BIOTROPIA Vol. 28 No. 3,2021: 204 - 213 DOI: 10.1 1598/btb.2021.28.3.1330 UNDERSTANDING NATURAL REGENERATION IN BURNED TROPICAL PEATLAND: A STRATEGY TO ACCELERATE THE FOREST RECOVERY PROCESS DWI FUJI LESTARY, TARYONO DARUSMAN, FRANSISICJS A. HARSANTO, DESRA ARRIYADI AND GINANJAR Katingan-Mentqa Project) PT Rimba Makmur Utama) Jalan Kantin No. 8, Bogor 16121, Indonesia Received 11 February 2020/Accepted 17 June 2020 ABSTRACT The 2015 massive forest fires across Central Icalimantan left large areas of burned peatlands that need to be restored, demanding substantial resources. To understand natural regeneration of burned peatland and how planting might accelerate its recovery process, the regrowth of burned peatlands was measured with different fire frequencies. Three transects were established. Each transect consisted of five 20 x 20 m plots with 30 m intervals. All woody species were recorded and classified into three categories as new regrowth, regrowth, and remnant trees that survive from the last fire. In addition, data from fifteen 2 x 2 m permanent natural regeneration plots and evaluation on survival rate of 2017 planting were also analyzed. Research results suggested that the absence of remnant trees due to frequent or severe fires does not always impede the emergence of new recruitments, although diversity of forest regrowth is likely to be affected by its proximity to forest remnants. The floristic composition also showed a domination of pioneer species, giving evidence that forest recovery is initiated. Our study indicated that the combination of fire frequency, fire intensity, and proximity to remnant forest will produce different degrees of forest recovery, and the result will be unique for each site. It is concluded that to support the recovery process through planting activity, the successional stage of the designated sites should be determined first. The common planting method on large areas with mixed climax-high valuable trees was not beneficial unless the restoration sites had reached the later stages of succession. Keywords: natural regeneration, peatland £ire, restoration, succession, tropical peatland INTRODUCTION Indonesian peatlands account for 14.91 milhon ha and contribute to more than 35% of the world's peatlands (Osala e t al. 2016). However, although peatlands store a substantial amount of carbon compared to other land uses and provide important hydrological services to the surroundmg areas, their presence is threatened by human disturbance, especially due to the need to clear vegetation for agricultural lands (Page e t al. 2009). Osaki e t al. (2016) stated that the agricultural activity on peatlands in Indonesia has a long and complex historical substance, with fires playing an important role in *Corresponding author, email: dwi.puji@ptrmu.com this story, although fire is not likely naturlly occurred on peat swamp forest ecosystem. Fires have been commonly used to clear peatland forests, and h s such burning become massive in areas whenever a long drought occurs, such as in commonplace during El Nino climatic phases (Page e t al. 2009; Shiodera e t al. 2016). From 1990 to 2015, about 61% of Indonesia's peatland forests were lost or damaged, with only 6% of virgin peatland forest remaining by 201 5 (Graham e t aL. 201 7). The last massive forest fire in Indonesia was in 2015, whch burned 2.6 millions ha of lands, where 33% of them were peatlands (Glauber & Gunawan 2015). Noxious haze and tonnes of greenhouse gases (GHGs) were released, catching national and international attention as well as raising awareness. Understanding natural regeneration in burned tropical peatland: A strategy to accelerate - Lestari et aL This disaster spurred the needs for better efforts in peatland restoration and fire prevention. Unfortunately, planting on such a remote and wide landscape requires a substantial amount of capital, ranging from 500 to 3500 USD per hectare (Giesen & Sari 2018). On the other hand, relying on natural regeneration unlikely to be enough as it may result in low diversity (Blackham et al. 2014). Moreover, unlike numerous studies on dryland or tropical forests, studies on the recovery process of peatland after fire are less common or still at their early stage (Graham et a/. 