Determinants of Needleleaf and Broadleaf Decomposition Rates Under and Outside the Parent Tree Stand

Abstract

We studied differences in the decomposition rate between Pinus merkusii Jungh. et de Vriese (tusam) leaves, a representative of needle leaf litter, and Diospyros celebica Bakh. (ebony) leaves, a representative of broadleaf litter, in three forest communities (Karst, Lowland, Pine) on the island of Sulawesi, Indonesia, and identified their determinants. Twenty-four 1 m × 1 m quadrats were set up in each forest community to observe the in situ decomposition process. Near each quadrat, 1 m² litter traps were set to monitor litter production. In addition, 30 litter bags containing tusam leaves and 30 litter bags containing ebony leaves were spread in the three forest communities, in both the dry and wet seasons, to observe their decomposition rate during each season. The ANOVA test showed that the one-year in situ Decomposition Rate Constant (k) was significantly highest in the Karst forest (0.0921/year), followed by the Lowland forest (0.0700/year), and the lowest in the Pine forest (0.0277/year). During the dry season, the mean k-value of tusam leaves was significantly faster than ebony leaves in Karst (0.7162/6 months for tusam, 0.3840/6 months for ebony) and Lowland forests (0.3472/6 months for tusam, 0.1017/6 months for ebony), but on the contrary, it is slower in the Pine forest (0.0498/6 months for tusam, 0.0745/6 months for ebony). During the wet season, there was no significant difference between the mean k-value of tusam leaves compared to ebony leaves in the Karst (0.5217/4 months for tusam, 0.4859/4 months for ebony) and Lowland (0.2397/4 months for tusam, 0.2098/4 months for ebony) forests, but in the Pine forest, the mean k-value of ebony leaves was significantly higher than that of tusam leaves (0.0942/4 months for tusam, 0.1650/4 months for ebony). This study explains that the decomposition process of leaf litter is complex, species-specific, and is controlled by a combination of factors. Extrinsic factors play a more critical role than intrinsic factors in determining the k-value. The low rate of decomposition of tusam leaves under its mother tree stands is not caused by intrinsic factors, but rather by extrinsic factors that inhibit the growth of decomposing agents.

1. Introduction

Litter decomposition is the main pathway of the nutrient cycling process in forest ecosystems [1]. During the decomposition process, litter undergoes a biogeochemical process [2]. Leaves are the primary component of forest litter, with almost 90% of the litter produced by young stands and about 70% of the litter produced by older stands consisting of leaves [3]. Due to their thin structure, leaves are a readily biodegradable component of litter.

Forest litter decomposition is a complex process, and many factors are involved [2]. Generally, these factors can be categorized into intrinsic factors in the form of the chemical composition of the litter itself and extrinsic factors (which are also called environmental factors) that consist of biological, chemical, and physical factors [4,5]. The diversity and abundance of macrofauna and microorganisms are biological decomposing agents that play the most important role in determining the rate of litter decomposition [6,7]. Meanwhile, the chemical composition of litter may influence the chemical and physical conditions of the environment [8,9], which in turn determine the diversity and activity of decomposing agents.

Research on forest litter decomposition rates has generally been conducted in subtropical and temperate regions and has reported various results. Most studies have revealed that intrinsic factors more influence decomposition processes than extrinsic factors [10,11] and are species-specific [11,12]. Among extrinsic factors, season has been found to have the most significant influence on decomposition rates, with decomposition occurring more rapidly in summer than in winter [13] and in warmer lowlands [14]. Meanwhile, rainfall has been reported to dampen decomposition rates [11]. In contrast to temperate regions, temperature variation across time and site is not significant in tropical regions [15]. Compared to temperature, rainfall is a key climatic element that determines seasonal patterns in tropical regions, resulting in dry and wet seasons throughout the year [16]. During the dry season, air humidity can drop to levels that cannot support optimal microbial life, resulting in a significant decrease in the diversity and abundance of decomposing microorganisms [17,18]. Conversely, during the wet season, high humidity supports the growth of microorganisms. Differences in environmental conditions between temperate and tropical climates can lead to variations in the rate of decomposition of forest litter in these two regions.

Studies on coniferous forests have revealed that pine leaves decompose slowly under the stand of their mother tree [19], usually forming thick piles of litter on the forest floor. There are various intrinsic and extrinsic factors that are considered to have the potential to cause the slow decomposition of pine leaf litter [20]. So far, research has focused more on the role of intrinsic factors than extrinsic factors in the slow decomposition rate of pine leaf litter. For example, Sheffer et al. [21] revealed that pine leaves generally contain higher concentrations of lignin than broadleaf species, and that high lignin content negatively affects the biodiversity and activity of decomposing agents [22,23]. In addition to lignin content, a higher carbon-to-nitrogen ratio (C/N) in pine leaves is also known to contribute significantly to the slower decomposition process of pine leaf litter [24]. Pine needles are also known to contain various allelopathic substances that can affect the activity of decomposing agents [25]. Higher lignin content, higher C/N ratio, and allelopathic effects may directly make it difficult for decomposing agents to decompose pine leaf litter or, indirectly, create environmental conditions less favorable for their survival [26].

