Global Change Biology (2016) 22, 432–448, doi: 10.1111/gcb.13099
Greenhouse gas emissions and global warming potential of traditional and diversified tropical rice rotation systems € G2, DAVID KRAUS1, S E B A S T I A N W E L L E R 1 , B A L D U R J A N Z 1 , L E N A J OR 3 H E A T H C L I F F S . U . R A C E L A , R E I N E R W A S S M A N N 3 , K L A U S B U T T E R B A C H - B A H L 1 , 4 and RALF KIESE1 1 Karlsruhe Institute of Technology, Institute of Meteorology and Climate Research (IMK-IFU), Kreuzeckbahnstr. 19 82467 Garmisch-Partenkirchen, Germany, 2Technical University Munich (TUM), Arcisstraße 21 80333 Munich, Germany, 3 ~os, Philippines, 4International Livestock Research Institute (ILRI) 30709 International Rice Research Institute (IRRI), Los Ban Naivasha Road, Nairobi, Kenya
Abstract Global rice agriculture will be increasingly challenged by water scarcity, while at the same time changes in demand (e.g. changes in diets or increasing demand for biofuels) will feed back on agricultural practices. These factors are changing traditional cropping patterns from double-rice cropping to the introduction of upland crops in the dry season. For a comprehensive assessment of greenhouse gas (GHG) balances, we measured methane (CH4)/nitrous oxide (N2O) emissions and agronomic parameters over 2.5 years in double-rice cropping (R-R) and paddy rice rotations diversified with either maize (R-M) or aerobic rice (R-A) in upland cultivation. Introduction of upland crops in the dry season reduced irrigation water use and CH4 emissions by 66–81% and 95–99%, respectively. Moreover, for practices including upland crops, CH4 emissions in the subsequent wet season with paddy rice were reduced by 54–60%. Although annual N2O emissions increased two- to threefold in the diversified systems, the strong reduction in CH4 led to a significantly lower (P < 0.05) annual GWP (CH4 + N2O) as compared to the traditional double-rice cropping system. Measurements of soil organic carbon (SOC) contents before and 3 years after the introduction of upland crop rotations indicated a SOC loss for the R-M system, while for the other systems SOC stocks were unaffected. This trend for R-M systems needs to be followed as it has significant consequences not only for the GWP balance but also with regard to soil fertility. Economic assessment showed a similar gross profit span for R-M and R-R, while gross profits for R-A were reduced as a consequence of lower productivity. Nevertheless, regarding a future increase in water scarcity, it can be expected that mixed lowland–upland systems will expand in SE Asia as water requirements were cut by more than half in both rotation systems with upland crops. Keywords: aerobic rice, maize, methane, nitrous oxide, paddy rice, rice rotation systems, economy, yield-scaled GWP Received 1 July 2015 and accepted 20 August 2015
Introduction The anthropogenic greenhouse gases (GHG) methane (CH4) and nitrous oxide (N2O) contribute about 17% and 6% to the overall global increase in radiative forcing, respectively (WMO, 2014). Paddy rice cultivation comprises significant CH4 emissions, accounting for about 10% of all anthropogenic CH4 emissions or about 1.5% of total global anthropogenic GHG emissions (Nazaries et al., 2013; FAOSTAT, 2014). N2O emissions from agricultural soils, representing approximately 5% of total global anthropogenic GHG emissions (WRI, 2014), are predominantly linked to inorganic and organic nitrogen fertilizer applications to arable upland systems (Davidson, 2009). Numerous studies report high CH4 but relatively low N2O emissions from flooded rice production (Linquist Correspondence: Ralf Kiese, tel. +49 (0)8821-183153, fax +49 (0) 8821-183294, e-mail:
[email protected]
432
et al., 2012) because anaerobic conditions limit nitrate availability and strict anaerobiosis favours complete denitrification to nitrogen gas (N2) (Zou et al., 2007). N2O emissions have their optimum in the range of 70– 80% water-filled pore space (WFPS) (Davidson et al., 2000; Zechmeister-Boltenstern et al., 2007). Therefore, in upland systems such as maize fields, N2O emissions are generally high after fertilization and/or rainfall/irrigation events as a result of tight coupling between nitrification and denitrification processes (Snyder et al., 2009). In South Asia, paddy rice cultivation is the most important cropping system (Devendra & Thomas, 2002). Nowadays, rice production is severely challenged since irrigation water is becoming increasingly scarce for agricultural production due to increased industrial and domestic demands and predicted lower availability as a consequence of climate change (Rijsberman, 2006). In regions with pronounced dry seasons, one option to adapt is to change traditional © 2015 John Wiley & Sons Ltd
