Logged peat swamp forest supports greater macrofungal biodiversity than large‐scale oil palm plantations and smallholdings

Abstract Intensive land expansion of commercial oil palm agricultural lands results in reducing the size of peat swamp forests, particularly in Southeast Asia. The effect of this land conversion on macrofungal biodiversity is, however, understudied. We quantified macrofungal biodiversity by identifying mushroom sporocarps throughout four different habitats; logged peat swamp forest, large‐scale oil palm plantation, monoculture, and polyculture smallholdings. We recorded a total of 757 clusters of macrofungi belonging to 127 morphospecies and found that substrates for growing macrofungi were abundant in peat swamp forest; hence, morphospecies richness and macrofungal clusters were significantly greater in logged peat swamp forest than converted oil palm agriculture lands. Environmental factors that influence macrofungi in logged peat swamp forests such as air temperature, humidity, wind speed, soil pH, and soil moisture were different from those in oil palm plantations and smallholdings. We conclude that peat swamp forests are irreplaceable with respect to macrofungal biodiversity. They host much greater macrofungal biodiversity than any of the oil palm agricultural lands. It is imperative that further expansion of oil palm plantation into remaining peat swamp forests should be prohibited in palm oil producing countries. These results imply that macrofungal distribution reflects changes in microclimate between habitats and reduced macrofungal biodiversity may adversely affect decomposition in human‐modified landscapes.


| INTRODUCTION
In the past decades, the expansion of agricultural plantations replacing peatland has occurred at an alarming rate, particularly in povertystricken regions of Southeast Asia. In 1990 to 2010 tropical peat swamp from Malaysia and Sumatra, Indonesia were reduced from 77% to 36%, with only 9% of peatland areas receiving protection (Miettinen, Shi, & Liew, 2012;Posa, Wijedasa, & Corlett, 2011). Despite providing essential services such as soil erosion control, ecosystem stabilization, and carbon storage (Yule, 2010), peat swamp forests have been treated as wastelands (Rijksen & Peerson, 1991). As trees are felled to establish oil palm plantations, the peat swamp forest is drained and resulting decomposition releases substantial carbon emissions into the atmosphere (Murdiyaso, Hergoualc'h, & Verchot, 2010). This is attributed to the fact that cumulative CO 2 emissions decrease during the high water table conditions, but the emissions may increase during the low water table conditions (Jauhiainen, Takahashi, Heikkinen, Martikainen, & Vasander, 2005). Nowadays, peat swamp forests are being threatened by extensive fire and human exploitations by logging and agricultural industry (Muhammad & Abdullah, 2015). Conversion of forest land cover to agricultural plantations is responsible for causing habitat destruction and loss of forest biodiversity (Rudel, Defries, Asner, & Laurence, 2009;Sodhi & Brook, 2006).
The impact of oil palm plantation expansion on peat swamp. biodiversity is understudied (Posa et al., 2011) which raises numerous concerns. Efforts to understand and conserve peat swamp biodiversity are limited by a lack of information about many taxonomic groups, particularly those groups that are most species rich. Species richness and species abundance are reduced when original land cover changes (Danielsen et al., 2009;Fitzherbert et al., 2008;Foster et al., 2011).
These baseline data are mostly available from well-known taxa such as birds and mammals (Azhar et al., 2011(Azhar et al., , 2013Hawa, Azhar, Top, & Zubaid, 2016;Mandal & Shankar Raman, 2016;Prabowo et al., 2016;Sasidhran et al., 2016;Syafiq et al., 2016). There is an urgent need to understand the effects of peat swamp forest conversion to oil palm agriculture on macrofungal diversity. Although macrofungi have important functions in decomposition, nutrient cycling, and nutrient uptake, little is known about macrofungal diversity, as it is ephemeral and has enigmatic growing patterns which make identification difficult (Schmit et al., 2005).
Macrofungi are sensitive to habitat modification (Brown, Bhagwat, & Watkinson, 2006;Halme et al., 2013). Previous studies have found that macrofungal diversity would be affected directly in countries experiencing modification in land cover (Kasel, Bennett, & Tibbits, 2008;López-Quintero, Straastsma, Franco-Molano, & Boekhout, 2012;Paz, Gallon, Putzke, & Ganade, 2015), especially when the floristic composition and structural characteristics are altered (Brown et al., 2006;Gómez-Hernández & Williams-Linera, 2011). The conversion of native forests to exotic crop plantations has been found to lower the number of macrofungal decomposer species, most likely due to changes in substrate availability and quality (Heilmann-Clausen et al., 2014;Paz et al., 2015). Despite their functional importance, knowledge of macrofungal diversity is still lacking in both peat swamp forest and oil palm production landscapes due to the paucity of information about these taxa. The data sets for macrofungal diversity in major geographical regions of the world are incomplete, and thus, the existing numbers of macrofungal species represent very conservative estimates for macrofungal diversity in each region (Lodge, 1997;Lodge & Cantrell, 1995;Mueller & Schmit, 2007;Piepenbring, 2007). Lacking data on macrofungal diversity with visible fruiting bodies in oil palm plantation is one of the largest knowledge gaps for fungi in peatlands. Thus, detailed studies of the effect of peat swamp and forest conversion to oil palm on macrofungal communities is a priority for conservation research.
In 2008, forest fires occurred extensively in Malaysia, degrading the forest area particularly in the North Selangor Peat Swamp Forest (NSPSF), and ultimately resulting in 1,231 ha of the country's second largest peat swamp forest being converted into an oil palm plantation (Yule & Gomez, 2008). This means that once this area became degraded, it was easier for oil palm to expand and use the area.
Encouraged by strong global market demand in oil palm products, large-scale plantations and smallholdings currently surround at least 60% of the NSPSF perimeter, and more NSPSF land is scheduled for clearance to make way for plantations (Azhar et al., 2011).
Our study provides new information related to the biodiversity patterns of macrofungal diversity in human-modified peat soil habitats ( Figure 1). This baseline information is essential to formulate successful conservation strategies (Brown et al., 2006;Lindenmayer et al., 2012). First, we compared morphospecies richness and number of macrofungal clusters between logged peat swamp forest, oil palm plantation (>50 ha; private business), and smallholdings (<50 ha; independent farmer) including those that were either monoculture or polyculture system. Secondly, we contrasted vegetation structure and substrate availability between the four habitat types. Thirdly, we determined key environmental and vegetation structure attributes which influence macrofungal diversity.

