**5. Excreta deposition and decomposition of dead plant tissues in the field**

A litterbag experiment (Fig. 10) was performed to follow the two-year decomposition of needles and twigs of *C. obtusa* on the forest floor and to compare them between Sites C and P to estimate the possible effects of excreta on the decomposition (Osono et al., 2006a). In another field survey, mass and N content of coarse woody debris (CWD: logs, snags, and stumps with diameter equal to or greater than 10 cm) were examined in the study sites to estimate the decomposition processes in cormorant-colonized forests.

Fig. 10. Litterbags to study long-term decomposition of dead plant tissues in the field. In the study of Osono et al. (2006a), needles and twigs collected at Site C were enclosed in polypropylene shade cloth (10 10 cm, mesh size of approx. 2 mm) and incubated on the forest floor at Sites C and P for two years. The litterbags were retrieved at 3- (the first year) or 6-month (the second year) intervals to analyze remaining mass and contents of organic chemical constituents and nutrients.

#### **5.1 Rate of mass loss of needles and twigs and recalcitrant compounds**

Over the two-year period, the mass loss was slower at Site P than at Site C and faster in needles than in twigs (Fig. 11). AUR mass loss in needles and twigs showed similar trends to mass loss of whole tissues and was slower at Site P than at Site C (Fig. 11). In contrast, mass loss of total carbohydrates in needles and twigs showed similar patterns between Sites C and P (data not shown; Osono et al., 2006a). These results support the hypotheses that the excreta deposition can lead to a reduction in decomposition rates and the accumulation of recalcitrant compounds in the decomposing plant tissues. The reduced AUR decomposition at Site P was primarily attributable to the reduced biomass and activity of ligninolytic basidiomycetes due to excess supply of excreta-derived N, as discussed above.

Excess Supply of Nutrients, Fungal Community, and Plant Litter

**5.3 Decomposition of coarse woody debris** 

Snag

during the period (Fig. 13).

Log Stump

CPAD

study of Katsumata (2004).

Mass (t/ha)

Decomposition: A Case Study of Avian-Derived Excreta Deposition in Conifer Plantations 185

under N-rich conditions (Berg, 1986, 1988). The 15N values (10.5 to 12.3‰) of AUR in needles and twigs at Site P, compared to those at Site C (-1.0 to 1.1‰), clearly indicated that excreta-derived N was incorporated into AUR during decomposition (Osono et al., 2006a). The formation of nitrogenous recalcitrant compounds registered as AUR resulted in the reduced net loss of mass of AUR, which in turn retarded the loss of mass of whole tissues.

Coarse woody debris (CWD) serves as a major pool and source of carbon and nutrients in forest ecosystems because of its long turnover time (Harmon et al., 1986). In Isaki peninsula, the mass of CWD ranged from 15.5 to 42.0 t/ha at Sites P, A, and D (Fig. 13). These values were 2 to 5.5 times that at Site C (7.7 t/ha, Fig. 13) and generally larger than CWD mass in most undisturbed coniferous forests (Harmon et al., 1986). The greater CWD mass in the colonized forests was due to the increased mortality of stems as snags in the colonized forest stands (Fujiwara and Takayanagi, 2001; see Section 1.3) which accounted for 68 to 87% of total CWD mass at Sites P, A, and D (Fig. 13). Most snags persisted as standing-dead for 10 years after the bird colonization at Site D, but gradually shifted from decay class I to II

CPAD

Snag

CPAD

I II

III IV

**distribution of snag**

**Decay class**

0

25

50

Proportion (%)

75

100

Site

Fig. 13. Mass and composition of coarse woody debris (CWD) and decay class distribution of snags at Sites C, P, A, and D at Isaki Peninsula (Katsumata, 2004). Sites are as in Table 1. CWD (diameter equal to or greater than 10 cm) were investigated in belt transects (4 m width, a total length of 2,030 m, 0.07 to 0.30 ha for each site) in 2003. CWD were recorded for each of three categories (log, snag, stump) and each of five decay classes [decay class I (recently dead and minimally decomposed) to V (strongly decomposed)] according to visual criteria for coniferous CWD (Sollins, 1982). No snag was classified into decay class V in the

The nitrogen content of CWD of *C. obtusa* was generally low regardless of the category (log, snag, or stump) and the decay class (I to V), mostly ranging from 0.8 to 1.5 mg/g (Fig. 14). The exceptions were logs in decay class IV at Sites P and A that had higher N content (mean values of 6.6 and 5.8 mg/g, respectively) (Fig. 14). However, the differences in N contents in CWD among the categories or the decay classes were not statistically significant (generalized linear model, P>0.05) because of a large variation in N content between CWD

0

25

50

Proportion (%)

75

100

**Biomass of CWD Composition of CWD**

Log Stump

Fig. 11. Changes in remaining mass of needles and twigs of *Chamaecyparis obtusa* (left) and of acid-unhydrolyzable residue (AUR) in needles and twigs (right) at Sites C and P examined for two years in the field (Osono et al., 2006a). needles at Site C; ● needles at Site P; twigs at Site C; ○ twigs at Site P. Sites are as in Table 1. Values indicate means ± standard errors (n=3).