G W P O F D I V E R S I F I E D R I C E R O T A T I O N S Y S T E M S 433 double-paddy rice systems into rotations with less water-intensive crops like aerobic rice or maize in the dry season. Aerobic rice cultivation is an emerging production system in South Asia where irrigation aims at reducing water inputs by keeping soil at field capacity rather than flooded or saturated (Belder et al., 2005; Bouman et al., 2005; Keating et al., 2010). Next to water scarcity, recent socio-economic developments with increased demands for livestock products and biofuels are becoming an additional impulse for rice farmers to grow upland crops in the dry season, for example maize as feed for poultry. Timsina et al. (2010) report the area in Asia grown with maize to be about 43.8 million ha, and although only a fraction of this maize is grown in rotation with rice (about 3.34 million ha), these systems are rapidly spreading. Shifting from paddy rice to maize or aerobic rice cultivation changes soil environmental conditions during the dry season and might result in ‘pollution swapping’, that is reduced emissions of CH4 at the expense of increases in N2O emissions (Stevens & Quinton, 2009). Moreover, soil organic C stocks might decrease due to the prolonged period with improved soil aeration and increased mineralization activity (Wassmann et al., 2004). To measure the proportions of such changes is of importance as the 100-year global warming potential (GWP) of N2O is about 12 times higher than of CH4 (Forster et al., 2007), indicating a risk for enhanced total GHG emissions after diversifying traditional rice cropping systems. Furthermore, changes from flooded to nonflooded cropping systems in the dry season can influence the emissions of CH4 from paddy rice soil in the following wet season cropping period. A study conducted under temperate climate conditions in Japan reported reductions in CH4 emissions in the range of 13–92% in fields converted from upland crop to paddy rice cultivation (Nishimura et al., 2011). However, detailed long-term measurements of combined CH4 and N2O emissions from paddy rice-upland crop rotations are still scarce, especially for ricemaize and aerobic rice systems in tropical and subtropical Asia. Therefore, the main objective of this study was to assess temporal changes of CH4 and N2O emissions as well as soil organic carbon (SOC) contents from diversified tropical rice systems. Observations of yield production and irrigation water usage allowed calculating yield-scaled GWPs and assessing economic benefits or risks of diversified rice cropping systems. To evaluate possible changes in SOC stocks in rotation systems with nonflooded crops, a comprehensive soil inventory of the field site was conducted directly before and after 3 years of crop diversification. GHG emissions were measured over a period of 2.5 years (Feb 2012–May © 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
2014) with two automated static chamber–GC systems set up at the International Rice Research Institute (IRRI), Los Ba~ nos, Philippines, for three different cropping systems across wet-dry seasons: 1 Traditional double-paddy rice rotation (R-R), 2 Paddy rice in rotation with aerobic rice (R-A) and 3 Paddy rice in rotation with maize (R-M).
Material and methods
Site description The field study was established at the Experimental Station of the International Rice Research Institute (IRRI), Los Ba~ nos, Philippines, about 66 km south of Manila. The study site (14°090 45″ N, 121°150 35″ E, 21 m a.s.l.) was cropped with paddy rice during wet season and dry season prior to the experimental set-up for over two decades. The long-term yearly (1979–2011) average sum of rainfall at this site is 2006 mm, with annual mean, maximum and minimum temperatures of 27.1, 30.7 and 23.5 °C, respectively (IRRI Climate Unit, 2011). Annual climate can be separated into dry season (DS) and wet season (WS) as rainfall is unevenly distributed over the year with a long-term average sum of rainfall of 300 mm in the DS (Jan–May) and 1706 mm in the WS (Jun– Dec). Soil is classified as ‘Andaqueptic Haplaquoll’ (USDA classification) with clay-dominated soil texture (54.2% clay, 32.8% silt, 13% sand) and a pH of 6.26.
Field experimental design Automated GHG exchange measurements were conducted at a field site cultivated with three-rice-based rotation systems. The field site was arranged in a split-block design with triplicate fields for each investigated rotation system. All systems were cropped with paddy rice in the wet season (WS) but differed in the dry season (DS), that is 1. paddy rice (R-R); 2. aerobic rice (R-A); and 3. maize (R-M). Each rotation system was further subdivided with regard to N fertilization: (i) zero N (ZN): no nitrogen at any time, (ii) site specific (SS): N fertilization determined by a crop manager tool (IRRI, 2014)/chlorophyll content, and (iii) conventional (CON): 130 kg N ha1 for all rotations and seasons, which lead to a total of 27 measuring plots (3 replicates 9 3 rotations 9 3 fertilizations) on 9 field sections. As preparation phase, all field plots were cropped with flooded rice under identical conventional N fertilization for one season before the beginning of the measurement campaign. For detailed information and plan of the field design, see Weller et al. (2015).