2017; Page et al. 2009; Shiodera et al. 2016). This results in considerable uncertainties around the effectiveness of current peatland restoration practice. The hypothesis was that the result of forest recovery over times w d vary depenchg on the fire frequency, fire intensity, and proximity to remnant forest. To assess the vegetation recovery process and succession on recently burned peatland, their natural regeneration on sites which have dfferent fire history and proximity to the nearest remnant forest was measured. Species composition and diversity after fires were compared, and the effects of the current practice of tree planting on peatland restoration was investigated. Specifically, this study aimed to investigate whether the common practices on peat forest revegetation were parallel with natural process of forest succession. Moreover, ths study was part of an ongoing vegetation survey focuses on peat swamp forest succession after fire. This study only examined vegetation or floristic component of peatland restoration, whde restoration related to the hydrological function and other components of biodiversity are not covered. MATERIALS AND METHODS Study Location The study was conducted along the big canal on the southern area of the ecosystem restoration concession of PT Rimba Makmur Utama, also known as the Icatingan-Mentaya Project, Ieatingan District, Central Icalimantan located at 2O32'36.8" S to 3O01'43.6" S and 113°00'29.7" E to 113°18'57.4" E (Fig. 1). The study location is a typical degraded peatland forest mostly damaged by logging activities in the late 1970s to early 2000s, and subsequent canal drainage mainly for agricultures and transportation network, as well as forest fires. Peat depth on the study location ranged from 300 cm to 450 cm, with annual precipitation of about 2820 mm (information was collected from the weather station at Haji Assan Sampit Airport by Rossita e t al. (201 8). In the late 1990s, the Public Works Agency (Dinas Pekejaan Urnurn) constructed a 24 km long canal to connect Icotawaringin Timur and Icatingan District. Nowadays, about 6 km of the canal cuts across the restoration concession and has become the main cause of the surrounding peat drainage. Before the concession was granted in 2013, fires occurred almost annually along the banks of the canal. Vegetation along the canal is dominated by ferns and shrubs, with a few clumps of pioneer species. BIOTROPIA Vol. 28 No. 3,2021 STUDY LOCATION Legend n ~0n-m BDudanes - Canal Figure 1 Map of the study location on the ecosystem restoration concession Note: The map was displayed using Planet Lab satellite image year of 2017. Plot Observation and Data Analysis Nata~a l Saccession Plots diameter at breast height (dbh, if plant height is 1.3 m or more). Local names were identified on the field by a well-trained local botanist and their Three transects within the natural succession scientific names were identified using the guide observation site (Fig. 1) were developed in April book of PT Rimba Makmur Utama. During the - 2018 on the southeast part of the canal to observation, all transects were covered by observe natural regeneration of peatland forest shrubs and ferns with patchy pioneer tree after fires. The first transect (transect FB, species (Fig. 2). Both ferns and shrubs are frequently burned) was located on an area that typical species that emerge on peat ecosystem was close to the canal (0.5 krn distance). It was after fire such as StenochlaenapaIastm's and Cypems burned three times in the period of 2010-2015 rotgndzls with a height of more than 2 m. and was far from the forest edge (1.5 km Analysis was then conducted by dividing all distance). The second transect (transect FBF, woody plants into three classes: 1. new frequently burned, close to forest edge) was also regrowth, which included all woody plants with burned three times in between those years but was located far from the canal (1 km distance) and near to the forest edge on the west side of the transect. The last transect (NB, newly burned) has never been burned before 2014 and only caught fire once in 2015. This transect was also isolated from the nearest canal and forest edge. Before the attack of frequent fires, the land cover of the three transects was a secondary forest. Details regarding the fire history of the study location can be seen in Table 1. Five 20 x 20 m plots were established on each transect (Fig. 2). Each plot was located 30 m away from each other. In total, 15 observation plots (0.6 ha) were measured in dus study. Within each plot, all woody plants were counted and measured in terms of their bole diameter (if plant height is less than 1.3 m) or height I 150 cm that were assumed to emerge later after fire; 2. regrowth, which included all woody plants with height > 1.5 m and dbh < 10 cm that were assumed to emerge soon after the fire in 2015; 3. survivor, which included the remaining trees surviving from the last fire with dbh 2 10 cm. Species composition, density, species richness (Shannon's diversity index), and species evenness (Pielou's evenness index) were analyzed to examine the structure and composition of the existing natural regeneration and stage of the succession process. Bray-Curtis dissidarity index and NMDS ordination were also calculated to understand the pattern of species composition among transects and acquire the notable species within each study site. All analyses