only when comparing plant litters of similar lignin contents (Taylor et al. 1989). Whether the widespread finding of strong relationships between litter nutrient content and decomposition rates reflects the existence of a general relationship, applicable to detritus originating from different photosynthetic organism, is not known as yet. The existence of such a general relationship is expected because all microbial decomposers have high nitrogen and phosphorus, in addition to carbon, needs in both aquatic (Goldman et al. 1987; Vadstein and Olsen 1989) and terrestrial (Findlay 1934; Thayer 1974; Swift et al. 1979) environments. Conversely, these relationships might differ between different sorts of plant detritus if they were indirect, resulting from covariation between carbon quality (e.g. contents of lignin, polyphenols, etc.) and nutrient content within plant types (e.g. Melillo et al. 1982; Abet et al. 1990; Upadhyay et al. 1989). Here we examine the strength and generality of the relationship between decomposition rates and plant nutrient concentrations by comparing published reports of decomposition rates and litter nutrient contents across a broad spectrum of plant detritus, from unicellular algae to trees. We first examine the variability in decomposition rates of litter from different sources, and then assess the power of differences in their nutrient concentration to statistically account for the observed variability. A subset of these data, for which lignin contents were available in addition to nitrogen and phosphorus contents, was used to compare the strength of the relationship between lignin and nutrient contents and litter decomposition rates. Because plant nutrient concentrations are often strongly intercorrelated (Garten 1976; Duarte 1992), we used path analysis (Williams et al. 1990) to statistically resolve the direct contribution of carbon, nitrogen, phosphorus, and, where available, lignin, to the observed relationship between nutrient content and detritus decomposition rate.
Methods We searched the literature for published reports of plant litter decomposition rates and chemical composition (carbon, lignin, nitrogen, and phosphorus concentrations) at initiation of decomposition. Decomposition rates (k, natural log units day-1) were described from the changes in plant dry weight (W) with time (t,
days) since the initiation of the experiments using the equation,
wt = Woe-kt which is the model most often used in the literature (Olson 1963) and simpler than the double-exponential model (e.g. O'Connell 1987). Because these decomposition rates have logarithmic units, we also described decomposition rates as the half-life of plant detritus (Ta/2, days), which, although a function of exponential decomposition rates (T1/2 = k - 1 . In 2), provides a more intuitive description of detritus turnover times. Decomposition rates were often reported in the studies, and were otherwise calculated from tabulated data or digitized graphs of weight remaining with time elapsed. We included in the data set (Appendix) all studies encountered during our search that included estimates of decomposition rates of plant litter (e.g. photosynthetic tissues, roots, rhizomes, stems), and any of the descriptor of tissue chemical composition needed to test our hypotheses (i.e. C, N, P, and lignin concentrations). Additional detail in the general description of the data set was obtained by grouping the data according to detritus origin (phytoplankton, macroalgae, seagrasses, freshwater angiosperms, amphibious plants, sedges, mangroves, grasses, shrubs, conifers, and broad-leaved deciduous and evergreen trees). The relationships between decomposition rates and nutrient concentrations were described using least-squares regression analyses of log-transformed data. Logarithmic transformation was found to be necessary to avoid heteroscedasticity in these analyses (Draper and Smith 1965). Differences in the relationship between plant litter decomposition rate and nutrient content depending on detritus origin (as defined above) were tested for using analysis of covariance (Draper and Smith 1966). The simultaneous influence of carbon (or lingin), nitrogen, and phosphorus on litter decomposition rates was tested for using multiple least squares regression analyses, instead of carbon/nutrient ratios, for the use of these ratios is conducive to statistical artifacts (cf. Chayes 1971; Atchley and Anderson 1978). The (statistical) influence of nitrogen, phosphorus, carbon (or lignin) contents on decomposition rates was partitioned into direct and indirect effects using path analysis (e.g. Williams et al. 1990). Separate path analyses were used to test the effects of C, N, and P, on the one hand, and those oflignin, N, and P, on the other, because lignin contents were only reported in a small subset of the studies, which did not include any study on phytoplankton or macroalgae.
Results and discussion
The data set comprised 256 reports of decomposition rates of plant litters originating from different photosynthetic organisms, from land an aquatic environments (Appendix). These data were gathered under a broad variety of conditions, from controlled laboratory experiments to field studies, and included decomposition of plant litter originating from photosynthetic tissues, roots, rhizomes, stems and branches, and mixtures of these (Appendix). Unfortunately, detailed descriptions of the experimental conditions (e.g. temperature, pH, oxygen tension) were only reported in a few studies and could not be included in the analysis. Decomposition rates ranged between 0.00019 day -1 for non-photosynthetic tissues of an Australian shrub (Leucospermun parile), and 0.098 day- i for the cells of a cyanobacterium (Anabaena sp.) and the leaves of a submerged freshwater angiosperm (Vallisneria spiralis), and differed significantly according to their origin (ANOVA, F=41.3, P < 0.0001; Fig. 1). Decomposition rates were faster for detritus derived from phytoplankton and
459 i
Microalgae Freshwater plants A m p h i b i o u s plants Macroalgae Seagrasses Grasses Sedges Mangroves Broad decid.tree leave: Shrubs Conifers Broad perennial tree leaves
**
I'---f]
I t
I
I
,
I
I
7--3---1
i
m
I t'---
b .
'
O
9
,
F ~
I
~
I
.
I
r-l----q f-y-]
-
9
O P P
0.0001
0.1 0
0.01
0.001
Decomposition
[
I
rates
(day
Half-life
-1)
Fig. 1. Box plots showing the distribution of detritus decomposition rates and half-lives for detritus of different sources. Boxes encom-
pass the 25 and 75% quartiles of all the data for each plant type, the central line represents the median, bars extend to the 95%
500
1000 of
1500
detritus
2000
(days)
confidence limits, asterisks-represent observations extending beyond the 95% confidence limits, and circles represent observations beyond the 99% confidence limits
Table 1. Regression equations between detritus decomposition rate (K, In units day x) and carbon (C), phosphorus (P), nitrogen (N), and
lignin concentrations (as % DW) in the plant litter Variable dependent
Intercept
Slope N
k k k k k k k
-2.45 - 1.42 1.17 - 1.38 - 1.89 -0.22 - 1.87
1.19+0.095
Slope P
Slope C
Slope lignin
0.93 • 0.066 -2.1• - 1.04• 0.20 0.80• 0.71 • 0.220 0.31 ~ 0.240
Submersed detritus: k -2.30 k - 1.22
1.33•
Terrestrial detritus: k -2.77 k - 2.20
0.48•
0.50• 0.66+0.154 0.39•
- 1.0•
1.01 •
0.46 • 0.09
-0.22•
n
rz
F
P
231 143 78 54 141 50 43
0.40 0.58 0.12 0.32 0.64 0.85 0.37
155 198 11.6 25.8 123 92 9.14