Agronomic practice The cropping period – from planting to harvest – was roughly January–May for the DS and July–October for the WS (Table 1). In the DS 2012 (February–June), paddy and aerobic rice was planted by direct seeding (60 kg seed ha1,
434 S . W E L L E R et al. be found in Weller et al. (2015). Due to severe damage of the experimental set-up caused by typhoon Rammasun at the beginning of the WS 2014 in July, annual measurements could not be completed for 2014. The GHG measuring system consisted of 27 chambers fixed to stainless steel frames, with 14 and 13 chambers connected to two gas chromatographs on site, respectively. Gas chromatographs equipped with a flame ionization detector (FID) and an electron capture detector (ECD) were operated on-site. Sample air was running through an ascarite column placed in front of the ECD to avoid bias caused by CO2 (Zheng et al., 2008). Chambers, made of transparent acryl glass, varied in size and modifications, depending on crop type and plant development stage (height). All chambers were equipped with fans and inlets for air mixture and pressure equalization during sampling. Crops were planted inside chambers and remained in place until harvest. Chamber soils were fertilized by hand according to field application rates and respective chamber areas. Chambers were closed in sets of pairs and sampled alternatively four times which resulted in a chamber closing time of 24 min, followed by two injections of a defined standard gas mixture for calibration purposes. Each chamber was sampled every four hours which allowed calculation of six fluxes per day. Complementary manual static chamber measurements, following the approach of Sander et al. (2014), were conducted at times when the operation of the automatic chamber systems was not possible. Partly during fallow periods and at field preparation, all plots were sampled manually at least weekly with two measurements per day (08:30–10:00 and 14:30– 16:00) and in higher frequency following field operations such as land soaking, ploughing, hydro-tilling and harrowing.
20 cm row spacing). Starting with the WS 2012, rice (paddy and aerobic) was always transplanted. Three-week-old seedlings were manually transplanted into puddled soil with 20 cm distance between hills. Planted rice varieties were NSIC Rc222 (paddy rice, high yielding) and NSIC Rc192 (aerobic rice, adapted to short-term drought) (Pinoyrkb, 2014). Paddy as well as aerobic rice received basal fertilization with solophos and muriate of potash at rates of 30 kg P2O5 ha1 and K2O ha1 between planting rows. Maize was hand-seeded using the Pioneer hybrid variety 30T80 with distances of 25 cm and 50 cm in and between rows, respectively. Maize received basal fertilization of 50 kg P2O5 ha1 and 30 kg K2O ha1. For all crops, nitrogen fertilization was applied in form of urea and was split into three applications. Fields were irrigated using a hydrant system equipped with water flow meters which supplied water from a nearby reservoir. Paddy rice was kept flooded with a water level of about 5 cm and was irrigated when flood water levels were lower than 2 cm. Aerobic rice was flushirrigated after 3–5 days without rainfall with the general aim maintaining soil water content around field capacity. Nevertheless, flush irrigation events can lead to short periods (1–3 days) of surface flooding. Maize was irrigated on demand by hose or flush irrigation after 5–8 days without rainfall.
Greenhouse gas measurements Automated GHG measurements were continuously conducted from February 2012 until June 2014, covering the cropping period of five consecutive seasons (dry–wet–dry– wet–dry). A detailed description of the measuring system can
Table 1 Length of cropping period (dates of seeding/transplanting until harvest and number of days) for the three rotation systems in the dry and wet season 2012–2014. Average irrigation (including land preparation) SD (n = 6) and total water input (including rainfall) are presented for each crop and season
Rotation system Dry season 2012
2013
2014
Wet season 2012
2013
Year
Cropping period (dd.mm)
Days on field (d)
Average irrigation (mm)*
Total water input (rain + irrigation) (mm)
R-A R-M R-R R-A R-M R-R R-A R-M R-R
10.02–28.05 24.02–13.06 01.02–24.05 03.01–02.04 08.01–29.04 03.01–13.04 07.01–04.04 08.01–05.05 07.01–13.04
109† 110 113† 89 111 100 87 117 96
378 266 1120 330 228 1444 645 273 1409
66 39 319 76 47 302 279 47 364
805 621 1524 647 520 1761 680 322 1444
R-A R-M R-R R-A R-M R-R
06.07–21.10 06.07–21.10 06.07–21.10 22.06–29.09 22.06–29.09 22.06–29.09
107 107 107 99 99 99
736 785 775 555 667 514
204 43 153 233 227 131
1778 1827 1817 1835 1947 1794
*Mean from all N-fertilized plots (CON + SS). †Planted by direct seeding. © 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
G W P O F D I V E R S I F I E D R I C E R O T A T I O N S Y S T E M S 435 Fluxes were calculated by linear regression procedures and were corrected by chamber temperature and atmospheric pressure. Average daily fluxes of any combined rotation and fertilization treatment were calculated by arithmetic means of sub daily fluxes using three replicated chambers. Cropping season emissions were computed by multiplying the arithmetic mean of all daily fluxes with the length of the cropping period (d). We further defined seasonal emissions (DS, WS) as the sum of cropping season (DS or WS crops) and corresponding fallow (after season) and land preparation (before season) emissions. Annual emissions are the sum of DS and WS emissions of a given year, including fallow and land preparation. N2O and CH4 emissions at fallow periods following DS could always be measured with the automated measuring system. Due to logistical constraints emissions in the fallow period, following wet season was only measured by manual chambers after WS 2014; thus, values were used also for WS 2012 and 2013. The same scaling approach was also applied for emissions during land preparation, which were measured by manual chambers in May/Jun 2014 and Dec/Jan 2015 (three replicates, for chamber details see Sander et al., 2014). For calculating the GHG balance, seasonal N2O and CH4 emissions were converted into CO2 equivalents (GWP) taking into account the specific radiative forcing potential of 298 for N2O and 25 for CH4 relative to CO2 for a 100-year time horizon (Forster et al., 2007). Further, we also included changes in soil organic carbon (SOC) stocks after 3 years as direct CO2 equivalents in the GWP. To relate yields of the different crop rotation systems to GHG emissions, annual yield-scaled GWP was computed (Van Groenigen et al., 2010; Linquist et al., 2012). For this, the seasonal GWP was divided by the annual grain production (rice and maize grain for R-M) from each rotation. Direct N2O emission factors (EFd %) of applied N fertilizer (kg N ha1) were calculated by using the following equation: EFd % ¼ 100ðEF E0Þ=N