In tropical Indonesia, there is only one species of pine, Pinus merkusii Jungh. et de Vriese [27]. This species grows naturally on the island of Sumatra, but because of its ability to adapt to new and extreme environmental conditions, it has been widely planted as a reforestation species throughout the Indonesian archipelago since the mid-nineteenth century. As is often reported in temperate climates [28,29], studies on the decomposition rate of forest litter in tropical Indonesia have shown that P. merkusii litter decomposes more slowly [30] compared to broadleaf litter [31,32]. This has implications for the nutrient cycles and restoration of fertility of forest land resulting from reforestation using P. merkusii. To manage the expanding pine forest stands in Indonesia, more comprehensive research is needed to uncover the factors underlying the slow rate of pine litter decomposition under the parent tree stands. Between extrinsic and intrinsic factors, which plays a greater role?

We assume that, if the decomposition process of pine leaf is more influenced by intrinsic factors rather than extrinsic factors, then the pine leaf litter should not only decompose slowly under its mother tree stands but also under the mixed broadleaf forest stands. Conversely, broadleaf litter will decompose more quickly than needleleaf litter under either pine or broadleaf forest stands. By observing the decomposition rate of the needle leaf of P. merkusii (hereafter referred to as tusam leaf) and the broadleaf of Diospyros celebica Bakh. (hereafter referred to as ebony leaf) placed under P. merkusii plantation forests and under mixed broadleaf forests (karst and lowland forests), this study aims to examine whether the decomposition rate of tusam and ebony is consistent across the three forest communities. Since many publications reveal that pine leaves contain chemical compounds that have the potential to inhibit the decomposition process, we predict that the decomposition rate of pine leaf litter placed in mixed broadleaf forest ground is also slow at the same rate as when placed on the ground of its mother tree stand. Additionally, we predict that the decomposition rate of ebony leaf litter will be faster than that of tusam leaf litter in all forest stands. The results of this study will contribute to the development of solutions aimed at accelerating nutrient cycles in Pine plantation forests in Indonesia.

2. Materials and Methods

2.1. Study Site

We conducted this research in three forest communities in the southern part of Sulawesi Island, Indonesia, as follows: broadleaf secondary karst forest, broadleaf secondary lowland forest and P. merkusii plantation forest (hereafter referred to as Karst forest, Lowland forest, and Pine forest, respectively). The flat distance between the three forest communities studied was about 3 km and all were in the same climate type, namely climate type C [33]. This climate type is characterised by two seasons in a year: the dry season (monthly rainfall < 60 mm) and the wet season (monthly rainfall > 100 mm; Figure 1).

Previously, Putra et al. [34] set up a 1 ha permanent plot for a vegetation survey in each forest community. They reported that the Karst forest was dominated by broadleaf pioneer tree species (e.g., Kleinhovia hospita L., Cananga odorata (Lamk.) Hook and Pterospermum celebicum Miq.). Lowland forest was dominated by palm tree species (e.g., Areca catechu L. and Arenga pinnata Merr.) and some broadleaf climax tree species (e.g., D. celebica and Palaquium obovatum (Griff.) Engl.). Meanwhile, P. merkusii is the most dominant species in Pine forests, covering up to 88% of the tree basal area. The Karst forest is located on the middle slope at an altitude 271 m asl and astronomically at 119°44′14.9″ E and 5°01′46.8″ S. The Lowland forest is located on the relatively flat area at the lower slope at an altitude 563 m asl and astronomically at 119°46′35.0″ E and 4°58′06.9″ S. Meanwhile, the Pine forest is located on the top of a hill at an altitude 501 m asl and astronomically at 119°45′56.7″ E and 5°00′17.3″ S.

2.2. Litterfall Observation

To observe the amount of litter production, we installed 24 litter traps in permanent plots in each forest community. The litter traps were purposely distributed under the forest canopy that best represents the density and species composition of the forest canopy. Each litter trap comprised a circle with a surface area of 1 m² (diameter: 112.9 cm). We collected the litter caught in each trap monthly. The litter samples collected from each trap were placed in separate vinyl bags and brought to the laboratory to be oven-dried at 60 °C. The mean dry weight of litter accumulated over a year was considered to be the amount of litter produced per square meter in that year.

2.3. In Situ Decomposition Experiment

We carried out an in situ decomposition experiment on 24 quadrats measuring 1 m × 1 m placed in each forest community to determine the natural rate of litter decomposition in the field. Each quadrat was placed next to a litter trap. Each quadrat was delimited using a PVC pipe to prevent litter movement from outside the quadrat into the quadrat and vice versa. At the beginning of the experiment, we removed all intact and partially decomposed litter components present in each quadrat. Then, we let forest litter fall into each quadrat and decompose naturally. A year later, we collected all intact and partially decomposed litter captured in the quadrats. We placed the litter samples from each quadrat in separate vinyl bags and brought them to the laboratory. After having been cleaned of adhering soil, each litter sample was dried in an oven at 60 °C; then, the dry weight was measured. We considered this weight to be the remaining undecomposed litter from one year of production. These observations began at the same time that we started the litterfall observation.

2.4. Decomposition Rate of Tusam and Ebony Leaves Under and Outside Their Stand Origin

Experiments for examining the decomposition rate of tusam and ebony leaves were conducted sequentially in the dry and wet seasons. The leaf litter used for each seasonal experiment comprised freshly fallen leaves collected from litter traps one month before each experiment began. This was done to prevent deterioration of the litter samples before the experiment began. The leaf litter samples were cleaned of the twigs they were attached to and the leaves of other species. The cleaned leaf litter samples of tusam and ebony leaves were then dried in an oven at 60 °C until they reached a constant weight.