| Study area description
We conducted this study at the North Selangor Peat Swamp Forest (NSPSF) and nearby oil palm planted areas. Surveys were conducted from November 2015 to January 2016 during the Northeast and Southwest monsoon season. The mean temperature of the monsoon months in the study area is 31°C; rainfall is between 200 and 500 mm annually (MMD (Malaysian Meteorological Department), 2015).
The NSPSF (N 3°40′26.56″, E 101°4′29.52″) is located at the north western part of Selangor with an elevation of 16 m above sea level. The NSPSF is a secondary forest embracing an area of 73,593 ha where 95% of the area is logged peat swamp forest and F I G U R E 1 Fruiting bodies of Lichenomphalia sp. are commonly encountered in peat swamp forest, but are absent in oil palm cultivation areas 5% is dipterocarp forest (Azhar et al., 2011(Azhar et al., , 2013 (Parish et al., 2014). Currently, the NSPSF is being threatened by forest fire and oil palm plantation expansion (Azhar et al., 2011;Sasidhran et al., 2016). Despite being designated as forest reserve, more than 1,000 ha of NSPSF has been cleared for plantation purposes (Sasidhran et al., 2016). Sungai Dusun Wildlife Reserve is the only hope for biodiversity conservation in the area as it has been formally appointed as a protected area (Adila et al., 2017;Sasidhran et al., 2016).
Agricultural plantations surveyed in this study were originally peat land but are now planted with oil palm (Elaeis guineensis), aged 8-years at the time of sampling. Large-scale plantations, covering an area of 2,000 ha, were managed by a large corporation with the use of advanced machinery (Azhar et al., 2011). Smallholdings, defined as semi-traditional cultivation area of less than 50 ha, were managed by small-scale farmers and were less dependent on modern infrastructure (Azhar et al., 2015). Two different smallholder landscapes were studied: monoculture and polyculture. Monoculture smallholdings were planted mainly with oil palms with no intercropping. Polyculture smallholdings, on the other hand, often practiced intercropping, where oil palm plants were planted side by side with subsidiary commercial crops such as banana, coconut, cassava, coffee, pineapple, mangoes, jackfruit, and tapioca ( Figure 2).

| Sampling design
We used systematic sampling with a random starting point (Morrison, Block, Strickland, Collier, & Peterson, 2008). First plot was randomly

| Macrofungal sampling
A team comprising five people conducted direct searches within each sampling plot. The team spent 20-30 min at each sampling plot to search for macrofungal specimens.
To increase the chances of encountering macrofungi, each sampling plot was visited after a rainy day because mushroom fruiting bodies are likely to appear in the most humid conditions (Henkel, Meszaros, Aime, & Kennedy, 2005;López-Quintero et al., 2012).
Throughout this study, macrofungi were identified following Lodge We concentrated our sampling on visible basidiomycetes and ascomycetes sporocarps detected on the forest floor, fallen logs, as well as on living and dead standing trees (Paz et al., 2015). At each sampling plot, we recorded the number of macrofungal clusters. A cluster was recorded as one observation irrespective of the number of sporocarps in that cluster (Brown et al., 2006). Voucher specimens were also collected and stored in a multiple partition plastic box for ex situ identification. Sporocarp surveys represent a cost-effective and reliable method to survey macrofungi, because they concentrate on the reproductive parts of the fungal species (Halme, Heilmann-Clausen, Rämä, Kosonen, & Kunttu, 2012;Paz et al., 2015).