ð1Þ
1
where EF (kg N ha ) is the cumulative N2O flux from the Nfertilized treatment, and E0 (kg N ha1) is the cumulative N2O flux from the unfertilized treatment (ZN).
Soil measurements An initial soil sampling of the field site was conducted in August 2011 during the field campaign preparation phase, in which all plots were cultivated with paddy rice and received equal amount of N fertilization. Second soil sampling was conducted in November 2014, shortly after WS harvest, during which all fields were also uniformly cultivated with paddy rice. In 2011, soil samples from two depths (0–10 and 10– 20 cm) were taken in two replicates per plot. Samples from each depth for the three respective plots per field were pooled. In 2014, soil samples from two depths (4–7 cm, 16–20 cm) were collected in 5 replicates per plot and pooled. Presented values for 2014 in this study derive only from fertilized plots (CON/SS), which received 260 and 270 kg N y1, respectively. Samples were sieved (d = 2 mm), grinded and analysed for soil total nitrogen (STN), and soil organic carbon (SOC) according to DIN ISO 13878 and 10694 by the analytical © 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
service laboratory of IRRI (2011) and a commercial laboratory (Dr. Janssen, Gillersheim, Germany; 2014). No bulk density (BD) values are available for 2011, while in 2014 bulk density (BD) in all sampled soil depths per plot was derived from six replicated soil cores (diameter 5 cm, 100 cm3). No differences in BD between systems were observed in 2014 (as all plots were equally puddled before transplantation) and BD from 2014 was used for both years to compute potential STN-/ SOC-stock changes for a fixed depth of 20 cm.
Auxiliary data Precipitation and air temperature was acquired from data sets provided by the IRRI Climate Unit from a weather station in close vicinity (10 mg CH4-C m2 h1) were reached about 4 weeks before harvest. As paddy field irrigation was usually stopped 4–6 weeks before harvest, CH4 fluxes decreased rapidly with drying of fields and decreasing soil water contents (Fig. 2c). However, also short lasting peak, CH4 emissions were occasionally observed at the beginning of these drying events due to release of CH4 entrapped in the soil matrix. As a consequence of low soil moisture contents, CH4 fluxes in the R-M system were negligible (0.1–0.18 mg CH4-C m2 h1) during maize cultivation in the DS (Fig. 3a, c). During paddy rice cultivation in the WS, CH4 emissions of the R-M system were substantial, but 60% lower when compared to R-R both in 2012 and 2013 (Fig. 3a). Compared with R-M, the R-A rotation showed considerable higher CH4 fluxes for aerobic rice cultivation in the DS, but fluxes were approximately 95% lower as compared to the permanently flooded R-R system (Fig. 4a). Seasonal development of CH4 fluxes in aerobic rice partly differed from wet rice cultivation. Only in the DS 2012, © 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
CH4 emissions steadily increased over the cropping season from approximately zero to 0.9 mg CH4 m2 h1 until irrigation was stopped to allow for crop ripening and soil water content declined. In contrast, in the DS 2014, CH4 emissions were characterized by short pulse emissions following irrigation events (Fig. 4c). In 2013, CH4 emissions were generally low and characterized by weak pulse emissions (80% (Fig. 6). General consent is that N2O emissions have their optimum in the range of 70–80% WFPS (Davidson et al., 2000). However, our results rather confirm the outcomes of a screening of 51 soils across Europe, which reported optimum of N2O emissions above 80% WFPS for most soils (Zechmeister-Boltenstern et al., 2007). Although we observed significant correlations of N2O and WFPS in our study, coefficients were rather low which may be due to the low spatial cover of soil moisture measurements. Therefore, for the ongoing experiments on our field site, more extensive soil moisture measurements are needed for better process understanding of coupled soil hydrological and N dynamics, which was also concluded by a model simulation study by Kraus et al. (2015), who used N2O emission and soil moisture data of this study for model validation. It also has to be noted that over the three observed years, N2O emissions from upland crops showed an increasing trend in general. As soil N contents in the diversified systems declined (Table 4), the background mineral N pool was most likely increased
by the mineralization of soil organic matter stocks, although we only found a strong trend of a decline in SOC stocks for R-M. Interestingly, examination of microbiological community also hinted at increased abundance of ammonia-oxidizing archaeons in the diversified systems (Breidenbach et al., 2015) which could further explain the increasing N2O emissions by enhanced nitrification processes. The comprehensive, long-term data set on N2O emissions of this study, to our knowledge a novelty for aerobic rice and tropical maize production, outlines the high importance but also the high inter-seasonal/annual variability of N2O emissions from tropical upland systems influenced by meteorological conditions and field-management practice.
