CO2 Emissions From Tropical Peat Soil Affected by Fertilization

The conversion of peat soils to agricultural uses has been thought to increase CO2 emission due to several factors, including fertilization. However, evidence on the effect of fertilization on CO2emissionsfrompeat soils is rareand often inconsistence. We measured the effects of different types of fertilizer, including N, P and K sources, and clay as an ameliorant on CO2 emission from a bare peat soil in Lubuk Ogong, Riau Province. Nutrients were added in the following combinations: 0 (unfertilized plot), N source (urea), slow-release N (slow release urea), N and Psource (Urea+SP-36), N, P and K sources (urea+SP-36+KCl) and combined NPK-Clay. Fertilization resulted in a decreasein CO2 emissions compared to that prior to fertilization except when slow-release urea was applied. Decreasing of CO2 emissions was probably due to pH-related effects because the pH in the N treatment was lower than in both the control and the unfertilized plot. A decreasein the level of CO2 emissions among the treatments followed the order NPK-Clay>NP>NPK>urea>slow-release urea. Covariance analyses showed that the difference in CO2 emissions prior to treatment was not significant. The application of individual and combined treatments of N, P, K and NPK mixed with 5 Mg ha-1 clay led to significantly reduced CO2 emissions from bare peat soil in Lubuk Ogong, Riau Province. In addition to fertilization, the water table depth was the only parameter that significantly affected the CO2 emissions (P<0.05). We conclude that the application of nutrient combinations, including N, P, K and clay, could reduce CO2 emissions because these treatments maintain a balanced nutritional condition in the soil with respect to the microbial activity.


INTRODUCTION
Tropical peatlands are considered to contribute greenhouse gas (GHG) emissions that accelerate global warming. High temperature and humidity levels in the tropics provide suitable conditions for various organic matter decomposition processes that cause a high loss of GHGs to the atmosphere. There are about 19.4 million hectares of peatland in Indonesia, and approximately 4 million hectares are located in Riau Province, which equals to 45% of the total area of Riau Province (Agus and Subiksa 2008). In 2000, land use and land use change and forestry (LULUCF) emissions in Indonesia were estimated to be 2.6 Mt CO 2 , corresponding to 34% of the global LULUCF emissions, and the majority of these emissions were caused by deforestation and forest degradation (PEACE 2007). Typically, the levels of CO 2 emissions from forested peatlands are lower than the levels of sequestration, and natural peatlands produce between 0.5 to 1.0 t C ha -1 yr -1 (Parish et al. 2007), whereas CO 2 sequestration from the atmosphere ranges from 0.008 to 0.80 t C ha -1 yr -1 (Harden et al. 1992).
The GHGs that are most commonly associated with agriculture are carbon dioxide (CO 2 ), methane (CH 4 ) and nitrous oxide (N 2 O). Among these GHGs, CO 2 is cycled and released into the atmosphere in the largest volume. Methane (CH 4 ) is mainly emitted due to rice cultivation and ruminant livestock, whereas N 2 O is released via the activity of soil microorganism (Snyder et al. 2007). Agriculture contributes less than 10% of the total emission of GHGs in the U.S. (US EPA 2007) and contributes less than 8% in Canada (Environment Canada 2007). Although total GHG emissions from agriculture are typically less than 10%, agriculture is considered to be the major source of CO 2 , CH 4 , and N 2 O.
The conversion of peatlands to agricultural lands has several consequences, such as subsidence of peat and the need to apply fertilizers and ameliorants. An appropriate management of drainage and fertilization on peat soils is expected to increase crop yield and maintain the soil condition. However, in case of peat soil, emission of GHGs should be considered. Concerns have been raised regarding an increase of CO 2 emissions on peat soils due to fertilization. Unfortunately, the effects of fertilizer application on peat soils on GHG emissions are not fully understood.
Nitrogen fertilizer, such as urea (CO(NH 2 ) 2 ), emits CO 2 when it is converted into ammonium (NH 4 + ), hydroxyl ion (OH -), and bicarbonate (HCO 3 -) in the presence of water and urease enzymes (Synder et al. 2007). A high efficiency of nitrogen use could increase C sequestration and reduce CO 2 emissions (Paustian et al. 1992). Nitrogen plays an important role in soil C storage and stabilizing C in the soil (Paustian et al. 1997;Lal 2004). The application of high levels of fertilizer significantly enhances carbon emissions mainly through the stimulation of root respiration process. Some reports on the effect of fertilization on soil CO 2 and N 2 O emissions showed a complex process involved (Sakata et al. 2015;Uchida et al. 2013;Nagano et al. 2012). Aerts and Caluwe (1999) suggested that CO 2 and CH 4 emissions are high in peat soils due to the low fertility of these soils. Therefore, mismanagement of the fertilizer rate, source, application placement and time can increase the overall N loss and N 2 O and CO 2 emissions. Several studies have reported that N 2 O and CO 2 emissions can be reduced using sufficient amounts of fertilizer and applying a high-efficiency fertilizer (Khan et al. 2007;Bufogle et al. 1998). However, several studies have shown contrasting results. The application of N fertilizer suppressed CO 2 emissions in several ecosystems, such as forests and grasslands (Kowalenko et al. 1978;Micks et al. 2004;Burton et al. 2004). These results could be explained as indirect effect of fertilization on CO 2 emissions. Nitrogen fertilization affects the soil C pool, resulting in an increase of biomass and microbial decomposition of crop residues (Green et al. 1995;Lal 2004). However, the reason for the suppression of soil CO 2 emissions due to N fertilization remains unclear.
In the present study, we investigated the effects of fertilizer application on the micro variability of CO 2 emissions using representative plots in a tropical bare peat soil. The study of the micro variability of CO 2 emissions may provide useful information for peat soil management.