The dry samples of tusam leaves were weighed to approximately 40 g each using a Henherr JCS-K (Changzhou Hener Weighing Equipment Co., Ltd., Changzhou, China) balance and this was recorded as the initial weight. Each leaf sample was then put into the 30 cm × 30 cm litter bag made of nylon netting with 1 mm mesh (modified from [35]), which was numbered using Dymo tape. We prepared 60 litter bags containing tusam leaves. In the same way, we also prepared 60 litter bags containing oven-dried ebony leaves. At the onset of the dry season (1 May 2020), we placed 10 pairs of litter bags (10 bags containing tusam leaves and 10 bags containing ebony leaves) next to litter traps. All four corners of each litter bag were pegged to the ground using wires to prevent the bag from moving. We monitored each litter bag every month and recorded any physical changes that occurred in the leaves. Six months after placement (1 November 2020), at the end of the dry season, we collected all of the litter bags and brought them to the laboratory. The remaining undecomposed leaves were carefully removed from each litter bag and cleaned of adhering soil using running water on a flour sieve (mesh size: 60 µm). The use of the flour sieve was intended to prevent leaf samples that had been crushed into small particles from being carried away by the water during cleaning. After being cleaned, leaf samples from each litter bag were dried in an oven at 60 °C and cleaned of mixed sand particles, and then, their dry weight was measured. This dry weight was recorded as the remaining weight of litter that had not decomposed. Thus, this experiment was called the dry season experiment.

Similar to the experiment on the decomposition rate during the dry season, at the onset of the wet season (1 December 2020), we placed the same number of litter bags in the same places the dry season experimental litter bags had been placed. This second experiment aimed to observe the decomposition rate of tusam and ebony leaves during the wet season. As in the dry season experiment, the wet season experiments were to be conducted also for six months. However, our fourth monthly observation found that the leaf samples in the two experimental bags had been wholly decomposed. Therefore, we collected all leaf bag samples simultaneously at that time, and the experiment was terminated at four months. The same laboratory treatment used in the dry season experiment was also applied to the leaf samples which remained undecomposed in the wet season experiment.

2.5. Chemical Analysis of Tusam and Ebony Leaves

According to the chemical compounds in the leaf samples to be examined, we used different analysis methods. The methods used were as follows: (a) the Van Soest method to analyse the content of lignin, cellulose and hemicellulose and the appliance used was Fibertec; (b) the Luff-Schoorl method to analyse the carbohydrate content and the appliance used was Buret; (c) the Colorimetric—AlCl₃ method to analyse the content of flavonoids; (d) the Colorimetric—Folin–Ciocalteu method to analyse the content of phenolics and tannins. The appliance used to analyse the content of flavonoids, phenolics, and tannins were UV–Vis Spectrophotometer, and (e) the methanol extraction method to analyse the resin content. The leaf samples used for the chemical analyses were randomly taken from freshly defoliated leaves. We collected 20 g of samples of tusam leaves (consisting of approximately 660 needles) and 20 g samples of ebony leaves (consisting of approximately 26 leaves). Each sample was finely ground for use in the chemical composition analyses. For efficiency, we did not analyse the chemical composition of multiple leaf samples. The analyses of the lignin, cellulose, hemicellulose, carbohydrates, tannin, and resin were carried out at the Laboratory of Feed Chemistry of the Faculty of Animal Husbandry, Hasanuddin University. Meanwhile, the analysis of the flavonoids and phenolics was carried out at the Biochemistry Laboratory of the Faculty of Mathematics and Natural Sciences, Hasanuddin University.

2.6. Decomposing Agent

Several studies [36,37] have shown that the diversity of macrofauna and microorganisms also determines the mass loss of litter in the decomposition process. In this study, we differentiated the decomposing agents into the following three groups: (1) microorganisms consisting of bacteria and microscopic fungi, (2) macroscopic fungi, and (3) macrofauna. Previously, Putra et al. [18] studied the diversity of microscopic decomposing agents at the study sites using PCR (Polymerase Chain Reaction) analysis. In this paper, we refer to Putra et al. [18] for the diversity of the microscopic decomposing agents (

Table 1. Diversity of microscopic decomposing agents in three forest communities (√ = exist) [18].

Species	Existence			Degraded Compound
	Karst	Lowland	Pine
Bacteria
Bacillus cereus Frankland and Frankland 1889.	√			Cellulose, hemicellulose, lignin
Burkholderia sp.	√			Cellulose, hemicellulose, lignin
Burkholderia ubonensis Yabuuchi et al., 2000	√			Phosphor
Burkholderia cepacia (Palleroni and Holmes 1981) Yabuuchi et al., 1993	√		√	Cellulose, lignin
Bacillus thuringiensis Berliner 1915		√		Cellulose, hemicellulose
Burkholderia cenocepacia Vandamme et al., 2003		√	√	Phosphor, potassium
Number of bacterial species	4	2	2
Fungi
Trichoderma virens (J.H. Miller, Giddens and A.A. Foster) Arx, 1987	√			Cellulose, hemicellulose
Aspergillus aculeatus Iizuka 1953	√			Cellulose, hemicellulose
Aspergillus terreus Thom 1918	√	√		Cellulose, hemicellulose, lignin
Aspergillus japonicus Saito 1906	√	√		Cellulose
Penicillium pinophilum Hedgc. 1907		√		Cellulose, hemicellulose, lignin
Trichoderma sp.		√		Cellulose, hemicellulose, lignin
Aspergillus sp.		√		Cellulose, lignin
Penicillium citrinum Thom 1910			√	Cellulose, lignin
Cladosporium tenuissimum Cooke 1878			√	Cellulose, hemicellulose, lignin
Talaromyces pinophilus (Hedgc.) Samson, Yilmaz, Frisvad and Seifert 2011	√	√	√	Cellulose, phosphor
Number of fungal species	5	6	3
Total Bacterial and Fungal	9	8	5

).

