Biogeosciences Drainage and land use impacts on changes in selected peat properties and peat degradation in West Kalimantan Province , Indonesia

Degradation of tropical peats is a global concern due to large Carbon emission and loss of biodiversity. The degradation of tropical peats usually starts when the government drains and clears peat forests into open peats used for food crops, oil palm and industrial timber plantations. Major properties of tropical peat forests are high in Water


Introduction
Land use change and tropical peat degradation are commonly discussed from a perspective of Carbon emission (Page et al., 2004;Hooijer et al., 2010;Miettinen and Liew, 2010).Most studies in tropical peats were based on remotely sensed data (Jaenicke et al., 2008), and general models of Carbon emissions associated with land use change and peat degradation (Page et al., 2004).A study focusing on direct measurements of peat properties associated with drainage and land use changes is not common.Therefore, we conducted this study by investigating peat properties in peat forests, industrial timber, oil palm plantations, and community agriculture.We aim to discuss impacts of drainage and land uses on selected peat properties and peat degradation in West Kalimantan Province, Indonesia.We also suggest important indicators of peat properties that could be used to indicate tropical peat degradation due to drainage and land use change.We also believe that global climate change leading to more Published by Copernicus Publications on behalf of the European Geosciences Union.rainfall variability in Borneo in the future would enhance the degradation of tropical peats (Li et al., 2007).
Peat degradation is characterized by a change of physical, biological and chemical properties leading to functional deterioration and ecological decline that harms environment and socio-economy development.Therefore, peat degradation is definitely a complex process associated with land uses and social perspectives.Unwise land uses drive significant changes of physical, biological and chemical properties towards peat degradation, and an increase of Carbon emission from land use change.
The underlying causes of peat degradation are commonly land use conversion from peat swamp forests into agriculture and other uses.As peat growth depends on inputs of fresh vegetation biomass, the removals of peat forming vegetation directly reduce the deposit of vegetation biomass into peat profile.Without consistent input of vegetation biomass, peats stop to grow (Moore, 1989;Clymo, 1984Clymo, , 1991)).In addition to vegetation removals, the peat forest conversion is usually associated with the construction of drainage canals that lower water table depths for creating favourable medium for growths of crops.A decomposition rate in drained peat with low water table depth is more rapid than in water-saturated peat with high water table depth.The decline of water table depths increases the thickness of oxidative layer (acrotelm), which is rich in oxygen (aerobic), fresh litter and moist.The acrotelm is active and more favourable for decays than the permanent water-logged layer (Catotelm).