Importance of GHG emissions during fallow periods and land preparation Our study is one of the few in which emissions during fallow and periods of land preparation are explicitly addressed for tropical rice-based systems. Regarding CH4, the prolonged fallow period following the WS paddy rice growing was of significant importance for the total CH4 balance as soils stayed waterlogged due to substantial rainfall events. As paddy rice fields were soaked for various days before ploughing, creating anoxic soil conditions, the contribution of land preparation phase to seasonal CH4 emissions was even higher in the WS for all systems. With contribution of fallow periods to annual CH4 emissions in the range of 8–11%, our results are in line with the few other existing studies which measured fallow emissions in rice cropping, reporting contributions of 9–30% (Bronson et al., 1997; Zhang et al., 2011). Regarding N2O emissions, cumulative emissions during fallow periods or periods of land preparation were mainly of importance for paddy rice. We assume that nitrate (NO 3 ) accumulation during dry soil periods (Davidson, 1992), that is the DS fallow period, and missing competition with plant N uptake, has © 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
G W P O F D I V E R S I F I E D R I C E R O T A T I O N S Y S T E M S 445 promoted N2O emissions when fields undergo first flooding for land soaking. In our study, the contribution of this event to seasonal N2O emissions varied largely (17–264%) but was in tendency higher as found in the study by Zhao et al. (2011), who reported that 80– 84% of total seasonal N2O emissions were released during preflooding of rice fields. Interestingly, for DS land preparation, the measured N2O emissions were insignificant in all systems in comparison, which might be connected to the higher N2O emissions during the preceding WS fallow and the depletion of available substrate (i.e. NO 3 ). Our results show that neglecting GHG emissions during periods of fallow and land preparation may lead to severe underestimations of seasonal and annual CH4/N2O budgets of paddy rice cultivation.
GWP and yield-scaled GWP In terms of GWP, land preparation/fallow periods were of much higher importance for WS cropping when compared to DS cropping (see Fig. 5a). The average GWP of paddy rice cultivation measured over all seasons (3868 2137 kg CO2-eq ha1 season1) was well in line with the average given by a meta-analysis (3757 kg CO2eq ha1 season1; Linquist et al., 2012). For the two complete years measured, the R-R control showed the significantly highest GWP values (Fig. 5a), while the R-M system showed the significantly lowest annual GWP of the three studied systems. Nevertheless, considering measurements of three consecutive seasons, our average GWP for tropical maize cultivation (1909 566 kg CO2eq ha1 season1) was higher than the average presented in the already mentioned meta-analysis (1399 kg CO2-eq ha1 season1; Linquist et al., 2012). The high emissions of N2O during the DS 2014 also contributed significantly to the GWP of aerobic rice cultivation (2029 1050 kg CO2-eq ha1 season1) in the R-A system. Ultimately, the main difference between traditional and diversified rice cropping system was the shift from N2O to CH4 as the predominant GHG emitted during the prolonged periods with aerated soil conditions. However, this ‘pollution swapping’ to the more potent GHG N2O was of lesser importance than the gain due to the reduction in CH4 emissions. Due to the continuous growth of global population resulting in an increasing demand of food production, more and more studies relate GHG emissions not only to area but rather to yields produced, for example on a hectare land. As we investigated double-cropping systems, we used the annual GWP and GY as computation base for the complete years measured. In line with the results for GWP and coupled to the high productivity of this system, our results identify the R-M system as a © 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
significantly lower emitter of GHG related to yieldscaled emissions when compared to R-R (Fig. 5c). Although the R-R system produced lower annual yields on average, it needs to be considered that annual yields of the R-M system include maize yields in the DS and are not directly comparable to yields of the R-R and RA systems with pure rice yields, as the used maize variety in this study is used for animal feed (yellow corn). The average seasonal YS GWP of maize and paddy rice cultivation in our study (180 48 and 616 383 kg CO2-eq. Mg GY1 for maize and rice, respectively) were again well in line with the values reported by the meta-analysis of 185 and 657 kg CO2-eq. Mg GY1 for maize and rice, respectively (Linquist et al., 2012). Considering aerobic rice cultivation, yields were reduced by 48 20% on average when compared to paddy rice in our study. Regardless of the lower productivity, the R-A system could still hold mitigation potential in terms of YS GWP in comparison with the R-R control, if one regards the observations in 2013. In this year, achieved DS yields were of the expected average and CH4 emissions during the WS were reduced significantly. Generally, the influencing values measured for R-A were highly variable over the five seasons (i.e. yields, WS CH4 and DS CH4/N2O emissions) and further studies seem necessary to assess the GWP of this cropping system when compared to double-cropping of paddy rice. It also has to be mentioned that other water-saving techniques might prove to be a better alternative to aerobic rice as, for example, alternative wetting and drying (AWD) recently has been reported to significantly reduce water use and GWP (CH4 + N2O) with lower yield penalties than observed in this study (Linquist et al., 2015; Nalley et al., 2015).