Macroscopic fungal diversity was observed in 25 sub-plots measuring 10 m × 10 m, which were systematically distributed throughout the permanent plot at each study site. For each macroscopic fungal colony we found in the sub-plots, we photographed it and recorded the specimen number. Photographs were used to identify the species’ names in the laboratory using several references. The abundance (colony cover area) of each macroscopic fungal species was not measured because the fruiting bodies of different species appear at different times. Data of macroscopic fungal colony cover from instantaneous surveys may not have reflected the predominance of fungal species in an area. During one period of survey, the clumps of some macroscopic fungi species may have only just reached the hyphal stage, while others have grown to pinhead, complete fruiting bodies, or have even decayed. Observations on macroscopic fungi were conducted twice: in the dry season and in the wet season.

Macrofauna diversity was observed using two different methods. The first method used pitfall traps made of plastic cups with a cup mouth diameter of 8.5 cm and a depth of 15 cm. The trapped macrofauna was collected daily for seven consecutive days. Macrofauna trapped in pitfall traps were collected and separated by species, placed in specimen bottles containing 70% ethanol, numbered, and then the number of individuals was counted. The second method used square ring samples with a side length of 20 cm and a depth of 10 cm. The square ring sample was plugged into the soil; then, the part of the soil that was inside the ring was excavated and placed in a plastic basin. The macrofauna species collected from the square ring samples were also put into separate specimen bottles containing 70% ethanol and numbered; then, the number of individuals was counted. Both the first method and the second method were replicated 10 times in each forest community and repeated twice: in the dry and in the wet seasons. Each trap was placed close to where the litter traps, experimental leaf litter bags, and in situ decomposition experiments were installed. The macrofauna species identification was carried out at the Forest Protection Laboratory and the Integrated Laboratory of the Faculty of Forestry, Hasanuddin University.

2.7. Soil Chemistry

A total of eight soil samples were taken near the location of the litter trap placement and other experiments on each permanent plot using a sample ring with a diameter of 7 cm and a depth of 10 cm. The soil samples were taken to the Laboratory of Chemistry and Soil Fertility, Faculty of Agriculture, Hasanuddin University, to be analysed for various chemical properties, such as pH, C, N, C/N, P, Ca, Mg, K, and Na content, cation exchange capacity (CEC), and base saturation (BS).

2.8. Climate and Soil Moisture

Rainfall data were obtained from the Indonesian government agency assigned to record climate data, that is, the Meteorology, Climatology, and Geophysics Agency (BMKG) in Maros Regency. Local maximum and minimum temperatures were measured by installing a maximum–minimum thermometer on tree trunks at 1 m above ground level on each permanent plot. Then, the temperature data were recorded monthly. Measurement of the soil surface moisture was also carried out every month near the litter trap locations using the Takemura Soil pH and Humidity Tester DM-5 (Takemura Electric Works Ltd., Nishi-Ikebukuro, Toshima-ku, Tokyo 171-0021, Japan). All monthly data collections were carried out between the first and third of every month.

2.9. Data Analyses

We assumed the difference between the mean weight of the litter collected in the litter traps over a year and the mean weight of the remaining undecomposed litter in the 1 m × 1 m quadrats to be the amount of in situ decomposed litter in 1 sqm during the experimental period. To determine the decomposition rate constant (k) of the tusam and ebony leaf litter in the experiment using the litter bags, we used the following formula [38]:

$${\displaystyle \frac{{\mathrm{W}}_{\mathrm{t}}}{{\mathrm{W}}_{\mathrm{O}}}}={\mathrm{e}}^{-k\mathrm{t}.}$$

(1)

where Wt is the weight of the litter at time t, Wo is the initial weight of the litter, k is the decomposition rate constant, e is the natural logarithm, and t is the time of decomposition.

Analysis of Variance (ANOVA) with Tukey Honest Significant Difference (HSD) method was used to detect the differences in (a) the mean litter production, mean unbiodegradable litter, mean estimated in situ decomposed litterfall and mean decomposition rate constant (k) of the in situ decomposed litterfall; (b) the mean k-value between the tusam and ebony leaves in each forest community and the mean k-value of the tusam and ebony leaves among the three forest communities; (c) the number of species of macroscopic fungi among the forest communities; (d) the number of species and density of soil macrofauna among the forest communities; and (e) the soil physical and chemical properties among the forest communities. Data that were not normally distributed were compared using the independent sample nonparametric K with the Kruskal–Wallis test. We tested the normality distribution of our data using the Shapiro–Wilk test. The correlation between biological and environmental factors and the decomposed mass of leaf samples was tested using Pearson analysis when the data were normally distributed and Spearman analysis when the data were not normally distributed. All statistical analyses were performed using the R version 2025.09.1+401 application [39].