#### **5.2 Immobilization of excreta-derived nitrogen**

The mass of N in needles at Site P increased rapidly during the first 3 months and was relatively constant thereafter, whereas that at Site C decreased during decomposition (Fig. 12). The mass of N in twigs also increased at Site P, whereas such an increase was not detected at Site C (Fig. 12). The net increase, i.e. net immobilization, of N at Site P indicates the incorporation of external N into decomposing plant tissues. 15N values of the plant tissues at Site P increased rapidly during the first 3 months to reach the value of cormorant's excreta (13.2 0.4‰, mean standard error, n=12; Kameda et al., 2006), whereas such an increase was not detected at Site C (Fig. 12). This stable isotope tracer successfully demonstrated that this exogenous N incorporated into the decomposing plant tissues was derived from excreta.

Fig. 12. Changes in remaining mass of nitrogen and nitrogen stable isotope ratio (15N, ‰) in needles and twigs of *Chamaecyparis obtusa* at Sites C and P in the field (Osono et al. 2006a). Symbols are as in Fig. 11. Values indicate means ± standard errors (n=3).

Causal relationships can be expected among the increased N immobilization, the AUR accumulation, and the reduced mass loss of whole tissues. The secondary formation of nitrogenous recalcitrant substances can be stimulated during plant litter decomposition under N-rich conditions (Berg, 1986, 1988). The 15N values (10.5 to 12.3‰) of AUR in needles and twigs at Site P, compared to those at Site C (-1.0 to 1.1‰), clearly indicated that excreta-derived N was incorporated into AUR during decomposition (Osono et al., 2006a). The formation of nitrogenous recalcitrant compounds registered as AUR resulted in the reduced net loss of mass of AUR, which in turn retarded the loss of mass of whole tissues.

#### **5.3 Decomposition of coarse woody debris**

184 International Perspectives on Global Environmental Change

**Needles & twigs AUR**

0

0 6 12 18 24

1

0.75

Incubation time (months)

Fig. 11. Changes in remaining mass of needles and twigs of *Chamaecyparis obtusa* (left) and of acid-unhydrolyzable residue (AUR) in needles and twigs (right) at Sites C and P examined for two years in the field (Osono et al., 2006a). needles at Site C; ● needles at Site P; twigs at Site C; ○ twigs at Site P. Sites are as in Table 1. Values indicate means ± standard

The mass of N in needles at Site P increased rapidly during the first 3 months and was relatively constant thereafter, whereas that at Site C decreased during decomposition (Fig. 12). The mass of N in twigs also increased at Site P, whereas such an increase was not detected at Site C (Fig. 12). The net increase, i.e. net immobilization, of N at Site P indicates the incorporation of external N into decomposing plant tissues. 15N values of the plant tissues at Site P increased rapidly during the first 3 months to reach the value of cormorant's excreta (13.2 0.4‰, mean standard error, n=12; Kameda et al., 2006), whereas such an increase was not detected at Site C (Fig. 12). This stable isotope tracer successfully demonstrated that this exogenous N incorporated into the decomposing plant tissues was

Fig. 12. Changes in remaining mass of nitrogen and nitrogen stable isotope ratio (15N, ‰) in needles and twigs of *Chamaecyparis obtusa* at Sites C and P in the field (Osono et al. 2006a).

Causal relationships can be expected among the increased N immobilization, the AUR accumulation, and the reduced mass loss of whole tissues. The secondary formation of nitrogenous recalcitrant substances can be stimulated during plant litter decomposition

Symbols are as in Fig. 11. Values indicate means ± standard errors (n=3).

1.25

0

**5.2 Immobilization of excreta-derived nitrogen** 

0 6 12 18 24

1.5 2 2.5 3

Remaining mass (g/bag)

errors (n=3).

derived from excreta.

Coarse woody debris (CWD) serves as a major pool and source of carbon and nutrients in forest ecosystems because of its long turnover time (Harmon et al., 1986). In Isaki peninsula, the mass of CWD ranged from 15.5 to 42.0 t/ha at Sites P, A, and D (Fig. 13). These values were 2 to 5.5 times that at Site C (7.7 t/ha, Fig. 13) and generally larger than CWD mass in most undisturbed coniferous forests (Harmon et al., 1986). The greater CWD mass in the colonized forests was due to the increased mortality of stems as snags in the colonized forest stands (Fujiwara and Takayanagi, 2001; see Section 1.3) which accounted for 68 to 87% of total CWD mass at Sites P, A, and D (Fig. 13). Most snags persisted as standing-dead for 10 years after the bird colonization at Site D, but gradually shifted from decay class I to II during the period (Fig. 13).

Fig. 13. Mass and composition of coarse woody debris (CWD) and decay class distribution of snags at Sites C, P, A, and D at Isaki Peninsula (Katsumata, 2004). Sites are as in Table 1. CWD (diameter equal to or greater than 10 cm) were investigated in belt transects (4 m width, a total length of 2,030 m, 0.07 to 0.30 ha for each site) in 2003. CWD were recorded for each of three categories (log, snag, stump) and each of five decay classes [decay class I (recently dead and minimally decomposed) to V (strongly decomposed)] according to visual criteria for coniferous CWD (Sollins, 1982). No snag was classified into decay class V in the study of Katsumata (2004).