Regional setting
West Kalimantan Province (±146 000 km 2 ) has predominantly lowlands, flat topography, swamps, and rivers.This region is rich in biodiversity living in both terrestrial and aquatic ecosystems.The climate of West Kalimantan is ever-wet all the year (see Fig. 1).The annual rain is highly abundant ranging from 3000 to 5000 mm, and the evaporation is around 1500 mm per year.This wet tropical climate supports the development of water surplus, humid, and water-saturated environment.Lowlands, flat topography and ever wet tropical climate favour peat formations.According to USDA soil taxonomy, peat belongs to Histosol that largely contains organic soil materials at least 40 cm depth (Soil Survey Staff, 2010).
Indonesia has approximately 183 176 km 2 of tropical peats (Agus and Subiksa, 2008), or almost 5% of the world peats.In tropical regions, Indonesia has the largest tropical peat deposits, which are mainly distributed in Kalimantan (Borneo), Sumatra and Papua.A total estimate of 17 500 km 2 of peats occurs in West Kalimantan Province (see Table 1).
The distribution of peats in this province greatly associates with rivers, lakes, and inundated depression region.The Kapuas River (1145 km) runs from Pontianak through the Kapuas Hulu Ranges, covering about 100 000 km 2 catchments area or equal to 68% of the total area of the province.Other major rivers are the Pawan (197 km), Sambas (233 km) and Mempawah (<100 km).
There are three types of peat genesis.These consist of coastal, basin and high peats.Coastal peats occurs near sea level (1-2 m a.s.l.), and are usually shallow (<3 m), with close association with mangroves, brackish water, and tidal influences.Further, inland along the river valleys with the altitudes between 5 and 20 m, basin peat domes occur.Podzolization in the mineral substratum supports the development of water-logged environment.Tides influence the periphery of basin peat dome, and the centre of dome is commonly flat, and seasonally flooded by rain water.Small hills and rivers may cut basin peat dome.High peats occur in  (Ramsey, 2010).the depressions at altitude greater than 20 m a.s.l.Sometimes, high peats are marginally developed between hills or undulating terrains of the upper river basin.High peats may form small domes, with variability of depths.The formation of hard-pan in the mineral substratum maintain inundated environment.
Biomass deposit into peat profile suffers from decomposition, and it is only between 5 and 10% of the biomass are preserved into peat.Water-logged environment, poor nutrients and low pH inhibit the growth of decomposers.High lignin content in the peat forming vegetation species slows biological and chemical decomposition.
In coastal regions of the rivers, coastal peats commonly occur in a variety of dome shapes on former sea beds (Diemont and Supardi, 1987).In the upper Kapuas river basin, inland peat formation is very complex, occurring on former lake beds and several depression areas between small hills.It is common that distribution of inland peats in the upper Kapuas River is marginally scattered and these peats form a variety of small peat domes, which may range from 1 to 10 km 2 (Anshari, 2009).
The formation of peats is very complex processes of incomplete decays of organic matters under anaerobic environment.Peat accumulation occurs when the decay rate is lower than the preservation rate.Therefore, a consistent input of partially decayed organic matters should be maintained in order to favour peat accumulation.When the decay rate is rapid and high, peat does not grow but suffers from decomposition that turns the organic matters into gases and dissolved organic acids and substances.
Selected radio Carbon ages of tropical peats in Kalimantan and Peninsular Malaysia are so variable, spanning from 30 000 to hundreds of years (Anshari et al., 2001(Anshari et al., , 2004;;Page et al., 2004;Wüst and Bustin, 2004;Wüst et al., 2008).The radio Carbon analysis indicates that coastal peat in the Kapuas River basin is formed around 3000-4000 years BP (see Table 2).The basal dates show that the formation of peats in this coastal region of West Kalimantan probably occurred at the same time (Diemont and Supardi, 1987).
In contrast, the peat age from the upper Kapuas River basin is very much older than peats from the lower Kapuas River basin.Anshari et al. (2004) reported that the age of peat from Danau Sentarum National Park in the upper Kapuas River is greater than 30 000 yrs BP (see Table 3).Further, a basal radio Carbon date of peat from Sebangau National Park is 20 000 yrs BP (Page et al., 2004).In Tasek Bera Basin, peat dates range from 3400 to 20 000 yrs BP (Wüst and Bustin, 2004).More radio Carbon dates are required in order to detect the initiation of peats in this region.Also, it is important to note that peat formations are intermittently in several phases of accumulations and decays of organic matters (Anshari et al., 2004).The rates of peat accumulation in Kalimantan and Tasek Bera of Peninsular Malaysia are estimated to occur from 0.1 to 2.5 mm yr −1 (Wüst and Bustin, 2004;Page et al., 2004;Hope et al., 2005).