GWP of rice-based systems considering changes in soil C stocks It takes long-term observations over years to decades before SOC changes are detectable (Smith, 2004). Accordingly, after the short time of only 3 years, our study could not show a significant change of SOC contents between cropping systems, also due to the high variability of SOC and limited sampling numbers. However, our results suggest a declining trend of SOC for the R-M system, which is likely to get significant in a few years. Converted to SOC-stock changes (0–20 cm, fixed depth), the potential SOC losses for R-M over 3 years would account for 1.43 0.62 Mg C ha1. Nishimura et al. (2008) state that drainage of paddy fields for upland crop cultivation causes significant carbon loss from the soil and report a decrease in SOC contents in Japanese paddy fields converted to soybean-wheat (1.77 Mg C ha1) and upland rice
446 S . W E L L E R et al.
Fig. 7 Accumulated global warming potential over 3 years for the three rotation systems. Parts of stacked columns represent the CO2 equivalents deriving from total CH4 and N2O emissions and potential changes in SOC stocks (either loss or accumulation) over the 3 year period. Capped lines represent SE (n = 3). Note that the average GWP from the preceding two WS (2012/ 2013) was added as a representative value for WS 2014 as no measurements are available for this period of time.
(1.48 Mg C ha1) cultivation after 2 years. Including the potential C stock changes in a GWP assessment of CH4 and N2O emissions over 3 years (with representative values for the WS 2014) would increase the GWP of R-M by 5.2 Mg CO2-eq ha1 (Fig. 7). This would still result in a lower GWP for R-M as compared to R-R, even if the highly uncertain SOC increases for R-A and R-R are considered as CO2 uptake. However, SOC stocks in R-M will likely continue to decrease for several years until a new equilibrium is reached. Therefore, further studies are needed to assess the long-term impact of SOC losses on the total GWP of diversified cropping systems.
Economic assessment of diversified rice cropping systems As future shares of crop cultivation of R-R, R-A and RM systems will also depend on the income situation for farmers, we assessed the possible economic impacts of a conversion to upland crop rotation. Our assessment showed that the R-M system produced slightly higher annual gross profit (5398 658 USD ha1 y1) then the R-R system (5154 292 USD ha1 y1). Regarding the much lower water requirements (Table 1), this system might be of increasing importance in SE Asia, if proper marketing channels (e.g. mills, local wholesalers etc.) are available (Timsina et al., 2010). Although
annual gross profits were about 25% lower in the R-A system (3850 880 USD ha1 y1), the reduced water requirements (66%) are an advantage which may balance out losses when irrigation costs are the main investment in rice production or sufficient water supply is simply not available (Bouman et al., 2005). Moreover, using other rice varieties, higher yields of over 6 Mg ha1 season1 have been reported in aerobic rice cultivation (Belder et al., 2005) and breeding programmes might soon deliver other higher yielding cultivars for aerobic rice production. It has to be noted that we only assessed the gross income, without considering any costs for labour, machinery, etc. Nevertheless, in addition to water savings, labour and time-intensive puddling of fields and transplanting of rice is not needed in maize cultivation. Regarding future constraints in water availability, our results show that especially R-M systems might be highly attractive to rice farmers and highlight the importance of studies in such systems for the assessment of land-use change impacts in rice producing areas.
Acknowledgements This study was conducted as part of the multidisciplinary research project ICON. We thank the German Research Foundation (DFG) for its generous funding (FOR 1701, ‘Introducing Non-Flooded Crops in RiceDominated Landscapes: Impacts on Carbon, Nitrogen and Water Cycles [ICON]’, BU1173/13-1 and KI1413). We thank Philipp Kraft (University Giessen) for providing data on irrigation water.