The nitrogen content of CWD of *C. obtusa* was generally low regardless of the category (log, snag, or stump) and the decay class (I to V), mostly ranging from 0.8 to 1.5 mg/g (Fig. 14). The exceptions were logs in decay class IV at Sites P and A that had higher N content (mean values of 6.6 and 5.8 mg/g, respectively) (Fig. 14). However, the differences in N contents in CWD among the categories or the decay classes were not statistically significant (generalized linear model, P>0.05) because of a large variation in N content between CWD

Excess Supply of Nutrients, Fungal Community, and Plant Litter

variation of data and the low number of study sites examined.

to 16400 (8780 on average) nests /ha/year for stems.

of bird colonization were from Fujiwara (2001).

0

2

4

Litterfall amount (t/ha/year)

6

8

**6.1 Nest number-litterfall amount (NNLA) models** 

are:

Decomposition: A Case Study of Avian-Derived Excreta Deposition in Conifer Plantations 187

Litterfall amount was measured for needles, twigs, and coarse woody debris (CWD) at the study sites and linearly related to the number of cormorant nests (Fig. 15). Here, CWD is sometimes equivalently referred to as stems when mentioning them as the living compartment of forest stands. The regression equations and coefficients of determination

Twig: LFTWG = 0.0104 NE - 0.164 (n=6, R2=0.74, P<0.05) (2)

 CWD: LFCWD = 0.0122 NE - 0.096 (n=4, R2=0.71, P=0.16) (3) where LFNDL, LFTWG, and LFCWD are litterfall amount (t/ha/year) of needles, twigs, and CWD, respectively, and NE is the number of cormorant nests (/ha/year). The coefficient of determination for CWD was not statistically significant at the 5% level because of the large

These regression equations provide useful implications about the relationship between forest decline and nest number. According to previous literature, the biomass of needles, twigs, and stems in plantations of *C. obtusa* varied with location and stand age, ranging from 10 to 20 (13 on average, n=22) t/ha for needles, 10 to 30 (15 on average) t/ha for twigs, and 50 to 200 (107 on average) t/ha for stems (Kawahara, 1974). Substituting the biomass data into equations 1, 2, and 3 yields the number of cormorant nests at which all biomass within a forest stand can be transformed into litterfall. Those are: 390 to 830 (520 on average) nests/ha/year for needles, 980 to 2900 (1460 on average) nests/ha/year for twigs, and 4100

0 100 200 300

Nest (number/ha/year)

Fig. 15. Litterfall amount as related to the number of cormorant nests. needles; twigs; coarse woody debris (CWD). Regression lines are for equations 1, 2, and 3, respectively.

simultaneously during the breeding season of cormorants in 1999 and 2000 at Sites T, P, and D (Fujiwara, 2001). For CWD, the litterfall amount was calculated by measuring diameters at breast height of trees that died (as standing-dead) during the periods of bird colonization starting from the years of colony establishment (1992 to 1997) to 2000 at Sites P, A, and D, as

For needles and twigs, the litterfall amount and the nest number were measured

well as Site C (Fujiwara, 2001) and by converting these data to mass according to an allometric equation of Toda et al. (1991). The nest numbers at these sites during the periods

Needle: LFNDL = 0.0226 NE + 1.180 (n=6, R2=0.94, P<0.01) (1)

samples. Measurements of N isotope ratio in log samples of decay class IV and V indicated that 15N was 0.6‰ for a log at Site C, whereas it ranged from 4.2 to 14.8‰ (mean = 8.6‰, n=10) for logs at Sites P, A, and D (Fig. 7), suggesting that excreta-derived N can be incorporated into logs during decomposition and that some logs served as a reservoir of excreta-derived N on the forest floor.

Fig. 14. Nitrogen content (mg/g) in coarse woody debris (CWD) of *Chamaecyparis obtusa*. snag, decay class I; snag, decay class II; snag, decay class III; log, decay class I; log, decay class II; log, decay class III; log, decay class IV. Sites are as in Table 1. Values indicate means ± standard errors.

#### **6. Predicting the dynamics of dead plant tissues and excreta-derived nitrogen in colonized forest**

The previous sections demonstrated that the excess supply of N as excreta altered the patterns of decomposition of dead plant tissues due to the changes in the ecological (i.e., abundance, diversity, and species composition) and physiological (growth and ligninolytic activity) properties of saprobic fungi. The impact of birds on forest stands, however, is not limited to the supply of a large amount of excreta to the forest floor and the concomitant changes in biological properties in soils. Cormorants break needles and twigs for nesting material and frequently drop these on the forest floor. Such behavior results in a high volume of litterfall in the colonized forests (Section 1.3), which can lead to tree mortality and forest decline. The fallen needles and twigs abundantly supplied to the forest floor are expected to serve as large reservoirs of carbon and excreta-derived nutrients (Section 5). Moreover, the amount of litterfall and excreta deposition is expected to depend on the density of bird colonization, which varies in time and space (Fujiwara and Takayanagi, 2001).