| Measurement of environmental variables and vegetation structure characteristics
At each sampling plot, environmental variables comprising air temperature, humidity, soil pH, soil moisture, and wind speed were recorded (

| Statistical analysis
We contrasted the observed morphospecies richness with the Chao 1 bias correction estimator for the species richness in EstimateS version 9.1 to assess the overall sampling effort (Colwell, Mao, & Chang, 2004). We used ACE (Abundance Coverage-based Estimator) to take into account imperfect detection of rare species (Colwell & Coddington, 1994 To determine the contribution of dominant morphospecies to the macrofungal community, we performed one-way analysis of similarity percentage (SIMPER). Prior to the analysis, the number of morphospecies count data was square root transformed. Bray-Curtis index was performed to calculate the compositional dissimilarity between different habitats. We used a nonmetric multidimensional scaling (NMDS) to visualize difference in species composition between habitat types.
We used PRIMER version 6 (PRIMER-E Ltd, Plymouth) to perform all multivariate analyses.
We examined the spatial autocorrelation in residuals by calculating Global Moran's Index in the ArcGIS™ version 10.1 (ESRI). The p value was used to reject or accept the null hypothesis which states that the analyzed attribute is randomly distributed among the features in the study area (Mitchell, 2005). We used inverse distance (nearby neighboring features have a larger influence on the computations for a target feature than features that are far away) to calculate Global Moran's Index.

| RESULTS
A total of 757 macrofungal clusters were collected, representing 127 morphospecies where 43.07% (n = 326 clusters) clusters were recorded from logged peat swamp forest. We recorded 19.68% (n = 149 clusters) in large-scale oil palm plantation. Monoculture and polyculture smallholdings had 15.59% (n = 118 clusters) and 21.66% (n = 164 clusters), respectively. With respect to the sampling completeness, the sampling effort in logged peat swamp forest was compared with the Chao 1 and ACE estimators, yielding 64% and 71% of the "true" species richness for macrofungi, respectively. In the plantation estate, the sampling effort yielded 46% (against ACE) and 65% (against Chao 1) of the "true" species richness. In monoculture smallholding, the sampling effort resulted in 71% (against ACE) and 84% (against Chao 1) of the "true" species richness. Similarly, the sampling effort in polyculture smallholdings produced 70% (against ACE) and 82% (against Chao 1) of the "true" species richness.

| Patterns of macrofungal cluster and morphospecies
To test our first hypothesis, we compared macrofungal cluster assemblages and morphospecies richness between logged peat swamp forest, large-scale plantation, monoculture, and polyculture smallholdings. Based on analysis of variance (ANOVA), we found that the number of macrofungal clusters was significantly greater (F 3,56 = 9.96;
Further analysis showed that polyculture smallholdings differed significantly in terms of relative humidity from monoculture smallholdings and large-scale plantations. No significant difference in relative humidity was detected between smallholdings and large-scale plantations.
Based post hoc Tukey test, all other comparisons of soil moisture were significant. We also found much more acidic soil in peat swamp forest (

| Macrofungal community
The macrofungal community differed significantly between the peat swamp forest and all other habitat types (ANOSIM, R global = 0.457; Number of permutations: 999; p < .001; Table 2). All three oil palm plantation habitats contained species that are not typical of peat swamp forest ( Figure 6).
One-way similarity percentages (SIMPER) was conducted to study the morphospecies richness contribution in all habitats. From the  F I G U R E 6 Nonmetric multidimensional scaling (NMDS) ordination comparing the macrofungal community between four different peatland habitats including those converted into oil palm agricultural areas analysis, a total of 13 morphospecies made up 90.04% (Table 3)

| DISCUSSION
This study compares the macrofungal biodiversity in four different habitats, namely logged peat swamp forest, large-scale plantation, monoculture, and polyculture smallholdings. Our study revealed that logged peat swamp forest supports higher abundance of macrofungal clusters and more diverse morphospecies. The current results