3. Results

3.1. Litterfall and In Situ Decomposition Rate Among the Forest Communities

The mean annual litter production (ALP) in the Lowland forest was significantly greater than that in the Karst and Pine forests (p = 0.0015; Figure 2 left). However, the mean of the remaining undecomposed litterfall (RUL) was significantly highest in the Pine forest, followed by the Lowland forest, and the lowest was in the Karst forest (p < 0.001; Figure 2 centre left). Assuming the ALP was the amount of litter produced in one year (Wo) and the RUL was the amount of litter that did not decompose after one year (Wt), the amount of litter that decomposed naturally in situ (Wo − Wt) did not differ between the Karst forest and the Lowland forest but was significantly lower in the Pine forest (p < 0.001; Figure 2 centre right). However, when we analysed the mean k-value among the forest communities, we found that the mean k-value was significantly highest in the Karst forest, followed by the Lowland forest, and was lowest in the Pine forest (p < 0.001; Figure 2 right).

3.2. Decomposition Rate of Tusam and Ebony Leaves Under and Outside Its Stand Origin

The mean k-value of the tusam leaves in the dry season was significantly higher than that of the ebony leaves in the Karst forest (p = 0.0086) and the Lowland forest (p = 0.0449) but significantly lower than that of the ebony leaves in the Pine forest (p = 0.0115; Figure 3, left). During the wet season, the mean k-value of the tusam leaves was not significantly different from that of the ebony leaves in the Karst forest (p = 0.7791) and the Lowland forest (p = 0.4212). Similar to the results from the dry season, the mean k-value of the tusam leaves in the wet season was also significantly lower than that of the ebony leaves in the Pine forest (p < 0.001; Figure 3, right).

Comparisons between forest communities showed that the mean k-value of the tusam leaves during the dry season was significantly highest in the Karst forest, followed by the Lowland forest, and was lowest in the Pine forest (p < 0.001). The mean k-value of the ebony leaves in the dry season showed the same pattern as that of the tusam leaves (p < 0.001) (Figure 3, left). The same pattern of the mean k-value was also confirmed for the tusam leaves in the wet season. It was highest in the Karst forest, followed by the Lowland forest, and lowest in the Pine forest (p < 0.001). The mean k-value of the ebony leaves was also highest in the Karst forest during the wet season but was not significantly different between the Lowland and Pine forests (p < 0.001; Figure 3, right).

3.3. Chemical Composition of Leaf Litter

The chemical contents of flavonoids, lignin, carbohydrates, hemicellulose, and cellulose in the tusam leaves were higher than those in the ebony leaves (Figure 4). On the contrary, the contents of phenolics, tannins, and resin were higher in the ebony leaves than in the tusam leaves.

3.4. Macroscopic Fungi

Among the three forest communities, the mean number of macroscopic fungi species in the dry season was significantly highest in the Lowland forest, followed by the Karst forest, and lowest in the Pine forest (p < 0.0001; Figure 5). Meanwhile, in the wet season, the mean number of macroscopic fungi species was not significantly different between the Karst forest and Lowland forest but was much lower in the Pine forest (p < 0.001).

3.5. Macrofauna

The mean number of species of macrofauna caught in the pitfall traps in the dry season was significantly highest in the Karst forest, followed by the Lowland forest, and lowest in the Pine forest (p = 0.035;

Table 2. Comparison of the abundance of soil macrofauna in the forest communities. Different letters in each row indicate significant differences.

Parameter	Forest Communities
	Karst	Lowland	Pine
Caught Using Pitfall Traps in the Dry Season (per 567.163 cm2)
Mean number of species	4.8 (±0.34) a	4.1 (±0.34) b	3.3 (±0.34) c
Mean density of macrofauna	36.4 (±4.02) d	34.2 (±4.02) d	17.2 (±4.02) e
Caught Using Pitfall Traps in the Wet Season (per 567.163 cm2)
Mean number of species	5.6 (±0.43) a	5.6 (±0.43) a	4.0 (±0.43) b
Mean density of macrofauna	73.6 (±5.45) c	60.1 (±5.45) c	38.2 (±5.45) d
Caught Using Square Ring Samples in the Dry Season (per 4000 cm2)
Mean number of species	5.8 (±0.27) a	6.5 (±0.27) a	4.3 (±0.27) b
Mean density of macrofauna	69.6 (±12.57) c	24.0 (±12.57) d	35.2 (±12.57) cd
Caught Using Square Ring Samples in the Wet Season (per 4000 cm2)
Mean number of species	10.0 (±0.44) a	9.1 (±0.44) a	6.4 (±0.44) b
Mean density of macrofauna	39.1 (±2.85) c	41.3 (±2.85) c	20.6 (±2.85) d

). The mean density of the macrofauna was not significantly different between the Karst forest and the Lowland forest, but was much lower in the Pine forest (p = 0.0039). In the wet season, no significant differences were detected in the mean number of species between the Karst forest and the Lowland forest, but it was significantly lower in the Pine forest (p = 0.0200). Similarly, the mean macrofauna density did not significantly different between the Karst and the Lowland forests; however, it was significantly lower in the Pine forest compared to the other two forest communities (p < 0.001).