In order to predict the impact of bird colonization on nutrient dynamics in soils, therefore, it is necessary to quantitatively relate the density of bird colonization to the amount of litterfall and to the amount of dead plant tissues and nutrients in the decomposing tissues. In this section, empirical linear models are constructed to describe the relationship between the number of cormorant nests (as an index of bird density) and (i) litterfall amount (denoted as Nest number-litterfall amount or NNLA model), (ii) amount of dead plant tissues remaining after a given period of decomposition (Nest number-residual mass or NNRM model), and (iii) amount of nutrients accumulated in dead plant tissues after a given period of decomposition (Nest number-residual nutrient or NNRN model).

#### **6.1 Nest number-litterfall amount (NNLA) models**

186 International Perspectives on Global Environmental Change

samples. Measurements of N isotope ratio in log samples of decay class IV and V indicated that 15N was 0.6‰ for a log at Site C, whereas it ranged from 4.2 to 14.8‰ (mean = 8.6‰, n=10) for logs at Sites P, A, and D (Fig. 7), suggesting that excreta-derived N can be incorporated into logs during decomposition and that some logs served as a reservoir of

CPAD

Site

Fig. 14. Nitrogen content (mg/g) in coarse woody debris (CWD) of *Chamaecyparis obtusa*. snag, decay class I; snag, decay class II; snag, decay class III; log, decay class I; log, decay class II; log, decay class III; log, decay class IV. Sites are as in Table 1. Values

The previous sections demonstrated that the excess supply of N as excreta altered the patterns of decomposition of dead plant tissues due to the changes in the ecological (i.e., abundance, diversity, and species composition) and physiological (growth and ligninolytic activity) properties of saprobic fungi. The impact of birds on forest stands, however, is not limited to the supply of a large amount of excreta to the forest floor and the concomitant changes in biological properties in soils. Cormorants break needles and twigs for nesting material and frequently drop these on the forest floor. Such behavior results in a high volume of litterfall in the colonized forests (Section 1.3), which can lead to tree mortality and forest decline. The fallen needles and twigs abundantly supplied to the forest floor are expected to serve as large reservoirs of carbon and excreta-derived nutrients (Section 5). Moreover, the amount of litterfall and excreta deposition is expected to depend on the density of bird colonization, which varies in time and space (Fujiwara and Takayanagi,

In order to predict the impact of bird colonization on nutrient dynamics in soils, therefore, it is necessary to quantitatively relate the density of bird colonization to the amount of litterfall and to the amount of dead plant tissues and nutrients in the decomposing tissues. In this section, empirical linear models are constructed to describe the relationship between the number of cormorant nests (as an index of bird density) and (i) litterfall amount (denoted as Nest number-litterfall amount or NNLA model), (ii) amount of dead plant tissues remaining after a given period of decomposition (Nest number-residual mass or NNRM model), and (iii) amount of nutrients accumulated in dead plant tissues after a given

period of decomposition (Nest number-residual nutrient or NNRN model).

**6. Predicting the dynamics of dead plant tissues and excreta-derived** 

0

2

4

mg/g

6

8

excreta-derived N on the forest floor.

indicate means ± standard errors.

**nitrogen in colonized forest** 

2001).

Litterfall amount was measured for needles, twigs, and coarse woody debris (CWD) at the study sites and linearly related to the number of cormorant nests (Fig. 15). Here, CWD is sometimes equivalently referred to as stems when mentioning them as the living compartment of forest stands. The regression equations and coefficients of determination are:

$$\text{Needle: LF}\_{\text{NDL}} = 0.0226 \times \text{NE} + 1.180 \text{ (n=6, R=0.94, P<0.01)} \tag{1}$$

$$\text{Twig:LF}\_{\text{TWG}} = 0.0104 \times \text{NE - 0.164 (n=6, R=0.74, P<0.05)} \tag{2}$$

$$\text{CWD: } \text{LF\_{CWD}} = 0.0122 \times \text{NE - 0.096 (n=4, R=0.71, P=0.16)} \tag{3}$$

where LFNDL, LFTWG, and LFCWD are litterfall amount (t/ha/year) of needles, twigs, and CWD, respectively, and NE is the number of cormorant nests (/ha/year). The coefficient of determination for CWD was not statistically significant at the 5% level because of the large variation of data and the low number of study sites examined.

These regression equations provide useful implications about the relationship between forest decline and nest number. According to previous literature, the biomass of needles, twigs, and stems in plantations of *C. obtusa* varied with location and stand age, ranging from 10 to 20 (13 on average, n=22) t/ha for needles, 10 to 30 (15 on average) t/ha for twigs, and 50 to 200 (107 on average) t/ha for stems (Kawahara, 1974). Substituting the biomass data into equations 1, 2, and 3 yields the number of cormorant nests at which all biomass within a forest stand can be transformed into litterfall. Those are: 390 to 830 (520 on average) nests/ha/year for needles, 980 to 2900 (1460 on average) nests/ha/year for twigs, and 4100 to 16400 (8780 on average) nests /ha/year for stems.