Sample collection and research site
The study was conducted in West Kalimantan Province, Indonesia (see Fig. 2).We collected peat core samples using a Russian type auger by Eijkelkamp, Holland.The peat sample campaigns were done between 2007 and 2010.We retrieved the core sample every 50 cm increment until reclaiming mineral substrate beneath peat.Every 50 cm peat core was transferred into half-cut PVC pipe ( = 2 inch), and carefully wrapped with home cling wrap.We took subsamples at a 50 cm interval up to 200 cm depths for laboratory analyses.growing species, community agriculture (CA), and oil palm plantations (see Fig. 2 and Table 4).
The inland peat forest is located in Danau Sentarum National Park in the upper Kapuas River basin, and is sufficiently intact and undrained.The coastal peat forest is located in Sungai Putri peat dome of Ketapang District, and it is also undrained.Both inland and coastal peat forests were selectively logged in the past.At present, no timber harvesting occurs in inland peats of Danau Sentarum National Park, and some illegal logging currently taking place in coastal peat forest of Sungai Putri peat dome.
Logged over peat forest is located in Antibar peat dome near Mempawah (about 70 km North of Pontianak).This coastal peat forest currently suffers from rapid rates of illegal logging activities, and slightly drained due to road construction.
The early industrial timber estate (EIT) is located in a small peat dome near Pontianak in the lower Kapuas River basin.The estate has just cleared the natural forest and replaced with a fast growing timber species (Acacia spp.) for one year.The timber plantation is slightly drained in order to lower water table depth.
The oil palm plantations consist of three age groups.One of the early oil palm (EOP2) is located in a scattered inland peat in the upper Kapuas River Basin.Other palm plantations are all located in coastal peat domes, and consist of early oil palm (<5 years), intermediate oil palm (5-10 years), and mature oil palm (15-20 years).All oil palm sites are heavily drained and regularly received fertilizers, particularly Nitrogen, Phosphorous, and Potassium.
The community agriculture (CA) is located in drained and open peat of Rasau Jaya peat dome.The forest of this peat dome was cleared and converted into agriculture by the government in 1972.Farmers usually plant maize and many kinds of vegetables in this peat.The use of chicken manures and limited amount of chemical fertilizers is common practice in this agriculture.

6
North of Pontianak).This coastal peat forest currently suffers from rapid rates of illegal logging activities, and slightly drained due to road construction.The early industrial timber estate (EIT) is located in a small peat dome near Pontianak in the lower Kapuas River basin.The estate has just cleared the natural forest and replaced with a fast growing timber species (Acacia spp) for one year.The timber plantation is slightly drained in order to lower water table depth.

Laboratory analysis
We studied selected physical and chemical property changes of peats due to different land uses and drainage disturbances.
Research variables are peat pH (H 2 O), Dry Bulk Density (DBD), Loss on Ignition (LOI), Gravimetric Water Content (GWC), Volumetric Water Content (VWC), Total Organic Carbon (TOC), Total Nitrogen (TN), and Atomic C/N Ratio (see Table 5).We took sub-samples for laboratory analysis at every 50 cm interval.The volumes of sub-samples for measurements are 50 cm 3 for pH, TOC, and TN, and 100 cm 3 for DBD, GWC, and LOI.We used Inolab pH meter (type 720 Bench to pH meter) to measure peat acidity.Fresh subsamples were diluted into distilled water at 1:5 for pH (H 2 O) measurements.GWC and DBD were determined after drying at 105 • C for 24 h.LOI values were calculated after combusting the oven dry samples at 550 • C for 5 h.LOI value represents the amount of organic matters.GWC, Ash Free DBD, and LOI were calculated using Eqs.( 1), ( 2), (3), and (4).

Water Content (g kg
Where: WS = wet sample weight, DOW 110 = constant weight after drying at 110 • C for 24-48 h, SV = sample volume, LOI = Loss on Ignition, DBD = Dry Bulk Density AW 550 = ash residue weight after combusting at 550 • C for 5 h.
The concentrations of TOC and TN were determined by dry combustion at high temperature with El Vario CHNS Analyzer.Samples were prepared by drying at 40 • C for at least 24 h until constant weights.About 20 mg of dry samples were oxidized and the evolved gases were measured.

Statistical analysis
We used SPSS version 17 for independent t-tests and Oneway ANOVA (SPSS, 2008).We compared average values of research variables according to contrasted locations (i.e.coastal and inland peats), and peat layers (i.e.acrotelm and catotelm).We assume that averages of acrotelm and catotelm depths are up to 100 cm, and greater than 100 cm, respectively.We realize that depths of acrotelm seasonally fluctuate following rainfall, and may be either less or greater than 100 cm.The determination of the acrotelm depth between 0-100 cm in this study is a conservative estimate.Takahashi (1999) reported that the water table depth of peat swamp forest in Central Kalimantan in a severe drought associated with El Niño of 1997 was 98 cm below the peat surface.
We used oneway ANOVA to analyze differences of peat properties according to the defined land use and drainage categories.As the numbers of samples collected from these groups are not equal, data distribution is not normal, and has unequal variances.To overcome this concern, we used Welch and Brown-Forsythe statistics in order to detect significant differences of the research variables with unequal variances (Field, 2005).If mean differences of research variables are statistically significant, we used Games-Howell (GH) procedure to detect specific differences in drainage and land use categories.