References Belder P, Bouman BAM, Spiertz JHJ, Peng S, Castaneda AR, Visperas RM (2005) Crop performance, nitrogen and water use in flooded and aerobic rice. Plant and Soil, 273, 167–182. Borken W, Matzner E (2009) Reappraisal of drying and wetting effects on C and N mineralization and fluxes in soils. Global Change Biology, 15, 808–824. Bouman BAM, Peng S, Castaneda AR, Visperas RM (2005) Yield and water use of irrigated tropical aerobic rice systems. Agricultural Water Management, 74, 87– 105. Breidenbach B, Blaser MB, Klose M, Conrad R (2015) Crop rotation of flooded rice with upland maize impacts the resident and active methanogenic microbial community. Environmental Microbiology, doi: 10.1111/1462-2920.13041. Bronson KF, Singh U, Neue HU, Abao EB (1997) Automated chamber measurements of methane and nitrous oxide flux in a flooded rice soil.2. Fallow period emissions. Soil Science Society of America Journal, 61, 988–993. Buendia LV, Neue HU, Wassmann R et al. (1998) An efficient sampling strategy for estimating methane emission from rice field. Chemosphere, 36, 395–407. Buresh RJ, Haefele SM (2010) Changes in paddy soils under transition to water saving and diversified cropping systems. In: CD, 19th World Congress of Soil Science, Brisbane, Australia, 1–6 August 2010. Butterbach-Bahl K, Baggs EM, Dannenmann M, Kiese R, Zechmeister-Boltenstern S (2013) Nitrous oxide emissions from soils: how well do we understand the processes and their controls? Philosophical Transactions of the Royal Society B: Biological Sciences, 368, 20130122 Conrad R (2002) Control of microbial methane production in wetland rice fields. Nutrient Cycling in Agroecosystems, 64, 59–69.
© 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
G W P O F D I V E R S I F I E D R I C E R O T A T I O N S Y S T E M S 447 Corton TM, Bajita JB, Grospe FS et al. (2000) Methane Emission from Irrigated and Intensively Managed Rice Fields in Central Luzon (Philippines). Nutrient Cycling
Rijsberman FR (2006) Water scarcity: fact or fiction? Agricultural Water Management, 80, 5–22.
in Agroecosystems, 58, 37–53. Davidson EA (1992) Sources of nitric-oxide and nitrous-oxide following wetting of dry soil. Soil Science Society of America Journal, 56, 95–102. Davidson EA (2009) The contribution of manure and fertilizer nitrogen to atmospheric nitrous oxide since 1860. Nature Geoscience, 2, 659–662. Davidson EA, Keller M, Erickson HE, Verchot LV, Veldkamp E (2000) Testing a conceptual model of soil emissions of nitrous and nitric oxides. BioScience, 50, 667–
Roy KS, Neogi S, Nayak AK, Bhattacharyya P (2014) Effect of nitrogen fertilization on methane and carbon dioxide production potential in relation to labile carbon pools in tropical flooded rice soils in eastern India. Archives of Agronomy and Soil Science, 60, 1329–1344. Sander BO, Samson M, Buresh RJ (2014) Methane and nitrous oxide emissions from flooded rice fields as affected by water and straw management between rice crops. Geoderma, 235, 355–362.
680. Devendra C, Thomas D (2002) Smallholder farming systems in Asia. Agricultural Systems, 71, 17–25. FAOSTAT (2014) Emissions - Agriculture. Emissions of methane and nitrous oxide produced from agricultural activities. Average 1990–2013. Forster P, Ramaswamy V, Artaxo P et al. (2007) Changes in atmospheric constituents
Shang Q, Yang X, Gao C et al. (2011) Net annual global warming potential and greenhouse gas intensity in Chinese double rice-cropping systems: A 3-year field measurement in long-term fertilizer experiments. Global Change Biology, 17, 2196– 2210. Smith P (2004) How long before a change in soil organic carbon can be detected? Global Change Biology, 10, 1878–1883.
and in radiative forcing. In: Climate Change 2007: The Physical Science Basis : Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL), pp. 129–234. Cambridge University Press, Cambridge; New York. IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Prepared by the National Greenhouse Gas Inventories Programme.
Snyder CS, Bruulsema TW, Jensen TL, Fixen PE (2009) Review of greenhouse gas emissions from crop production systems and fertilizer management effects. Reactive Nitrogen in Agroecosystems: Integration with Greenhouse Gas Interactions, 133, 247–266. Stehfest E, Bouwman L (2006) N2O and NO emission from agricultural fields and soils under natural vegetation: summarizing available measurement data and modeling of global annual emissions. Nutrient Cycling in Agroecosystems, 74, 207–228.