were performed using R version 3.4.0 with vegan package 2.4-3 (Oksanen etal. 2017). Understanding natural regeneration in burned tropical peatland: A strategy to accelerate - Lestari e t all Table 1 Fire history on the three transects from 2010 to 2018 Fire History (4 symbol means there was fire in the conspecific year) Years Transect FB Transect FBF Transect NB Notes: Fire and hotspot data were analyzed from Landsat 5, 7, 8 and Sentinel 2 satellite images. Each image on each year were displayed on composite mode using similar band combination of SWIR, NIR and Green, and hotspot historical data acquired from National Institute of Aeronautics and Space of Indonesia (LAPAN) website were overlaid. N a t ~ r a l Regeneration Plots the typical species that appear after fires on peatland forest. In total, fifteen 2 x 2 m permanent plots were established within the ecosystem restoration concession area of PT Runba Makrnur Utama, which was distributed on the southern part of the canal (Fig. 1, natural regeneration plots). Plots were located on degraded peatland that burned almost annually before 2015. The last fire incidence was in 2014. In 2015, these plots were established, and all seedlings less than 1.5 m in height were recorded every six months. The trend of natural regeneration composition from year to year was then analyzed to examine Tree Plantings on Degrdded Peatland As an ecosystem restoration concession, the concession is responsible for planting activity on their areas, and the result is monitored periodically. In January 2017, in total, 19,670 seedlings of local tree species were planted. About 2.5% of seedlings were monitored and the survivor rate was calculated in 9 months and 17 months after planting. BIOTROPIA Vol. 28 No. 3,2021 RESULTS AND DISCUSSION Species Composition on Different Sites Fires that attacked the study area resulted in the low density of remaining trees, as shown in Figure 3. Transect FB likely received a higher degree of fire incidence as only 5 trees/ha were left on this site, while transects FBF and NB had hgher density and more diverse remnant tress (Fig. 4). The study results suggested that the proximity to main canal influenced the intensity of fire, where fires normally started from the surroundmg canal. Transect FBF had the highest density and more surviving tree species, possibly due to the lower severity of fire on ths site, as well as its proximity to the forest edge. In addition, based on our observation, limited number of surviving trees indicated that those native climax-species were mostly not equipped with natural mechanism to survive under fires, as fire was not a natural phenomenon in tropical peat swamp environment unlike in dry sclerophyll forests where fires could occur naturally. The only protection was the wet and inundated peat environment that mostly absent when peat was drained. FB FBF N B -New regrowth .Regrowth Survivor Figure 3 The average density of woody species on different locations Notes: transect FB: frequently burned, close to canal, far from forest; transect FBF: frequently burned, far from canal, close to forest edge; transect NB: only burned once in 2015, far from forest and canal, divided by three size classes: new regrowth, regrowth, and survivor. Error bars indicated standard error. PHP New regrowth .Regrowth .Survivor Figure 4 Number of species on different locations Notes: transect FB: frequently burned, close to canal, far from forest; transect FBF: frequently burned, far from canal, close to forest edge; transect NB: only burned once in 2015, far from forest and canal, divided by three size classes: new regrowth, regrowth, and survivor. Understanding natural regeneration in burned tropical peatland: A strategy to accelerate - Lestari et al: It is likely that the availability of survivor recruitments after disturbance is often trees does not guarantee the emergence of new recruitments, although Cleary and Priadjati (2005) stated that the presence of remnant trees might be important to accelerate the succession process. Three years after the last fire incidence in 2015, recruitments of woody species were abundant on the three transects. Even though only a small number of remaining trees were present on transect FB, recruitments on this site were very dense (in total 800 recruitments/ha appeared after the 2015 fire) compared to transect FBF and transect NB. On the other hand, although higher recruitment density can be found on transect FB, the density itself is not parallel to the species diversity, which is relatively low at transect FB. The 800 recruitments/ha on &IS transect were composed of 21 species only, while transects FBF and NB contained 28 and 24 species of recruitments, respectively. This indicates that the density of the remnant determined by features of its local landscape. Bray-Curtis