Independent t-tests
Table 6 presents means of research variables as classified into acrotelm-catotelm, and inland-coastal peats.We summarized results of independent t-tests for acrotelm (0-100 cm) versus catotelm (>100 cm), and inland versus coastal peats in Table 7a and b.
Table 6 shows that pH and DBD values are relatively low.Peat pH is slightly higher in catotelm than acrotelm, and pH in inland and coastal peats is very much similar to each other.DBD in acrotelm is slightly higher than catotelm, and DBD in inland peats is lower than in coastal peats.LOI values are generally lower in acrotelm than catotelm, and are higher in inland peats than in coastal peats.As expected, water contents in catotelm and inland peats are higher than acrotelm and coastal peats.The average TOC value is 52% in all categories.TN values are more concentrated in acrotelm and coastal peat.Accordingly, atomic C/N ratios in acrotelm and coastal peats are lower than catotelm and inland peats.
The t-tests for acrotelm-catotelm and inland-coastal peat show that LOI and TOC are indifferent.Significant differences between acrotelm and catotelm are found in pH, DBD, GWC, VWC, TN, and Atomic C/N Ratio.Peat pH, LOI and TOC between inland and coastal peats are statistically indifferent.Other variables (DBD, GWC, VWC, TN, and atomic C/N ratio) between coastal and inland peats are significantly different.

Drainage groups
Table 8 presents a summary of peat properties according to drainage categories.Values of peat pH, DBD, and TN increase following drainage disturbance.TOC values seem to decrease due to drainage influences.Therefore, atomic C/N ratios substantially decrease in drained peats.LOI seems unaffected by drainage.
Peat drainage influences pH, DBD, VWC and TN, showing some increasing trends towards drained peat.LOI and GWC do not indicate either an increasing or decreasing trend.TOC is slightly higher in undrained peat than in drained peat.It is obvious that atomic C/N ratio show a decreasing trend from undrained peat to drained peat (see Table 8).Table 11 presents Games-Howell (GH) multiple comparisons.Highly significant differences between undrained peat forest and drained peats are found in pH, DBD, VWC, TN and atomic C/N ratio.As shown, DBD in slightly drained peats (i.e.logged over peat forest (LF) and industrial timber plantation (EIT)) shows significant different from drained peats used for oil palm and agriculture.Significant different in TOC is shown between slightly drained peat and both drained and undrained peats.TN in drained peat is significantly different from undrained peat forest and slightly drained peat.Atomic C/N ratios show significantly different in all drainage groups.