Study Area
The study was conducted in an agricultural peatland in Lubuk Ogong Village, Pelalawan District, Riau Province, which is the second largest peatland in Riau Province (Wahyunto and Suryadiputra 2008). The study sites are located between latitude 0 o 20'44.5"N and longitude 101 o 41'10.9"E. The annual rainfall of this area is 2,500-3,000 mm and the daily temperature is between 22-31°C. The observation plots were located in bare land peat soils.
The observation area comprised of 24 plots in which the size of each plot was 2 m x 2 m. There were 6 fertilizer treatments and 4 replicates for each treatment, as shown in Table 1. The rate of fertilizer application was determined based on the fertilization recommendation for oil palm plantations. Fertilizers were spread evenly on the soil surface and then thoroughly mixed into the soil.

Measurement of Soil Properties and CO 2 Emissions
The level of maturity and determined thickness of peat and mineral substrate were determined in the field according to the method proposed by Agus et al. (2011). Soil samples were taken from 24 plots both in the top soil (0-20 cm) and sub soil (20-50 cm) before fertilizatation.
The soil samples were air-dried and ground to pass through a 2-mm sieve. Soil pH was measured using a glass electrode with soil:water ratio of 1:5 (McLean 1982). Total C content of the soil samples was determined using Walkley and Black method (Nelson and Sommers 1982). Total nitrogen (N) content was determined using Kjeldahl method (Bremner and Mulvaney 1982)). Available soil phosphorus (P) was extracted using Bray 2 method (Bray and Kurtz 1945). Cation exchange capacity (CEC) and exchangeable calcium (Ca), magnesium (Mg), potassium (K) and sodium (Na) were extracted using 1 M ammonium acetate (NH 4 OAc) at pH 7.0. Total P and K were extracted using 25% HCl. Calcium (Ca) and Mg concentrations were determined using an atomic absorption spectro photometer (Hitachi Z-5000), and K and Na concentrations were determined using a flame photometer. The microelements were extracted using DTPA, and their concentrations were measured by inductively coupled plasma (ICP) (Spectro Genesis Ametek). Samples for soil bulk density (BD) measurement were taken using a peat auger by drilling the 0-20 cm soil layer. Each sample in the tube was transferred quantitatively to a plastic bag and the BD was determined using the gravimetric method (Agus et al. 2011).
CO 2 emissions were monitored using anLi-COR 820 portable infrared gas analyzer (IRGA). The depth of water table, soil temperature and soil water content were measured at each measurement time point. Other parameters, including the soil pH, peat maturity, organic C content, bulk density and other soil chemical properties of peat soil were also observed.
The CO 2 gas emission in the surface of the peatland was trapped using a PVC closed chamber with a diameter of 25 cm and height of 23 cm. The gas fluxes were monitored every second for approximately 2 minutes. The relationship between the time and CO 2 concentration was exponential. The CO 2 flux is estimated using the following equation:

Peat Properties
The peat maturity in the surface soil (0-50 cm depth) is sapric, whereas hemic (50-100 cm) and fibric (100-150 cm) peats are present in the deeper layers ( Table 2). The properties of the peat soil in the study site are presented in Table 3. The soil pH in both the surface and subsurface is highly acidic with pH values less than 3.5. The soil carbon contents in the surface and subsurface of the peat soil range from 49 to 50 %. The CEC on a weight basis of peat soil was greater than that of typical mineral soils; however, if the values are considered on a volume basis, the CEC of peat soil would be significantly less than that of mineral soil. Organic soil commonly has 55% of the exchange capacity saturated with hydrogen. The contents of potassium, Mg and Na are low due to a high loss of these elements through leaching. The exchange capacity of organic soil materials depends on the number of negatively charged exchange sites. The sites adsorb cation Ca, Mg, K, and Na that replace the adsorbed hydrogen ions, as reported by Andriesse (1988). The total and available phosphorus contents are high, probably due to low levels of P sorption by Al and Fe. In general, the availability of micronutrient extracted with DTPA is low because trace elements, including Fe, Mn, Cu and Zn could be bound by organic acids.
The chemical properties of the soil before and after fertilization did not show a significant difference (data not shown), except for the soil C content. At the end of the experiment, the soil C content has decreased by approximately 3-10% in the fertilized plots. The loss of soil C in the top (0-20 cm) peat soil is higher compared to that in the subsoil (20-50 cm). The loss of soil C is likely related to the microbial activity. However, for the long term of observation (6 months) employed in this study, the effect of fertilizer to the decrease of soil C content could be ignored, and the microbial activity is likely controlling the soil carbon content through decomposition process.