Fig. 15. Litterfall amount as related to the number of cormorant nests. needles; twigs; coarse woody debris (CWD). Regression lines are for equations 1, 2, and 3, respectively. For needles and twigs, the litterfall amount and the nest number were measured simultaneously during the breeding season of cormorants in 1999 and 2000 at Sites T, P, and D (Fujiwara, 2001). For CWD, the litterfall amount was calculated by measuring diameters at breast height of trees that died (as standing-dead) during the periods of bird colonization starting from the years of colony establishment (1992 to 1997) to 2000 at Sites P, A, and D, as well as Site C (Fujiwara, 2001) and by converting these data to mass according to an allometric equation of Toda et al. (1991). The nest numbers at these sites during the periods of bird colonization were from Fujiwara (2001).

Excess Supply of Nutrients, Fungal Community, and Plant Litter

respectively.

respectively.

gas into the atmosphere.

Decomposition: A Case Study of Avian-Derived Excreta Deposition in Conifer Plantations 189

Excreta-derived N was immobilized in decomposing needles and twigs (Section 5.2) and in CWD (Section 5.3). Osono et al. (2006a) estimated the potentials of these plant tissues to immobilize excreta-derived N (denoted here as the immobilization potential, which means the maximum amount of exogenous N immobilized per initial material) and the duration of this immobilization phase according to the method described in Mellilo and Aber (1984). Firstly, the linear relationships between the percent remaining mass of plant tissues and N

Needles: MRNDL/LFNDL 100 = -20 NITNDL + 120 (n=7, R2=0.76, P<0.01) (10)

Twigs: MRTWG/LFTWG 100 = -23 NITTIG + 110 (n=7, R2=0.32, P=0.18) (11)

 CWD: MRCWD/LFCWD 100 = -174 NITCWD + 105 (n=6, R2=0.13, P=0.48) (12) where NITNDL, NITTWG, and NITCWD are N content (%, w/w) of needles, twigs, and CWD,

0123

Nitrogen content (%)

Fig. 16. Linear relationships between the percent remaining mass of decomposing plant tissues (% of the original mass: MR*t*/LF100) and the N content (%, w/w) in the remaining

Substituting the intercepts and slopes of equations 10-12 and the decomposition constants of equations 4-6 into the equations of Mellilo and Aber (1984), the immobilization potential was calculated to be 6.6, 8.6, and 0.8 mg N/g initial material and the duration of the immobilization phase to be 1.6, 19.9, and 32.2 years for needles, twigs, and CWD,

Using these values, Osono et al. (2006a) estimated the N immobilization potential of litterfall to be 10.3 and 7.2 kg/ha/month for needles and twigs, respectively. These values accounted for 4.1% and 3.0% of total N input as excreta during the breeding season at Site P (i.e., 240 kg/ha/month; Kameda et al., 2006). This tentative calculation thus suggests that the increased litterfall at Site P due to breeding activity of the cormorants has a potential to immobilize only a total of 7% of total excreta-derived N deposited on the forest floor during 2 (needles) and 20 (twigs) years of decomposition. The major fate of excreta-derived N thus can be leaching into deeper soil layers (Hobara et al., 2005) and volatilization as ammonia

Finally, substituting equations 10-12 into equations 7-9 yields equations describing the relationship between the nest number and N mass (kg/ha) in the remaining needles, twigs,

and CWD, respectively, at a given decomposition time *t* (NNRN model):

**6.3 Nitrogen immobilization and Nest number-residual nutrient (NNRN) model** 

content in the remaining materials were described as regression equations (Fig. 16):

Remaining mass (%)

tissues for needles, twigs, and CWD. Symbols are as in Fig. 15.

This calculation suggests that all needles in a forest stand can fall when the annual number of nests reaches 400 to 800, or when the cumulative number of nests over some years reaches those values. Because the needle is a photosynthetic organ and because *C. obtusa* is known to lack the ability to sprout (i.e. re-grow) after mechanical loss, 400 to 800 nests per ha will be a critical level at which the forest stand cannot maintain primary production and will start to decline. This prediction is in agreement with the observation at Site D, where the cormorants colonized intensively at least for 4 years, from 1992 to 1996, and then declined (Fujiwara, 2001). The number of cormorant nests in 1992 was 269 /ha at Site D (Fujiwara, 2001), which corresponds to a litterfall rate of needles of 7.3 t/ha/year according to equation 1. This estimated litterfall rate would be high enough to result in forest decline at Site D if similar colonization density of cormorants was maintained for 4 years.