Land use groups
We present means of research variables in land use groups in Table 12.Peat pH shows inconsistent variability in these groups.We believe that the variability of organic substances has a greater role in determining pH values.Without chemical fractionation of these organic matters, it is difficult to explain why peat pH has small variability in these land use groups.
DBD show an increasing trend from peat forests to disturbed and converted peats used for logging, oil palms, and industrial timber.A small sample number in CA influences low DBD values in this research.TOC values are relatively constant in the peat forests than the disturbed and converted peats used for logging, oil palms, and industrial timber.In general, TOC values range from 49 to 54%.The quality of organic matters must have a greater influence in the variability of TOC.
TN concentrations tend to be slightly higher in peats used for oil palms and agriculture due to fertilizer application.In general, TN values in peat forest range from 0.8 to 1.5%.In peats used for oil palms and agriculture, TN concentrations range from 2.4 to 3.4%.In consequence, atomic C/N ratios in peats used for oil palms and agriculture are lower than in peat forests.
Tables 13 and 14 present results of test homogeneity of variances, and robust tests of equality of means of research variables.We found that variances in land use classes are significantly unequal, and all research variables are significantly different.Results of GH multiple comparisons are presented in Table 15.
Based on GH multiple comparisons, DBD in inland peat forest are significantly different from DBD in selected peats used for agriculture, intermediate and mature oil palm plantations, the timber estate, and logged over peat forest.It is important to note that DBD in inland peat forest is not different from DBD in coastal peat forest.DBD in logged over peat forest is significantly different from DBD in EIT, PF1, EOP2, and CA.DBD in Coastal peat forest is notably different from most DBD in peats used for oil palms in coastal region.Further DBD in community agriculture is significantly different from DBD in PF2, LF, EOP1, IOP, and MOP.DBD in all oil palms of coastal peat is significantly different from DBD in early oil palm in inland peat.
LOI in intermediate oil palm plantations (IOP) significantly differs from PF2, EIT, LF, EOP1, MOP, and CA.TOC concentrations are not very much different in the land use categories.TOC in inland peat forest is only significantly different from TOC in EIT, which is also significantly different from TOC in IOP, CA and PF1.TOC in coastal peat forest (PF1) is significantly different from TOC in LF and MOP.TOC values in oil palms are all indifferent.
TN concentration in inland peat forest (PF2) is significantly different from other land uses in coastal peats and early oil palm in inland peat.TN in early industrial timber (EIT) in coastal peat is indifferent from TN in coastal peat forest, but is significantly different from other land use types.TN in coastal peat forest (PF1) is not significantly different from TN in logged over forest (LF), and is significantly different from other land uses.In peats used for oil palms and agriculture, TN concentrations in early oil palm of coastal peat (EOP1) and intermediate oil palm of coastal peat (IOP) are highly significantly different.In coastal area, peats used for oil palms show significant different between TN in IOP and MOP, and between TN in MOP and EOP2.TN in community agriculture (CA) is significantly different from TN is IOP and MOP.
As TN concentrations significantly increase in peats used for oil palms and agriculture, atomic C/N ratios between the peat forest and converted peats are mostly different.The C/N ratio in inland peat forest (PF2) is also different from the  C/N ratio in coastal peat forest (PF1), and the C/N ratios in PF1 and PF1 are significantly different from the C/N ratios in converted peats used for oil palms and agriculture.
Values of pH between coastal and inland peat forests are not significantly different according to GH multiple comparisons.Logged over forest shows significant different in pH from all peats used for oil palms, and early industrial timber (EIT).Means of pH in EIT is significantly different in all oil palms, except in EOP2.Further, pH value in IOP is significantly different from MOP and EOP1.It is important to note that pH in CA was not recorded in this analysis.
GWC values show significant differences between GWC values in EIT and EOP1, IOP and MOP.In oil palm land uses, GWC in MOP is significantly different from GWC in EOP1, EOP2, and IOP.GWC in PF1 was not recorded, and therefore was not compared.
VWC in coastal peat forest (PF1) is significantly different from VWC in peats used for MOP, IOP, and EOP1; and peats used for EIT and LF.In oil palms, there is significantly different between VWC in EOP 1 and EOP2; between IOP and MOP; between IOP and EOP2; between MOP and EOP2.
Figure 3 presents a scatter plot TOC versus TN according to drainage groups, and Fig. 4 presents scatter plots DBD versus TN according to land use groups.We recorded high TN concentration in drained peats used for oil palms and agriculture.
DBD in peat forests is generally lower than 0.1 g cm −3 .Rieley et al. (2008) recorded an average DBD value of 0.08 g cm −3 in peats from Central Kalimantan.Other authors commonly report the average DBD value of 0.1 g cm −3 for fibrist peats (Andriesse, 1988;Satrio et al., 2009;Soil Survey Staff, 2010).
Increases in TN in peats used for oil palms are mostly associated with application of fertilizers (see Figs. 3 and 4).All oil palms need regular and heavy inputs of fertilizers, particularly Nitrogen (N), Phosphorous (P) and Potassium (K).The direct impact of Nitrogen is to reduce C/N ratio as TOC val- ues are relatively constant in all land use groups, and are not greatly affected by drainage.We believe that the reduction of C/N ratio greatly influences a rate of peat decomposition.The availability of N would enhance decomposers to oxidize peats into dissolved organic Carbon, dissolved organic acids, and CO 2 gas.
Figure 5 depicts scatter plots of LOI versus C/N ratio in land use groups, and Fig. 6 presents scatter plots LOI versus TOC, and LOI versus TN.These figures show that LOI values in both peat forest and converted peats for oil palms and agricuture are consistently high, and values of C/N ratios in converted peats are substantially lower than in peat forest.Values of C/N ratio in IOP are comparably as high as in peat forest.This occurs because regular inundation in IOP causes rapid loss of Nitrogen, and retards the decomposition rate in this oil palm site.