The Effects of Fertilization on CO 2 Emissions
The CO 2 emissions from the peat soil before treatment ranged from 54 to 72 Mg ha -1 yr -1 ( Figure  1). The initial level of CO 2 emission in the peat soil observed in this study is higher compared to that in the peat soils from other sites in Indonesia (Husnain et al. 2014). The relatively high temperature of the soil and air as well as the relatively young age of the peatland may be responsible for these high levels  (Dariah et al. 2013;Marwanto et al. 2013) and 53 to 65 Mg ha -1 yr -1 in Central Kalimantan ). The cultivation systems and land uses have changed significantly CO 2 emissions as reported by Wilson and Al-Kaisi (2008). The levels of CO 2 emissions from the bare peat soil before and after fertilization are summarized in Table 4. After fertilization, the average level of CO 2 emission varied from 37.38 to 78.23 Mg ha -1 yr -1 . Covariance analysis showed that there was no significant difference in CO 2 emissions before fertilization among the treatments. However, after fertilization, the highest CO 2 emission was observed in the plots with nitrogen application. The application of N, P, K and the combination of these nutrients resulted in a decrease of CO 2 emissions. Among the treatments, the highest CO 2 flux was observed in plots with urea SL, although this flux was not significantly different from that in the Control and Urea plots. The lowest CO 2 flux   was observed in the NPK-Clay plot (Table 4). In summary, after fertilization, the order of the level of CO 2 emission was as follow: Urea SL> Urea> NPK> NP>NPK-Clay. In general, CO 2 emissions in the fertilized plots varied with the fertilizer type and measurement time. In contrast, in the control plot, the CO 2 emissions before and after fertilization were not significantly different. CO 2 emissions were various at each time point of measurement and affected by changes in the soil temperature and depth of water table (Table 5).The deeper of watertable exposes a thicker peat layer to aerobic conditions (Agus et al. 2011), which can contribute to the increase of CO 2 emissions. The levels of CO 2 emissions were relatively high and varied widely in the Urea and slow-release Urea plots. Urea is commonly used as nitrogen fertilizer source. Nitrogen plays a role in stabilizing C in soil by increasing the soil carbon level through biomass production. In the long term, nitrogen fertilization can lead to an increase of net C storage via humification processes (Paustian et al. 1992). However, the application of N fertilizer in the form of urea (CO(NH 2 ) 2 ) leads to a loss of CO 2 because urea is converted into NH 4 + , OHand HCO 3 -. All C in urea is considered to be emitted as CO 2 unless the proper management of urea is performed (Snyder et al.2007). Nitrogen fertilization also leads to emission of N 2 O as a result of the microbial processes of nitrification and denitrification. However, in this study, we did not observe N 2 O emission because the emission of N 2 O was expected to be small compared to the CO 2 emissions. According to IPCC (2006), approximately 1% of N fertilizer will be emitted as N 2 O-N. In adition, N 2 O emissions vary widely on a site-specific basis (Snyder et al. 2007).
A slow-release N fertilizer used in this study was a coated urea. In principle, slow-release nitrogen fertilizer would minimize GHG emissions due to the presence of a nitrification inhibitor that would minimize the conversion of NH 4 + into NO 2 -. As a result, N would be retained as NH 4 + for a longer period and plants would have enough time to assimilate NH 4 + and thereby prevent N 2 O emission. Uchida et al. (2013) reported that coated urea reduced CO 2 emissions. In contrast, our study observed that CO 2 emissions were high and fluctuated when slow-release urea was applied. Although nitrogen was released in bare peat soil, the nitrogen was not utilized by plants. An additional factor that caused high levels of CO 2 emissions was the fertilizer placement. Fertilizer was applied on the soil surface. The loss of N fertilizer is greater when N fertilizer is applied on the soil surface compared to band placement (Snyder et al. 2007;Hultgreen and Leduc 2003). These two factors may explain the high levels of CO 2 emissions from the plots applied with nitrogen fertilizer.
In contrast CO 2 emissions in the plots applied with NP, NPK and NPK+Clay were lower than in the plots applied with nitrogen fertilizer. The application of combined N, P and K fertilizer should improve the nutrient balance and cycling in the soil compared to the single application of nitrogen fertilizer. For example, when mixed N and P fertilizer was applied, the soil pH increased, whereas acidic conditions were observed when nitrogen fertilizer in the form of urea was added. In general, a balanced application of N, P, K and trace elements fertilizers is believed to be the best management practice for agricultural soils, including peat soils. Certain studies have reported a positive effect of applying P, K and trace elements together with N fertilizer. Shlegel et al. (1996) found that application of P fertilizer increased yields and reduced soil NO 3 -N levels, thereby leading to reduced GHG emissions. Application of P fertilizer inhibits the decomposition process of soil organic matter (Amador and Jones 1993) and soil respiration (Franklin et al. 2003) and does not affect CO 2 emissions (Aerts and Toet 1997). Johnson et al. (1996) suggested that application of K fertilizer improved crop N uptake and NO 3 -N retention in the upper soil profile. The condition above could have occurred as application of fertilizer can inhibit CH 4 oxidation by competing for methane monooxygenase, the enzyme used by methanotrophs to oxidize CH 4 . In addition, nutrient can also directly influence microbial communities at timescales reflecting the turnover of microbial populations (Keller et al. 2006). The lowest level of CO 2 emissions was observed with the application of NPK fertilizer combined with 5 Mg ha -1 clay. In this case, clay plays a significant role in reducing CO 2 emissions. This phenomenan may be due to the addition of clay is related to the complexation and binding of organic acids with metal ions contained in the clay. The incorporation of clay minerals within the peat soil could reduce the concentration of organic acids through chelation processes with metal ions. As a result, the peat soil would be more resistant to decomposition processes, therefore reducing the production of CO 2 and CH 4 (Sabiham 2010).
The fluctuations of CO 2 emissions from the fertilized plots are probably caused by chemical and biological reactions in the peat soil. These changes varied within the type of fertilizer and time of sampling. Harrison and Webb (2001) suggested that the emission factors of different fertilizer types depend on the conditions during the period after fertilization. However, Stehfest and Bouwman (2006) found that differences among fertilizer types largely vanish after fertilization due to the equal effects of the application rate, crop type, climate, soil organic carbon, soil pH and length of the experiment. Similarly, Bridgham et al. (1996; reported that different peatland types may respond differently to similar fertilization and liming applications. The variation in the CO 2 emissions is also affected by the depth of water table (P<0.05). The matrix correlation of several possible factors affecting CO 2 emissions are presented in Table 4. An increase of the depth of water table observed in the early rainy season (Nov-Dec 2011) corresponds to the decrease of CO 2 emissions.

CONCLUSIONS
The results of current study showed that application of combined N, P, K fertilizer and clay could reduce CO 2 emission. After fertilization, the relative levels of CO 2 emissions were as follows: Urea SL> Urea> NPK> NP>NPK Clay. Combined fertilization, including N, P, K and clay, could maintain a balance nutritional condition in the soils, including pH, the loss of ammonium to the air, decomposition processes, soil respiration, and other factors controlling peat soil reactions that are responsible for reducing CO 2 emissions. The average level of CO 2 emission varied from 37.38 to 78.23 Mg ha -1 yr -1 . However, large variations in the levels of CO 2 emissions were recorded during the experiment. Therefore, an intensive and long-term measurement of CO 2 emissions should be continued.