dissimilarity index calculation among the three transects also indicated that transect FB was the least similar compared to the other two transects, while transects FBF and NB share more similarities (Table 3). The NMDS ordnation displayed clearer segregation by showing that transect FB tends to segregate from the rest of the transects. This transect was characterized with more long-lived pioneer species, such as Melalezlca leacadendron, Melicope lanzl-ankenda, Jyggiam sp., and Macarangapminosa, whde transects FBF and NB were also rich with other generalist and late successional species, such as Alstonia scholaris, Ficzls spp., and Nepheliam mangd_yi. Unhke other researchers who reported that burned peatland forests in Icalimantan were normally dominated by pioneer species especially Combretocapzls and Cratoxylm species (Blackham et al. 2014; Graham et al. 2017; Shiodera e t al. 2016), there were no indications about the existence of these trees after fire might not affect the emergence of species in our study sites. However, it is new recruitments on peatland forest, as a source confirmed that our study sites were still at the of seeds might come from various sources. Seedbanks were likely absent due to frequent fires, but sources of seeds were possibly supplied from the nearest sites by their dispersal agents. Therefore, proximity to the remnant forest is expected to play an important role to increase diversity of these recruitments. This is supported by higher Shannon' diversity index Fable 2) on transect FBF compared to other transects. This is consistent with Chazdon (2008) who stated that the nature of forest early stage of forest succession as most species that were supposed to be present in undisturbed peat swamp forests as mentioned by Mirmanto (201 0) were absent. Moreover, Mrmanto (201 0) also reported that at least 2,000 trees/ha with more than 30 species could be found within 0.25 ha area of burned peatland. This indicated that the density of regrowth on our study site was still relatively low although the species richness (especially on FBF) demonstrated a valuable sign of recovery. Table 2 Shannon's diversity index (H') and Pielou's evenness (E3 on different locations Site New regrowth Regrowth Survivor H' FB FBF NB All sites E' FB FBF NB AU sites Notes: Transect FB: frequently burned, close to canal, far from forest; transect FBF: frequently burned, far from canal, close to forest edge; transect NB: only burned once in 2015, far from forest and canal, divided by three size classes: new regrowth, regrowth, and survivor. BIOTROPIA Vol. 28 No. 3,2021 Table 3 Bray Curtis dissimilarity index among the three transects Bray Curtis Dissimilarity Index FB FBF FBF 0.610 NB 0.607 0.467 Figure 5 Non-metric multidimensional scaling (NMDS) ordination with stress value < 0.2, showing that this ordination displays a fair representation of species composition on each plot. Plots 1-5 are plots on FB transect, while plots 6-10 and 11-15 are located on FBF and NB transects, respectively Moreover, our study indicated that fires frequencies are not the only determining factor on forest recruitment on this study site, although according to Shiodera e t al. (2016), intense and repeated fires reduced the ability of forests to regenerate. The combination of fire frequency, fire intensity, and proximity to remnant forest will produce different degrees of forest recovery, and the result will be unique for each site (Graham e t al. 2017). Our study showed that frequent fires might not impede new recruitments. However, proximity to the nearest forest edge might impact the diversity of regrowth. This is because proximity to forest remnant plays an important role in producing seeds to ensure the continuous emergence of recruitments. For example, despite receiving frequent fires in the last nine years, transect FBF recruits more diverse regrowth compared to transect NB which was burned just once in 2015. Another study on ex Mega Rice Project in Central I(alimantan also stated that natural regeneration on isolated degraded peatlands resulted in slow and patchy regrowth with low diversity (Blackham e t al. 201 4). pioneer species, which seeds were mostly dispersed by wind or birds, or sourced from dormant seedbank within the peat layer. Only few resproutings were found and mostly appeared from Ficzls spp. It is supported by Chazdon (2008) that initial succession is normally composed of long-lived pioneer species that change slowly over times. Moreover, Table 4 also displayed a phenomenon that most recruitments were not conspecific to the remaining trees that survive after fire. For example, the presence of a few Dipterocarp trees on transect NB was not followed by the emergence of seedlings from these species. Once again, our study showed that abundant mother trees would not give a substantial advantage