#### **6.2 Nest number-residual mass (NNRM) model**

The exponential equation of Olson (1963) is used to describe the changes in remaining mass of needles, twigs, and CWD with respect to the period of decomposition. Data of the 2-year decomposition experiment at Site P (Fig. 11; Osono et al., 2006a) were used to estimate decomposition constants (*k*, /year) for needles and twigs (Equations 4 and 5). Katsumata (2004) showed that more than 68% of CWD was present as snags in the study sites and that most of the snags persisted as standing-dead for more than 10 years but gradually shifted from decay class I to II (Fig. 13). Thus, a total of 32 snags, including those in decay class I at Sites C (no bird colonization; i.e., 0 year after colonization) and P (3 years) and in decay class I and II at Sites A (6 years) and D (10 years), were sampled to measure mass per volume. The mass per volume data of CWD were used to construct the pattern of changes in remaining mass per volume of snags and to estimate a decomposition constant for CWD by means of a chronosequence approach (Equation 6). The exponential equations are expressed as:

$$\text{Needless: } \text{MR}\_{\text{NDL}, \ell} = \text{LF}\_{\text{NDL}} \times \exp^{0.27t} \left( \text{n} = 7, \text{R} \text{2} = 0.28 \right) \tag{4}$$

$$\text{Twigs: MR}\_{\text{TWG},l} = \text{LF}\_{\text{TWG}} \times \exp^{0.034} \left(\text{n} = 7, \text{R} \newline 2 = 0.75\right) \tag{5}$$

$$\text{CWD: MR}\_{\text{CWD},t} = \text{LF}\_{\text{CWD}} \times \exp\text{\text{\textquotedbl{}}}\text{\textquotedbl{}}\left(\text{n} \equiv 6, \text{ R} \text{\textquotedbl{}} \text{\textquotedbl{}}\right) \tag{6}$$

where MRNDL,*t*, MRTWG,*t*, and MRCWD,*t* are the remaining mass (t/ha) of needles, twigs, and CWD, respectively, at time *t* and *t* is the time in years. The decomposition constants (*k*) are 0.27, 0.03, and 0.02 /year, respectively. The coefficient of determination for needles was low because of asymptotic pattern of changes in remaining mass over the study period (Fig. 11). Substituting equations 1-3 into exponential equations 4-6 yields the equations describing the relationship between the nest number and the remaining mass of needles, twigs, and CWD, respectively, at a given decomposition time *t* (NNRM model):

$$\text{Needless: MR}\_{\text{NDL},t} = \left(0.0226 \times \text{NE} + 1.180\right) \times \exp^{0.27t} \tag{7}$$

$$\text{Twings: MR}\_{\text{TWG},t} = (0.0104 \times \text{NE - } 0.164) \times \exp^{0.03t} \tag{8}$$

$$\text{CWD: MR}\_{\text{CWD,t}} = \left(0.0122 \times \text{NE - } 0.096\right) \times \exp{\text{\textdegree{0.02t}}}\tag{9}$$

This calculation suggests that all needles in a forest stand can fall when the annual number of nests reaches 400 to 800, or when the cumulative number of nests over some years reaches those values. Because the needle is a photosynthetic organ and because *C. obtusa* is known to lack the ability to sprout (i.e. re-grow) after mechanical loss, 400 to 800 nests per ha will be a critical level at which the forest stand cannot maintain primary production and will start to decline. This prediction is in agreement with the observation at Site D, where the cormorants colonized intensively at least for 4 years, from 1992 to 1996, and then declined (Fujiwara, 2001). The number of cormorant nests in 1992 was 269 /ha at Site D (Fujiwara, 2001), which corresponds to a litterfall rate of needles of 7.3 t/ha/year according to equation 1. This estimated litterfall rate would be high enough to result in forest decline at Site D if

The exponential equation of Olson (1963) is used to describe the changes in remaining mass of needles, twigs, and CWD with respect to the period of decomposition. Data of the 2-year decomposition experiment at Site P (Fig. 11; Osono et al., 2006a) were used to estimate decomposition constants (*k*, /year) for needles and twigs (Equations 4 and 5). Katsumata (2004) showed that more than 68% of CWD was present as snags in the study sites and that most of the snags persisted as standing-dead for more than 10 years but gradually shifted from decay class I to II (Fig. 13). Thus, a total of 32 snags, including those in decay class I at Sites C (no bird colonization; i.e., 0 year after colonization) and P (3 years) and in decay class I and II at Sites A (6 years) and D (10 years), were sampled to measure mass per volume. The mass per volume data of CWD were used to construct the pattern of changes in remaining mass per volume of snags and to estimate a decomposition constant for CWD by means of a chronosequence approach (Equation 6). The exponential equations are expressed

Twigs: MRTWG,*t* = LFTWG exp-0.03*t* (n=7, R2=0.75) (5)

where MRNDL,*t*, MRTWG,*t*, and MRCWD,*t* are the remaining mass (t/ha) of needles, twigs, and CWD, respectively, at time *t* and *t* is the time in years. The decomposition constants (*k*) are 0.27, 0.03, and 0.02 /year, respectively. The coefficient of determination for needles was low because of asymptotic pattern of changes in remaining mass over the study period (Fig. 11). Substituting equations 1-3 into exponential equations 4-6 yields the equations describing the relationship between the nest number and the remaining mass of needles, twigs, and CWD,

respectively, at a given decomposition time *t* (NNRM model):

Needles: MRNDL,*t* = LFNDL exp-0.27*t* (n=7, R2=0.28) (4)

CWD: MRCWD,*t* = LFCWD exp-0.02*t* (n=6, R2=0.75) (6)

Needles: MRNDL,*t* = (0.0226 NE + 1.180) exp-0.27*t* (7)

Twigs: MRTWG,*t* = (0.0104 NE - 0.164) exp-0.03*t* (8)

CWD: MRCWD,*t* = (0.0122 NE - 0.096) exp-0.02*t* (9)

similar colonization density of cormorants was maintained for 4 years.