| Macrofungal morphospecies richness and abundance
Based on ANOVA analysis, we found that logged peat swamp forest contains higher macrofungal clusters and more diverse morphospecies smallholdings. This indicates that logged peat swamp forest supports more diverse macrofungal biodiversity, which might be due to environmental factors that determine the morphospecies and number of macrofungal cluster produced (López-Quintero et al., 2012).
However, between the cultivated habitats themselves, the macrofungal morphospecies and number of macrofungal cluster do not differ significantly. The patterns in the spatial distribution of fungal species hereby provide important clues about the underlying mechanisms that structure ecological communities and these are central for setting conservation priorities (López-Quintero et al., 2012;Mueller & Schmit, 2007). Brown et al. (2006) reported that forest patches had the highest sporocarp abundance and the greatest morphospecies richness per sample area. In contrast, coffee plantations had the lowest (Brown et al., 2006). Nevertheless, coffee plantation samples were more diverse for a given number of sporocarps than a sample of a similar size from forest patches (Brown et al., 2006). Morphospecies richness and the number of macrofungal clusters might be attributed to vegetation structure characteristics; canopy cover and canopy closure, substrate availability, and environmental factors. This is due to the changes in taxonomic and chemical composition of plant diversity which influence macrofungal productivity (Swift, Heal, & Anderson, 1979 Although we did not detect a significant difference in the macrofungal abundance and richness between oil palm management systems, we did find a higher richness and abundance in polyculture smallholdings compared to large-scale plantations and monoculture smallholdings. A possible explanation is that polyculture smallholdings exhibit a higher level of habitat heterogeneity created by crop diversity (Azhar et al., 2013(Azhar et al., , 2015. Habitat heterogeneity provides suitable microclimate which influences macrofungal productivity (Gómez-Hernández & Williams-Linera, 2011) especially in tropical countries (Lodge, 1997). Besides that, the production of macrofungal morphospecies abundance and richness in plantation areas might also be influenced by intensive use of pesticides and fungicides (Paz et al., 2015).
Canopy cover is important in shaping macrofungi diversity and productivity (Villeneuve et al., 1989). Closed canopy cover provides suitable environmental condition for macrofungal production because it reduces temperature and increases relative humidity for fungal growth (Belsky, 1994;Belsky et al., 1989). Due to humid microclimate, the soil moisture is higher in peat swamp forest than converted agriculture land. However, fungi are an adaptable species; they are able to withstand stress and adjust themselves to adapt to the most oligotrophic environments (Bergero, Girlanda, Varese, Intilli, & Luppi, 1999;Dighton, 2003;Wainwright, Al-Wajeeh, & Grayston, 1997

| Macrofungal morphospecies composition
Our results indicate that peat swamp forest and oil palm plantation supported different macrofungal communities. Macrofungal community in the logged peat swamp forest was diverse compared to oil palm production areas. Our findings are consistent with studies carried out in Western Ghats, India (Brown et al., 2006), south-eastern Australia (Kasel et al., 2008), and southern Brazil (Paz et al., 2015), which found that converted land does not support the original macrofungal composition of peat swamp forests.
There was no significant difference in macrofungal composition between monoculture and polyculture smallholdings. This implies that smallholdings show a similar macrofungal composition. Our SIMPER analysis shows that Schizophyllum commune was found in all three oil palm habitats and was the most abundant species in largescale oil palm plantation and monoculture smallholdings. A study in Southwestern Nigeria also found that S. commune is the most dominant macrofungi in rubber plantations (Osemwegie & Okhuoya, 2011). S. commune is not only an edible mushroom with medicinal value but is also reported to contain etiological agents (Saha et al., 2013).
Ganoderma boninense was the most common species in polyculture smallholdings. This could be due to the polyculture smallholdings management practices. Ganoderma are most commonly found on the remains of coconut trunks which were left on the site to rot. Retaining dead trees can potentially harbor diseases such as white rot fungus, G. boninense (Hushiarian et al., 2013). Antagonistic fungi, applying chemical treatments, and planting legume cover crops have been used to control G. boninense (Hushiarian et al., 2013).

| CONCLUSIONS
To our knowledge, our study has provided the first empirical evidence that modification in peat swamp forest to agricultural plantation leads to definite changes in the macrofungal biodiversity. Macrofungal biodiversity was reduced when the peat swamp forest was converted into oil palm plantations due to changes in environmental factors driven by vegetation structure modification and substrate availability. Available data also indicate logged peat swamp forests are essential to the persistence of macrofungal biodiversity in tropical human-modified landscapes. Further expansion of oil palm plantation on forest land should be prohibited in palm oil producing countries. The sustainable management of the peat swamps forest requires retaining diverse substrates such as dead wood of all sizes, species, and decay stages to maintain wood-inhabiting fungi diversity (Gates, Mohammed, Wardlaw, Ratkowsky, & Davidson, 2011). As oil palm expansion is inevitable in the tropics, palm oil stakeholders should be encouraged to use management practices that can enhance this biodiversity (Fischer et al., 2008;Foster et al., 2011).

ACKNOWLEDGMENTS
This work was supported by Universiti Putra Malaysia Grant GP-IPM/2017/9517300. The authors would like to thank Rasdianah Dahali for assisting us in the field. We are grateful to plantation managers for granting us permission to conduct this study in their oil palm plantations.