on recruitments unless they are able to regenerate. The presence of climax species such as Shorea spp., would not likely support initial forest recovery as these species are not able to produce continuous seeds for regeneration due to h t e d pollination (Ghazoul 2005), and if they are able, seedlings of climax species might find it hard to survive due to extreme heat and sun radiation on a typical open peatland. Vegetation Recoveries Over Times and the forest recovery process starts with Impact on Peatland Restoration colonization (Chazdon 2008), the key to Table 4 shows that new recruitments after peatland vegetation recovery after fires is to fires on the three transects were dominated by enable vegetation colonization as soon as Understanding natural regeneration in burned tropical peatland: A strategy to accelerate - Lestari et al. Table 4 The four most dominant species within each transect Transect FB Transect FBF Transect NB New Regrowth Alstonia scholaris Melicope lung-ankenda Campnospemza coridceum Macaranga przlinosa Regrowth Melicope lunu-ankenda Melaleuca sp. Campnospem coriaceum Alstonia scholaris Survivor Campnosperma coriaceum Campnosperma coridceum Alstonia scholaris Ficzs spp. Ntphelium mangqi Elaeocarpus acmocarpus Macaranga prtlinosa Campnosperma coriaceum Tetractomia o bouata Tetractomia o bouata Alstonia scholaris Campnospema coriacem Elaeocarpus acmocarpus Alstonia scholaris Caqbnosperma coriaceum Melicope lunu-ankenda Jjpgium spp. 2 Alstonia scholaris Melicope lunu-ankenda Ficus spp. Sypgium spp 2 Myristica iners Shorea te_ysmanianna Shorea ulginosa Tetractomia obovata possible, and this depends on the availability of This finding is supported by our observation regeneration sources (seeds or resproutings). on 15 of the 2 x 2 m permanent plots of natural Unfortunately, heavily degraded peatlands are regeneration (Fig. 6). These plots were burned commonly dominated by high and dense ferns almost at an annual basis. with the last fire and shrubs that impede other woody species to incidence being in 2014. From this figure, grow (Page et 2009)' Given this several pioneer species (Meialeuca ieucadendmon, only pioneer species are able to grow and supply Meiicope iunu-ankenda, S_yggiam sp.) dominated the continuous seeds for further colonization (Hapsari et al. 201 8; Shiodera et al. 2016). Only whole study area. However, in the third year, late when ths condtion is achieved, late successional and generalist species successional species might then be able to (CaqfloQe~ma cokacer*ni, AAonia ~cboiahJ, emerge dispersed by birds or bats, and bring the Cte~o~@~~flpa~~ZfO~aJ) started to appear, although recovery process to the next stage. pioneer species still dominated. Melole~~ca Ieucadendron cope iunlt-ankeda -Campno sperma coriaceum --.c Ctenolophon parvifolius - Melaleuca leucadendmn -Melicope lunu-ankeda --F Syzygium sp. -Ampelocis sus sp. -eSyzyigrurn sp. cf campanulaturn ~ A l s f o n r a scholaris Figure 6 Trend of recruitments on the smaller natural regeneration permanent plots, observed from 2015 to 2018 Note: The last fire attack was in 2014, and before that year, all plots were almost annually burned. BIOTROPIA Vol. 28 No. 3,2021 The evidence that only pioneer species are able to secure the stand initiation process asks a question to the common technique of vegetation restoration on degraded peatland forest in Indonesia. Common practice normally involves line or blanket planting on a large area with a mix of pioneer and climax species regardless of their ability to produce continuous seed sources and resprouting ability for rapid colonization. Planting also normally prioritizes high economic value species or rare species which are beneficial only when they are purposed for enrichment planting after the first stage of the successional phase (stand initiation) is achieved. A paleocological study by Hapsari et al. (2018) stated that floristic composition in degraded peat-swamp forest in Sumatra can passively recover, and this is marked by initial domination of rapidly generating trees such as Gnetzam, Calopkyllum, Sapotaceae, and Ficus to assure tree colonization and finally enable other late successional species to naturally establish either dispersed by bats or birds. Our examination in 9 and 17 months after planting various mixed species, showed that pioneer species had a much higher survival rate and so dominated the revegetation (including species such as Alstonia) Combretocarpu~, and Syugium), and only a few late successional-high valuable tree species (such as Dyera, Dio.