**6.2 Nest number-residual mass (NNRM) model** 

as:

### **6.3 Nitrogen immobilization and Nest number-residual nutrient (NNRN) model**

Excreta-derived N was immobilized in decomposing needles and twigs (Section 5.2) and in CWD (Section 5.3). Osono et al. (2006a) estimated the potentials of these plant tissues to immobilize excreta-derived N (denoted here as the immobilization potential, which means the maximum amount of exogenous N immobilized per initial material) and the duration of this immobilization phase according to the method described in Mellilo and Aber (1984). Firstly, the linear relationships between the percent remaining mass of plant tissues and N content in the remaining materials were described as regression equations (Fig. 16):

$$\text{Needles: MR}\_{\text{NDM}} / \text{LF}\_{\text{NDL}} \times 100 = \text{--} 20 \times \text{NTL}\_{\text{NDL}} + 120 \text{ (n=7, R=0.76, P<0.01)} \tag{10}$$

$$\text{Twings: MR}\_{\text{TWG}} / \text{LF}\_{\text{TWG}} \times 100 = \text{-} 2\text{\\$} \times \text{NT}\_{\text{TWG}} + 110 \text{ (n=7, R=0.32, P=0.18)} \tag{11}$$

$$\text{CWD:}\,\text{MR}\_{\text{CWD}}/\text{LF}\_{\text{CWD}} \times 100 = \text{-174} \times \text{NIT}\_{\text{CWD}} + 105 \text{ (n=6, R=0.13, P=0.48)} \tag{12}$$

where NITNDL, NITTWG, and NITCWD are N content (%, w/w) of needles, twigs, and CWD, respectively.

Fig. 16. Linear relationships between the percent remaining mass of decomposing plant tissues (% of the original mass: MR*t*/LF100) and the N content (%, w/w) in the remaining tissues for needles, twigs, and CWD. Symbols are as in Fig. 15.

Substituting the intercepts and slopes of equations 10-12 and the decomposition constants of equations 4-6 into the equations of Mellilo and Aber (1984), the immobilization potential was calculated to be 6.6, 8.6, and 0.8 mg N/g initial material and the duration of the immobilization phase to be 1.6, 19.9, and 32.2 years for needles, twigs, and CWD, respectively.

Using these values, Osono et al. (2006a) estimated the N immobilization potential of litterfall to be 10.3 and 7.2 kg/ha/month for needles and twigs, respectively. These values accounted for 4.1% and 3.0% of total N input as excreta during the breeding season at Site P (i.e., 240 kg/ha/month; Kameda et al., 2006). This tentative calculation thus suggests that the increased litterfall at Site P due to breeding activity of the cormorants has a potential to immobilize only a total of 7% of total excreta-derived N deposited on the forest floor during 2 (needles) and 20 (twigs) years of decomposition. The major fate of excreta-derived N thus can be leaching into deeper soil layers (Hobara et al., 2005) and volatilization as ammonia gas into the atmosphere.

Finally, substituting equations 10-12 into equations 7-9 yields equations describing the relationship between the nest number and N mass (kg/ha) in the remaining needles, twigs, and CWD, respectively, at a given decomposition time *t* (NNRN model):

Excess Supply of Nutrients, Fungal Community, and Plant Litter

(Section 5.3), a process not incorporated in the model.

dead plant tissues as reservoirs of carbon and excreta-derived N.

**7. Conclusion** 

nutrients.

activity.

Decomposition: A Case Study of Avian-Derived Excreta Deposition in Conifer Plantations 191

mass reached its maximum value at 3 years of decomposition but decreased thereafter due to the net release from needles. The N mass in needles becomes smaller exponentially, and it almost disappears before 20 years of decomposition. This suggests that needles serve as a temporary reservoir and then as a source of N thereafter up to 20 years of decomposition. In contrast, twigs immobilize N slowly for 20 years to become the dominant reservoir of N thereafter. The model predicts that CWD (snags) plays a negligible role in N retention. Note, however, that a part of CWD can be a reservoir of N when it falls down to become logs

It should also be noted that some equations in the models have low coefficients of determination. This especially holds true for the litterfall amount of CWD (Eq. 3), the changes in remaining mass of needles (Eq. 4), and the dynamics of N in twigs and CWD (Eqs. 11 and 12). When the decomposition of needles is assumed to follow an asymptotic function, for example, needles become a more important, longer-term reservoir of dead plant tissues and N in the detritus pool. Obviously, longer-term studies of tree mortality and decomposition of needles, twigs, and CWD will be necessary to construct more accurate empirical models. Still, the present models provide useful insights into the effects of the density of cormorant colonization on the amount of litterfall and into the differential roles of

