Changes in Aboveground Carbon Stocks
Tropical peatland forests provide numerous ecosystem services including carbon storage, and thus they are important for climate change mitigation (Murdiyarso et al. 2009; Novita et al. 2021). However, in this study, it was found that aboveground carbon storage of peatland forest particularly in standing trees was significantly altered by land use conversion. The carbon storage of standing trees was extremely low in grassland (0.20 ± 0.14 t ha− 1) and cropland (0.49 ± 0.42 t ha− 1) compared to the wetland forest (29.78 ± 4.79 t ha− 1). Such loss in carbon storage is attributed to the removal of vegetation mainly for the purpose of agricultural production (Anshari et al. 2010). In the LSBP, a large portion of the wetland forest has been deforested and converted to agriculture (primarily for rice production) but later abandoned due to poor yield leaving extensive grasslands (ASEAN Peatland Forests Projects, 2018) which stores far less carbon. In other tropical peatland forests especially in Indonesia and Malaysia, the removal and conversion of forests have been associated mainly from the establishment of oil palm plantations and logging (Miettinen and Liew 2010; Tonks et al. 2017; Jaafar et al. 2020) which subsequently resulted in the significant carbon loss. Hergoualc’h and Verchot (2011) were able to estimate the carbon loss in vegetation associated with the conversion of virgin peat swamp forest into other land uses in Southeast Asia. When converted to logged forest, mixed croplands and shrublands, rice field, and oil palm plantation, the carbon loss are estimated at 116.9 ± 39.8, 204.1 ± 28.6, 214.9 ± 28.4, and 188.1 ± 29.8 t ha− 1, respectively (Hergoualc’h and Verchot 2011).
In addition, the potential for carbon storage of the peatland forest under study should not be underestimated for its relatively higher average carbon stock of 29.78 ± 4.79 t ha− 1 with a range of 15.74 to 67.45 t ha− 1. The average standing tree carbon stock observed in this study was higher when compared to the average carbon stock of standing trees in the intermediate forest (14.42 t ha− 1) but lower compared to the tall pole forest (87.01 t ha− 1) of the Caimpugan peatland of Agusan Marsh in Mindanao Island, Philippines (Alibo and Lasco, 2012). When compared to the estimates in previous studies in other parts of the world, it was comparable to the average tree carbon stock of Thalawathugoda and Kolonnawa peatlands of Colombo Sri Lanka (30.38–47.99 t ha− 1, Dayathilake et al. 2020) and mangrove forest at Sofala Bay, Central Mozambique (28.02 ± 9.20 t ha− 1, Sitoe et al. 2014) but was lower than in riverine wetlands (peat swamp) of Encrucijada Biosphere Reserve, Mexico (95.1 ± 15.7 t ha− 1, Adame et al. 2015), Indonesian mangrove forest (64.40 t ha− 1, Murdiyarso et al. 2009) and Micronesian mangrove forest (104.4 ± 12.9–169.20 ± 28.2 t ha− 1, Kauffman et al. 2011).
Similar to the pattern of carbon stocks of trees, carbon stocks associated with understory and grass layer generally decreased with land use conversion which was highest in the wetland forest (5.98 ± 0.43 t ha− 1) and lowest in the cropland (2.54 ± 0.22 t ha− 1). The greater understorey and grass layer carbon stock in the wetland forest can be directly associated with the presence of a thick layer of sedges mainly composed of M. sumatrana and S. scrobiculata, and a vine species S. palustris. The observed average carbon stocks for understorey in wetland forest was higher compared to that of understorey carbon stocks of forest areas (0.66–2.33 t ha− 1) in Caimpugan peatland of Agusan Marsh, Mindanao (Alibo and Lasco, 2012). In the study of Dayathilake et al. (2020), it was pointed out that the contribution of understorey species to the overall aboveground and belowground carbon stock in wetland forest may be insignificant. However, the present investigation demonstrated that understorey and grass layers accounted for a relatively larger percentage of the total aboveground carbon even at forest sites (17.79 ± 2.39%), indicating the significance of its contribution to the overall aboveground carbon stock. Whereas, the low carbon stock in the cropland can be explained by the fact that these areas are of course dominated crop species (e.g. O. sativa) and other grasses which are either subjected to harvesting or cultivation. The cultivation process can slow down the growth or immediately remove vegetation and they store carbon in a relatively much shorter period of time than those in the wetland forest or grassland.