Conclusions
We found that drainage and the conversion of peat forests into agriculture, oil palm and industrial timber plantation causes major changes of selected peat properties.In general, LOI and TOC values are constant in both peat forests and converted peats.Peat pH, DBD, and TN sufficiently tend to increase.Consequently, C/N ratio in the converted peats is significantly lower than in the peat forest.We believe that drainage largely influences an adequate increase of DBD and peat pH.Peat compaction is a direct consequence of drainage, and the removals of organic acids may substantially increase peat pH.An increase of TN concentrations in peats used for oil palms and agriculture is a direct impact of fertilizer uses, and subsequently leads to significant reduction of C/N ratio.We believe that increases in peat pH, DBD and TN concentration, and low C/N ratio (<20) in the drained and converted peats in this study are important indicators that indicate some processes of peat degradation.
Further studies are required in order to fully assess tropical peat degradation.We suggest to measure and develop models of decomposition rates under various land uses covering peat forests and drained peats used for agricultures, oil palms, and industrial timber plantations.Direct measurements of Carbon flux from these different land uses may improve our understanding on Carbon balance in tropical peats.Eddy co-variance and closed gas chamber techniques for measuring CO 2 emissions from these peats are urgently required.

Figure 2 .
Figure 2. West Kalimantan Province, showing study sites in selected coastal peat domes and two sites in inland peats

Fig. 2 .
Fig. 2. West Kalimantan Province, showing study sites in selected coastal peat domes and two sites in inland peats.

Fig. 3 .
Fig. 3.A scatter plot TN versus TOC according to Drainage Groups.

Table 1 .
Peat Distribution and Associated Peatland Draining Rivers in West Kalimantan Province (Source: Modified from BPS Kalimantan Barat, 2007).

Table 2 .
Diemont and Supardi (1987)r Ages of coastal peats from the lower Kapuas River basin.The last four age is adopted fromDiemont and Supardi (1987).1The coversion of 14 C age into Calendar year age was done by Oxcal 4.1 program

Table 3 .
Selected Radio Carbon Dates of Peats from Kalimantan and Peninsular Malaysia.

Table 4 .
Land Use and Drainage Categories of Study Sites.

Table 5 .
A list of research variables in this study.

Table 6 .
Means of Research Variables according to Acrotelm versus Catotelm, and Inland and Coastal Peats.

Table 7a .
Results of Independent t-tests for Acrotelm versus Catotelm (a) and Inland versus Coastal Peats (b).

Table 8 .
Means of Research Variables in Drainage Groups.

Table 9 .
Test of Homogeneity of Variances in Drainage Groups.Table 9 presents a test result of homogeneity of variances in drainage groups.All variables, except TOC, have unequal variances.Table 10 depicts results of robust tests of equality of means of the variables.The robust tests of equality of means show that LOI and GWC are indifferent and other variables are significantly different.

Table 10 .
Robust Tests of Equality of Means of Research Variables in Drainage Groups.
a Asymptotically F distributed

Table 12 .
Means of Research Variables according to Land Use Groups.

Table 13 .
Test of Homogeneity of Variances.

Table 14 .
Robust Tests of Equality of Means of DBD, LOI, TOC, TN, Atomic C/N Ratio, pH, GWC and VWC.

Table 15 .
Games-Howell Multiple Comparisons of DBD, LOI TOC, TN, and Atomic C/N Ratio according to Land Use Groups.