For the sampling using square ring samples taken during the dry season (Table 2), the mean number of species of macrofauna between the Karst forest and the Lowland forest was not significantly different but was significantly lower in the Pine forest (p < 0.001). Meanwhile, the mean density of the macrofauna was highest in the Karst forest, but the difference was only statistically significant when compared with the mean density in the Lowland forest (p = 0.042). During the wet season, the mean number of species was not significantly different between the Karst forest and the Lowland forest, but was significantly lower in the Pine forest (p < 0.001). Similarly, the mean density of the macrofauna in the wet season showed the same trend as the mean number of species. It was not significantly different between the Karst forest and the Lowland forest, but was significantly lower in the Pine forest (p < 0.001).

3.6. Soil Chemical Properties

Aside from the pH, C, and P, the content of all other soil chemicals (N, C/N, P, Ca, Mg, K, Na) and the CEC and BS did not differ significantly among the forest communities (

Table 3. Soil chemical properties in the dry and wet seasons in permanent plots in three forest communities. Different letters in each row indicate significant differences.

Parameter	Forest Communities
	Karst	Lowland	Pine
Chemical Properties
pH	6.28 (±0.05) a	6.39 (±0.05) a	5.93 (±0.05) b
C (%)	2.30 (±0.11) c	2.33 (±0.11) c	1.61 (±0.11) d
N (%)	0.22 (±0.01) e	0.22 (±0.01) e	0.17 (±0.01) e
C/N (%)	10.75 (±0.32) f	10.63 (±0.32) f	9.63 (±0.32) f
P (ppm)	12.11 (±0.48) gh	10.40 (±0.48) g	12.73 (±0.48) h
Ca (kg−1)	6.42 (±0.71) i	6.10 (±0.71) i	5.40 (±0.71) i
Mg (kg−1)	0.90 (±0.44) j	1.95 (±0.44) j	2.01 (±0.44) j
K (kg−1)	0.48 (±0.05) k	0.36 (±0.05) k	0.43 (±0.05) k
Na (kg−1)	0.47 (±0.05) l	0.41 (±0.05) l	0.36 (±0.05) l
CEC (kg−1)	19.79 (±0.67) m	20.33 (±0.67) m	18.15 (±0.67) m
BS (%)	41.63 (±4.29) n	44.88 (±4.29) n	45.63 (±4.29) n

). Soil pH and C content did not differ significantly between the Karst and the Lowland Forest, but they were significantly lower in the Pine forest (p < 0.001 for pH; p = 0.0023 for C). P content in the Pine forest was significantly higher than in the Lowland forest (p = 0.0197), but did not differ significantly from the Karst forest.

3.7. Soil Moisture and Maximum and Minimum Temperature

The mean soil moisture in the Karst forest (p = 0.0187) and in the Lowland forest (p = 0.0416) was lower in the dry season than in the wet season, but this was not the case in the Pine forest (p = 0.1765) (Figure 6, left). Neither the maximum nor the minimum temperature showed significant differences between seasons in all forest communities (Figure 6, centre and right).

When compared among the forest communities, the mean soil moisture showed no significant difference between the Karst and Lowland forests in the dry season. However, in the Pine forest, it was significantly lower than in the other two forest communities (p = 0.0054). In the wet season, the mean soil moisture was significantly highest in the Karst forest and lowest in the Pine forest (p < 0.001). Meanwhile, no significant differences were detected between the mean maximum and minimum temperatures among the forest communities in either the dry or wet seasons.

3.8. The Correlation Between Biological Factors and the Decomposed Leaf Sample

No significant correlation was detected between the macroscopic fungal colony cover and the decomposed mass of tusam leaf samples across forest communities during the dry season. Meanwhile, during the wet season, a significant positive correlation was detected in the Karst and Lowland forests, but was not detected in the Pine forest. As well as between the colony cover of macroscopic fungi and the decomposed mass of tusam leaf sample, correlation analyses between the colony cover of macroscopic fungi and the decomposed mass of ebony leaf samples during the dry season also did not show a correlation in all forest communities. However, during the wet season, a significant positive correlation was detected in the Lowland forest, and was not detected in the other two forest communities (see

Table 4. Correlation analysis between extrinsic biological factors and the decomposed mass of leaf samples.

Leaf Sample	Forest Type	Season	Biological Factors
			Macroscopic Fungi Colony Cover	Macrofauna Abundance
Tusam	Karst	Dry	r = 0.433 p = 0.211	r = 0.681 * p = 0.030
		Wet	r = 0.830 ** p = 0.003	r = 0.861 ** p = 0.001
	Lowland	Dry	r = 0.340 p = 0.336	r = 0.975 ** p < 0.0001
		Wet	r = 0.871 ** p = 0.001	r = 0.817 ** p = 0.004
	Pine	Dry	r = −0.207 p = 0.567	r = −0.361 p = 0.305
		Wet	r = 0.037 p = 0.918	r = 0.653 * p = 0.041
Ebony	Karst	Dry	r = 0.629 p = 0.051	r = −0.222 p = 0.538
		Wet	r = 0.096 p = 0.792	r = 0.309 p = 0.386
	Lowland	Dry	r = 0.171 p = 0.637	r = −0.118 p = 0.745
		Wet	r = 0.632 * p = 0.050	r = 0.700 * p = 0.024
	Pine	Dry	r = −0.197 p = 0.586	r = −0.361 p = 0.305
		Wet	r = 0.057 p = 0.876	r = −0.507 p = 0.140

Note: * = Correlation is significant at the 0.05 level; ** = Correlation is significant at the 0.01 level.

for statistical values).