$yros) and Shorea) were found (Figure 7). Moreover, our findings showed higher overall survival rates compared to another trial planting experiment by Tata (2017), and slightly lower rates compared to a study by Lampela e t al. (2017). Our study suggested that planting late successional as well as high valuable tree species, without examining first which the successional stage the site is in, is unnecessary. Although some species with high commercial value are able to grow on the initial phase of forest recovery, however, as previously mentioned, it will not give any beneficial value if those species are not able to produce continuous regeneration for stand initiation process. However, late successional or climax species could sull be incorporated in the initial planting but with smaller number compared to the pioneer species. Again, these climax species are beneficial for enrichment planting only, where the planting purpose is to increase species diversity on a site that has passed the first phase of the successional process. Therefore, to increase the effectiveness of the forest recovery process on recently burned peatland, planting rapidly regenerating or pioneer species to ensure stand colonization is highly recommended. Figure 7 Survival rate of tree seedlings calculated in 9 months after planting Q?+9m) and 17 months after planting (P+ 17m) Understanding natural regeneration in burned tropical peatland: A strategy to accelerate - Lestari et al: CONCLUSION To ensure the forest recovery process, forest colonization with species that can produce continuous species accumulation, either by sprouting or producing seed sources, needs to be addressed first. The common method of restoration practice using expensive species that cannot guarantee continuous self-regeneration is unbeneficial unless the restoration sites have reached the later stage of succession. The successional stage of the designated area should be determined first, as planting should focus on species that meet the needs of the successional stage on the designated sites. To ensure forest recovery, vegetation restoration on peatland should be parallel with hydrological restoration and fire prevention. ACKNOWLEDGEMENTS We wish to acknowledge all field staff of PT Rimba Makmur Utama who helped the authors undertake routine data collection in the field, and local communities who enrich our insight on the peatland forest ecosystem with their important-original knowledge. REFERENCES Blackham GV, Webb EL, Corlett RT. 2014. Natural regeneration in a degraded tropical peatland, Central IWmantan, Indonesia: Implications for forest restoration. Forest Ecology and Management, 324, 8-15. https://doi.org/lO.1016/ j.foreco.2014.03.041 Chazdon RL. 2008. Chance and determinism in tropical forest succession. In: Carson KT, Schnitzer SA, editors. Tropical Forest Community Ecology. London: Willey-Blackwell. p. 384-408. Cleary DFR, Priadjati A. 2005. Vegetation responses to burning in a rain forest in Borneo. Plant Ecology, 177(2), 145-163. https://doi.org/l0.l007/s11258- 005-2107-0 Ghazoul J. 2005. Pollen and seed dispersal among dispersed plants. Biological Reviews, 80(3), 413- 443. https://doi.org/10.1017/S1464793105006731 Giesen W, Sari ENN. 2018. Tropical Peatland Restoration Report : the Indonesian case. https://doi.org/ 10.13140/RG.2.2.30049.40808 Glauber AJ, Gunawan I. 2015. The cost of fire. An economic analysis of Indonesia's 2015 fire crisis. In: The World Bank (Vol. 17). Graham LLB, Giesen W, Page SE. 2017. A common- sense approach to tropical peat swamp forest restoration in Southeast Asia. Restoration Ecology, 25(2), 312-321. https://doi.org/10.1111/ rec.12465 Hapsari ICA, Biagioni S, Jennerjahn TC, Reimer P, Saad A, Sabiham A, Behling H. 2018. Resilience of a peatland in Central Sumatra, Indonesia to past anthropogenic disturbance: Improving conservation and restoration designs using palaeoecology. Journal of Ecology, 106(6), 2473- 2490. https://doi.org/10.1111/1365-2745.13000 Lampela M, Jauhiainen J, Sarkkola S, Vasander H. 2017. Promising native tree species for reforestation of degraded tropical peatlands. Forest Ecology and Management 394:52-63. https://doi.org/ 10.101 6/j.foreco.2016.12.004 Mirmanto E. 2010. Vegetation analyses of Sebangau peat swamp forest, Central Icalimantan. Biodiversitas, Journal of Biological Diversity 11 (2):82-8. https://doi.org/l0.l3057/biodiv/dl10206 Oksanen AJ, Blanchet FG, Friendly M, IGndt R, Legendre P, Mcglinn D, . . . Szoecs E. 2017. Package 'vegan'. Osaki M, Nursyamsi D, Noor M, Wahyunto, Segah H. 2016. Peatland in Indonesia. In: Osaki M. Tsuji N, editors, Tropical Peatland Ecosystem, Tokyo UP):: Springer. p. 33-48. Page S, Hosciko A, Wosten H, Jauhiainen J, Silvius M, Rieley J, . . . Limin S. 2009. Restoration ecology of lowland tropical peatlands in Southeast Asia: Current knowledge and bture research directions. Ecosystems, 12(6):888-905. https://doi.org/ 10.1007/~10021-008-9216-2 Rossita A, Witono A, Darusman T, Lestari DP, Rlsdiyanto I. 2018. Water table depth fluctuations during E N S 0 phenomenon on different tropical peat swamp forest land covers in I