Litter materials such as leaves and woody debris are important components of aboveground carbon stocks where the decompositions of these materials serve as an important control on carbon accumulation in tropical peatlands (Hoyos-Santillan et al. 2015). However, land use conversion results in the disappearance of the litter layer and a decrease in the rate of litter fall (Hairiah et al. 2006). Such effects on litter are evident in this study where leaf litter and downed wood carbon stocks decreased significantly from wetland forest through cropland. The study of Upton et al. (2018) showed that litter inputs determined carbon storage in tropical peatlands where there were greater litter inputs and organic carbon accumulations in mixed forest and mangrove forest. Therefore, the reduction in leaf litter and downed wood carbon stocks in LSBP would directly mean a reduction on carbon accumulation in peat soils in much disturbed land uses (cropland and grassland) as compared to wetland forest. In addition, the reduced accumulation in litter biomass and carbon stock in the study area can be directly associated also with the accelerated decomposition due to exposure of moist litter materials to air brought about by the decline in the water table. Instead of accumulating, plant organic matters turn into gasses and dissolved organic acids and substances (Anshari et al. 2010).
Finally, the combined aboveground carbon stocks in the LSBP have declined significantly where the conversion of wetland forest to grassland and cropland had resulted in carbon loss of as much as 86.59 and 90.45%, respectively. The overall average aboveground carbon stock in wetland forest (38.56 ± 4.58 t ha− 1) in the present study was similar to that of the carbon stock in the intermediate forest of Caimpugan peatland, Philippines (31.16–43.40 t ha− 1, Alibo and Lasco, 2012) and mangrove forest at Sofala Bay, Central Mozambique (33.30 t ha− 1, Sitoe et al. 2014) although lower than in the peat swamp of Tanjung Puting National Park, Indonesia (~ 200.00 t ha− 1, Murdiyarso et al. 2009) and in intact peat forest of Central Kalimantan, Indonesia (73.48 t ha− 1, Petrova et al. 2008). The finding implies a greater need for the preservation of the remaining forested portion of the peatland as carbon sink.
Changes in Belowground Carbon Stocks
As expected, belowground root carbon stocks were significantly higher in wetland forests (5.05 ± 0.64 t ha-1) as compared to grassland and cropland. It can be speculated that carbon loss in roots is directly associated with the forest removal in the LSBP. However, caution must be taken into account in interpreting this finding of the study, that perhaps there is still the presence of undecomposed tree roots in the grassland and cropland areas forming the belowground carbon pool. The observed belowground root carbon stock of the wetland forest in the present study was comparable to that found in Thalawathugoda and Kolonnawa peatlands of Colombo Sri Lanka (4.87–7.44 t ha-1, Dayathilake et al. 2020) but lower than in the mangrove forest of Sofala Bay, Central Mozambique (25.22 ± 5.30 t ha-1, Sitoe et al. 2014). Root systems should be considered as they are an important part of the total forest biomass and eventually carbon storage (Verwer and van der Meer 2010) where belowground root carbon accounted for 11.49 ± 0.50% of the total biomass carbon for wetland forest in this study. Apart from that, roots have been found to be crucial in the peat formations in tropical peat swamp forest (Hoyos-Santillan et al. 2015, Verwer and van der Meer 2010).