The correlation between the abundance of macrofauna and the decomposed mass of tusam leaf samples during the dry season showed a significant positive correlation in the Karst and Lowland forests, but did not in the Pine forest. During the wet season, a significant positive correlation was detected in all forest communities. Meanwhile, there was no significant correlation detected between the abundance of macrofauna and the decomposed mass of ebony leaf samples during the dry season in all forest communities. However, during the wet season, a significant positive correlation was detected only in the Lowland forest (see Table 4 for statistical values).

3.9. The Correlation Between Extrinsic Physical Factors and Decomposed Leaf Samples

Due to the non-uniformity in the number of soil samples analysed for chemical properties and the number of leaf samples examined for decomposition rate, we could only test the correlation between soil pH and decomposed mass of leaf samples, as well as the correlation between soil moisture and decomposed mass of leaf samples. There was no significant correlation detected between soil pH and the decomposed mass of tusam leaf samples during the dry season in all forest communities. During the wet season, significant positive correlations were detected in the Karst and Lowland forests, but was not detected in the Pine forest. Similar to tusam leaf samples, the decomposed mass of ebony leaf samples was not detected to correlate with soil pH during the dry season in all forest communities. During the wet season, significant positive correlations were detected in the Karst and Lowland forest, but was not detected in Pine forest.

In the dry season, neither the mass of decomposed tusam leaf samples nor the mass of decomposed ebony leaf samples was detected to be correlated with soil moisture. In the wet season, a significant positive correlation between soil moisture and the decomposed mass of tusam leaf sample was detected only in the Karts forest, and was not detected in the Lowland and Pine forests. For the ebony leaf experiment, a significant correlation between soil moisture and the decomposed mass of leaf samples was not detected in all forest communities, in both the dry and wet seasons (see

Table 5. Correlation analysis between physical factors and the decomposed mass of leaf samples.

Leaf Sample	Forest Type	Season	Physical Factors
			Soil pH	Soil Moisture
Tusam	Karst	Dry	r = −0.141 p = 0.698	r = −0.158 p = 0.663
		Wet	r = 0.677 * p = 0.032	r = 0.797 ** p = 0.006
	Lowland	Dry	r = 0.437 p = 0.206	r = 0.088 p = 0.81
		Wet	r = 0.666 * p = 0.036	r = 0.543 p = 0.105
	Pine	Dry	r = 0.479 p = 0.161	r = −0.553 p = 0.098
		Wet	r = 0.008 p = 0.983	r = −0.275 p = 0.442
Ebony	Karst	Dry	r = −0.357 p = 0.311	r = 0.613 p = 0.059
		Wet	r = 0.774 ** p = 0.009	r = 0.364 p = 0.301
	Lowland	Dry	r = 0.033 p = 0.928	r = 0.012 p = 0.973
		Wet	r = 0.744 ** p = 0.014	r = 0.44 p = 0.203
	Pine	Dry	r = −0.007 p = 0.984	r = −0.273 p = 0.445
		Wet	r = −0.173 p = 0.634	r = −0.278 p = 0.436

Note: * = Correlation is significant at the 0.05 level; ** = Correlation is significant at the 0.01 level.

for statistical values).

4. Discussion

The main objective of this study was to comprehensively identify the factors that determine differences in decomposition rate between needle leaf litter and broadleaf litter in a tropical climate region. As the first step, we conducted an in situ experiment to confirm whether the rate of decomposition in Pine forests is slower than that in broadleaf forests, as has been frequently reported in temperate regions [28,29,40]. The results of the in situ experiment showed that the mean k-value of the natural decomposition of the local-origin litter in the Pine forest was slower than that of the local-origin litter in the mixed broadleaf forests (Karst forest and Lowland forest). Meanwhile, between the mixed broadleaf forest communities, the mean k-value of the natural decomposition of local-origin litter in the Karst forest was significantly higher than that in the Lowland forest.

In line with the mean k-value of the in situ experiments, but contrary to our predictions, the k-value of the tusam leaf litter in the mixed broadleaf (Karst and Lowland) forest was significantly faster than that in the Pine forest during both the dry and wet seasons. Aside from that between the Lowland forest and Pine forest during the wet season, the mean k-value of the ebony leaf litter was also significantly higher in the mixed broadleaf forests compared to that in the Pine forest. This seemed to be related to the diversity and abundance of decomposing agents, such as the microbes (bacteria and microscopic fungi), macroscopic fungi, macrofauna on the soil surface and macrofauna in the soil, which were significantly higher in the Karst and Lowland forests compared to the Pine forest. Correlation analysis between macroscopic fungal colony cover, as well as macrofauna abundance, and the mass of decomposed leaf samples showed a significant positive correlation, but the correlation coefficient appeared stronger in the Karst and Lowland forests compared to the Pine forest. The correlation coefficient also appeared stronger in tusam leaf samples compared to ebony leaf samples.