The series of studies demonstrated that excess supply of excreta-derived N changed the community structure, nutrition, and substrate utilization of saprobic fungi, which by altering the decomposition processes led to carbon sequestration, accumulation of excretaderived N, and thus a slow turnover of carbon and N in forest soils affected by the cormorant (Fig. 18). Most of the previous studies examined the effects of excess supply of nutrients on fungal communities, microbial activities, decomposition processes, or soil carbon accumulation separately, and the interactions and possible causal relationships between these biological and ecosystem processes have rarely been explored. This case study of cormorant-derived excreta deposition in conifer plantations at Isaki Peninsula thus can provide useful implications for the understanding of biological mechanisms underlying the N-induced sequestration of soil carbon in forest ecosystems supplemented with excess

The past century is the first time since the evolution of modern N cycle linked to microbial processes with robust natural feedbacks and controls that human activities may have produced the largest impact on global N cycle (Canfield et al., 2010). Disrupted N cycles due to excess supply of N of anthropogenic origin and the concomitant buildup of N and carbon sequestration in forest soils are one of major global issues because of its potential influence on the evolution of carbon dioxide and feedback to global warming (Nadelhoffer et al., 1999; de Vries et al., 2006). The present study highlights potential importance of fungi and their indispensable roles linking N deposition with carbon sequestration in soils. Future research directions include the dynamics of phosphorus (Conley et al., 2009), another major nutrient that is abundantly contained in excreta (Hobara et al., 2005), and which limits primary production more frequently than N and has different effects on soil processes and fungal

Needles:

$$\text{NT}\_{\text{NID},t} = \left[ (0.0226 \times \text{NE} + 1.180) \times \exp^{0.27t} \right] \times \left[ (100 \times \exp^{0.27t} - 120) / (\text{-} 20) \right] \times 10 \tag{13}$$

Twigs:

$$\text{NIT}\_{\text{NID},l} = \left[ (0.\ 0104 \times \text{NE} + 0.164) \times \exp^{\text{q.039}} \right] \times \left[ (100 \times \exp^{\text{q.039}} - 110) / (\text{-23}) \right] \times 10 \tag{14}$$

CWD:

NITNDL,*t* = [(0. 0122 NE + 0.096) exp-0.02*t*] [(100 exp-0.02*t* – 105)/(-174)] 10 (15)

#### **6.4 Dynamics of dead plant tissues and excreta-derived nitrogen in colonized forest**

Using the empirical models in equations 7-9 and 13-15, long-term patterns of the remaining mass of dead plant tissues and in the N mass during decomposition were estimated for forest stands colonized by cormorants (Fig. 17). The models show the different roles of plant tissues as components of the forest floor and reservoirs of excreta-derived N.

At the time of litterfall (i.e., 0 year), needles, twigs, and CWD account for approx. 50%, 25%, and 25% of total litterfall, respectively (Fig. 17). However, needles almost disappear before 20 years of decomposition because the mass loss for them is much faster than that for twigs and CWD. After 20 years twigs and CWD constitute the dominant components of the detritus pool in the forest stand. Although not shown in Fig. 17, CWD becomes two times more important quantitatively than twigs at 60 years of decomposition.

Fig. 17. Estimated long-term changes in the remaining mass of dead plant tissues and the nitrogen mass in the remaining tissues during decomposition in forest stand colonized by cormorants at 200 nests/ha, the mean value at Isaki Peninsula in 2000 (Fujiwara, 2001). The models start with litterfall in a single year, and the figures show decomposition for 30 years.

Needles account for 87% of N in dead plant tissues at the time of litterfall because the initial N contents are 3 and 15 times higher than in twigs and CWD, respectively (Fig. 17). During the first two years of decomposition, N mass in needles increased 1.8 times compared to that of the initial mass due to the net immobilization of excreta-derived N (Section 5.2). The N mass reached its maximum value at 3 years of decomposition but decreased thereafter due to the net release from needles. The N mass in needles becomes smaller exponentially, and it almost disappears before 20 years of decomposition. This suggests that needles serve as a temporary reservoir and then as a source of N thereafter up to 20 years of decomposition. In contrast, twigs immobilize N slowly for 20 years to become the dominant reservoir of N thereafter. The model predicts that CWD (snags) plays a negligible role in N retention. Note, however, that a part of CWD can be a reservoir of N when it falls down to become logs (Section 5.3), a process not incorporated in the model.

It should also be noted that some equations in the models have low coefficients of determination. This especially holds true for the litterfall amount of CWD (Eq. 3), the changes in remaining mass of needles (Eq. 4), and the dynamics of N in twigs and CWD (Eqs. 11 and 12). When the decomposition of needles is assumed to follow an asymptotic function, for example, needles become a more important, longer-term reservoir of dead plant tissues and N in the detritus pool. Obviously, longer-term studies of tree mortality and decomposition of needles, twigs, and CWD will be necessary to construct more accurate empirical models. Still, the present models provide useful insights into the effects of the density of cormorant colonization on the amount of litterfall and into the differential roles of dead plant tissues as reservoirs of carbon and excreta-derived N.