Peat soil carbon stocks within 1 m depth increased along with increasing land use conversion intensity (Fig. 3b) despite the lower peat organic matter/carbon in the cropland areas of LSBP (Fig. 5e). This finding corroborated with the result of the study of Bader et al. (2018) where carbon stocks in managed peat soils (1 m depth) was significantly higher in cropland than those in grassland and forest. This pattern can be explained by the higher overall average peat bulk density in cropland (0.15 ± 0.01 g cm-3) which can be the result of mixing with mineral soils (Murdiyarso et al. 2009) and compaction due to the use of heavy machineries for cultivation. However, when the total peat depth is taken into consideration in the determination of carbon stock, it would most likely turn out that the wetland forest would have the greatest carbon storage than grassland or cropland as indicated in peat depth data (Fig. 4a). As part of the limitation for this study, as much as it was desired to sample the whole peat profile, it was only possible to extract a complete and undisturbed core samples up to 1 m depth with the kind of peat sampler used. Nevertheless, to have an overview on the potential peat carbon stock for all the land uses in LSBP, approximate carbon stock was computed for each location using the peat depth data and the overall average value for organic carbon and bulk density for each land use classification. The estimated carbon stocks were 2050.48 ± 317.15, 1655.74 ± 372.95, and 1100.57 ± 213.35 t ha-1 for wetland forest, grassland and cropland, respectively. These estimated values imply that LSBP especially in wetland forest stores significant amount of carbon in peat.
Changes in Peat Physico-chemical Properties
The water content of peat in the wetland forest was higher than the reported values in the tropical peatland forest of Southeast Asia (Anshari et al. 2010, Tonks et al. 2017). The peat moisture content decreased from forest to cropland along with the decreasing trend of water table height (Fig. 4b) which can be linked to compaction from farm machinery (Tonks et al. 2017). Peat bulk density in the forest areas (0.05 ± 0.002–0.07 ± 0.004 g cm-3) of this study were comparable to tropical peatland forests of SE Asia ranging from 0.07–0.15 g cm-3 (Anshari et al. 2010, Lampela et al. 2014, Könönen et al. 2015, Tonks et al. 2017) although lower than in other studies (0.46–0.69 g cm-3) (Murdiyarso et al. 2009, Aribal and Fernando 2018). Peat bulk densities in this study linearly increased with land use conversion that was highest in the cropland and which indicates peat degradation or decomposition (Krüger et al. 2015, Guillaume et al. 2016). High peat bulk density in the cropland can be a consequence of compaction linked to pressure applied on the peat by agricultural equipment, and shrinkage that occurs through the contraction of organic fibres when drying (Hooijer et al. 2012). Those above-mentioned processes might as well explain why croplands have very shallow peat. Similar to the finding of this study, high bulk densities were observed for converted peat in mature oil palm plantations (Anshari et al. 2010). In addition, aside from the possible effect of compaction from equipment, low porosities in cropland is an indication of peat decomposition which reduces the proportion of large pores by breaking down plant debris into smaller fragments (Rezanezhad et al. 2016), thereby reducing the water-holding capacity of peat soil.
The observed organic matter in the wetland forest (87.39 ± 0.74%) of the present investigation was comparable to North Selangor Peat Swamp Forest (NSPSF), Malaysia as reported by Tonks et al. (2017) with a value of 94.1 ± 1.5% but slightly higher than in Caimpugan peat swamp forest on Mindanao Island, Philippines (65.74–73.70%, Aribal and Fernando 2018). The reduction of organic matter particularly in the surface peat in the cropland area provides evidence on peat decomposition (Tonks et al. 2017). Enhanced peat decomposition follows after drainage, which is characterized by microbial respiration and peat oxidation, and as well the application of fertilizer, induce organic matter losses (Anshari et al. 2010, Leifeld et al. 2020). Such influence of drainage on organic matter loss is depicted by the observed direct positive relationships between water content and organic matter (Fig. 6a and b, Fig. 7) which was exactly observed also in the study of Tonks et al. (2017). Moreover, the loss of organic matter due to land use conversion was observed to be explained by increasing peat bulk density (Anshari et al. 2010, Hooijer et al. 2012) as found also in this study (Fig. 6c).