Previously, Begum et al. [41] and Qi et al. [42] reported that the abundance of soil microorganisms and macrofauna is generally high at high pH levels. In addition, higher soil C content [43] as well as higher soil moisture [44] substantially increase the microbial biomass and activity on the forest floor. This study found that the soil pH, soil C content, and soil moisture in the mixed broadleaf forests were significantly higher than those in the Pine forest. This may have been the cause of the high diversity and abundance of microorganisms and macrofauna acting as decomposing agents in the Karst and Lowland forests. Correlation analyses showed that soil pH significantly correlated with the decomposed mass of leaf samples during the wet season in the Karst and Lowland forests. Soil moisture was detected to significantly correlate with the decomposed mass of leaf samples only in the Karst forest during the wet season. Meanwhile, significant correlations between macroscopic fungal colony cover, as well as macrofauna abundance, and the mass of decomposed leaf samples generally occur in the wet season. This finding explains that soil extrinsic factors, such as soil pH, soil C content, and soil moisture, indirectly accelerate the decomposition process by supporting the abundance of decomposing agents, particularly in the wet season.

The finding that the tusam leaf litter decomposed faster than the ebony leaf litter in the mixed broadleaf forests, which was not in accordance with our prediction, may have been related to the lower content of secondary metabolites such as phenolics, tannins and resins in the tusam leaf litter compared to the ebony leaf litter. Previous studies have found that plants produce phenolic compounds, tannins and resins as a defence against herbivores [45,46] and, therefore, are often used as indicators of litter quality [47]. Fenner and Freeman [48] explained that phenolic compounds dissolved during the decomposition process could indirectly inhibit the enzymatic activity of decomposing bacteria and fungi. Although not considered toxic, high tannin content often has a lethal effect on several species of aquatic decomposing agents [49]. In addition, a higher resin content can also result in a slower decomposition rate [50].

Several previous studies have stated that lignin is the main factor behind the slow decomposition of tusam leaf litter [21,40], but our study found differently. The results of this study showed that it is true that tusam needle litter decomposes slowly under the stand of its mother tree. However, in the mixed broadleaf forests, it decomposed very quickly and even faster than the ebony leaf litter, which comprised broadleaf species. The high decomposition rate of the tusam needles in the mixed broadleaf forests could be attributed to the disparity in the number of species of decomposing microorganisms found in the three forest communities studied. More bacterial and microfungal species were found in the mixed broadleaf forests than in the Pine forest, and most of the species were lignin-degrading species.

Surprisingly, however, in the Pine forest, this study found that the mean k-value of the tusam leaf litter was significantly slower than that of the ebony leaf litter in contrast to the mean k-value of the tusam leaf litter in the mixed broadleaf forests (Karst and Lowland). This happened consistently in both the dry season experiments and the wet season experiments. No data on any observed intrinsic or extrinsic factors could be attributed to these findings. This shows that there are still factors that affect the rate of litter decomposition that have not been revealed in our study. It is not impossible that Pine forests contain microorganisms that produce enzymes capable of degrading the high phenolic and tannin content of ebony leaves. This will be an important topic for future studies. This study explains how extrinsic and intrinsic factors are complexly interconnected in determining the rate of decomposition in a forest community. Extrinsic factors consist of chemical, physical, and biological factors that influence each other to create conditions in which mixed broadleaf forests can support a higher diversity and abundance of decomposing agent organisms. The higher soil pH, C concentration, and soil moisture in mixed broadleaf forests provide a suitable environment for many species of both microorganisms and macrofauna decomposing agents to thrive [42,43]. Needle leaves and the lack of trees in the sub-canopy layer in Pine forests allow the intensity of sunlight to reach the forest floor to be greater than in broadleaf forests [51], and this, in turn, can cause a decrease in soil moisture due to high evaporation [52]. High-intensity sunlight and low soil moisture may be unfavorable for the growth of decomposing agents, especially microorganisms [53,54].

This finding does not fully support the results of previous studies which have found that broadleaf litter decomposes faster than needle leaf litter [24,40]. This study illustrates how the decomposition process of leaf litter is species-specific, site-specific, and controlled by a combination of multiple extrinsic factors. Our findings highlight that slow decomposition in Pine forests is not predominantly caused by intrinsic factors (i.e., the chemical composition) but rather occurs to the environmental conditions of the forest ecosystem itself. To illustrate how intrinsic factors may not play an important role in determining the rate of decomposition, Lorenz et al. [28] revealed that based on their experiments on the litter decomposition of four European tree species, including Scottish pine, neither the tannin nor phenolic concentrations could explain the significant variation in litter mass loss.

5. Conclusions

From the findings of this study, it can be concluded that the process of litter decomposition in forests is complex and involves various intrinsic (chemical content of the litter) and extrinsic (biological, physical, and chemical) factors. The rate of litter decomposition is determined by a reciprocal relationship between intrinsic and extrinsic factors. All chemical compounds contained in litter can be degraded if there is an appropriate decomposing agent (biological factors) at the site. Meanwhile, the biodiversity and the abundance of decomposing agents (extrinsic biological factors) are influenced by the physical and chemical properties of the soil (extrinsic physical and chemical factors), and those properties are determined by the chemical content of the litter (intrinsic factors). The slower decomposition rate of tusam leaf litter in the Pine forest compared to mixed broadleaf forest is not primarily due to intrinsic factors (e.g., litter chemical composition) but rather to extrinsic factors within the forest ecosystem itself. This is because tusam leaf litter decomposed more quickly in the mixed broadleaf forests than Pine forest. Among extrinsic factors, biological factors play a more dominant role compared to physical and chemical factors.