IRRI (2014) Crop Manager. Available at: http://cropmanager.irri.org/ (accessed 15 December 2014). Itoh H, Ishii S, Shiratori Y, Oshima K, Otsuka S, Hattori M, Senoo K (2013) Seasonal transition of active bacterial and Archaeal communities in relation to water management in paddy soils. Microbes and Environments, 28, 370–380. Jackel U, Schnell S, Conrad R (2001) Effect of moisture, texture and aggregate size of
Stevens CJ, Quinton JN (2009) Policy implications of pollution swapping. Physics and Chemistry of the Earth, 34, 589–594. Timsina J, Jat M, Majumdar K (2010) Rice-maize systems of South Asia: current status, future prospects and research priorities for nutrient management. Plant and Soil, 335, 65–82. Van Bodegom PM, Leffelaar PA, Stams AJM, Wassmann R (2000) Modeling methane
paddy soil on production and consumption of CH4. Soil Biology & Biochemistry, 33, 965–971. Keating BA, Carberry PS, Bindraban PS, Asseng S, Meinke H, Dixon J (2010) Eco-efficient agriculture: concepts, challenges, and opportunities. Crop Science, 50, S109– S119. Kraus D, Weller S, Klatt S, Haas E, Wassmann R, Kiese R, Butterbach-Bahl K (2015) A new LandscapeDNDC biogeochemical module to predict CH4 and
emissions from rice fields: variability, uncertainty, and sensitivity analysis of processes involved. Nutrient Cycling in Agroecosystems, 58, 231–248. Van Groenigen JW, Velthof GL, Oenema O, Van Groenigen KJ, Van Kessel C (2010) Towards an agronomic assessment of N2O emissions: a case study for arable crops. European Journal of Soil Science, 61, 903–913. Wassmann R, Neue HU, Ladha JK, Aulakh MS (2004) Mitigating greenhouse gas emissions from rice-wheat cropping systems in Asia. In: Tropical Agriculture in
N2O emissions from lowland rice and upland cropping systems. Plant and Soil, 386, 125–149. Linquist B, van Groenigen KJ, Adviento-Borbe MA, Pittelkow C, van Kessel C (2012) An agronomic assessment of greenhouse gas emissions from major cereal crops. Global Change Biology, 18, 194–209. Linquist B, Anders M, Adviento-Borbe M, Chaney R, Nalley L, da Rosa E, van Kessel
Transition — Opportunities for Mitigating Greenhouse Gas Emissions? (eds Wassmann R, Vlek PG), pp. 65–90. Springer, Netherlands. Weller S, Kraus D, Ayag K, Wassmann R, Alberto MCR, Butterbach-Bahl K, Kiese R (2015) Methane and nitrous oxide emissions from rice and maize production in diversified rice cropping systems. Nutrient Cycling in Agroecosystems, 101, 37–53. WMO (2014) World Meteorological Organization Greenhouse Gas Bulletin: the state
C (2015) Reducing greenhouse gas emissions, water use, and grain arsenic levels in rice systems. Global Change Biology, 21, 407–417. Liu D, Ishikawa H, Nishida M, Tsuchiya K, Takahashi T, Kimura M, Asakawa S (2015) Effect of paddy-upland rotation on methanogenic archaeal community structure in paddy field soil. Microbial Ecology, 69, 160–168. Miller AE, Schimel JP, Meixner T, Sickman JO, Melack JM (2005) Episodic rewetting enhances carbon and nitrogen release from chaparral soils. Soil Biology and Bio-
of greenhouse gases in the atmosphere based on observations through 2013. Available at: http://www.wmo.int/pages/prog/arep/gaw/ghg/GHGbulletin.html (accessed 8 September 2014). WRI (2014) World Resources Institute World greenhouse gas emissions in 2005. Available at: http://www.wri.org/publication/world-greenhouse-gas-emissions2005 (accessed 13 Oktober 2014). Zechmeister-Boltenstern S, Schaufler G, Kitzler B (2007) NO, NO2, N2O, CO2 and CH4
chemistry, 37, 2195–2204. Nalley L, Linquist B, Kovacs K, Anders M (2015) The economic viability of alternative wetting and drying irrigation in Arkansas rice production. Agronomy Journal, 107, 579–587. Nazaries L, Murrell JC, Millard P, Baggs L, Singh BK (2013) Methane, microbes and models: fundamental understanding of the soil methane cycle for future predic-
fluxes from soils under different land use: temperature sensitivity and effects of soil moisture. Geophysical Research Abstracts, 9, 07968. Zhang GB, Ji Y, Ma J, Xu H, Cai ZC (2011) Case study on effects of water management and rice straw incorporation in rice fields on production, oxidation, and emission of methane during fallow and following rice seasons. Soil Research, 49, 238– 246.
tions. Environmental Microbiology, 15, 2395–2417. Nishimura S, Yonemura S, Sawamoto T, Shirato Y, Akiyama H, Sudo S, Yagi K (2008) Effect of land use change from paddy rice cultivation to upland crop cultivation on soil carbon budget of a cropland in Japan. Agriculture Ecosystems & Environment, 125, 9–20. Nishimura S, Akiyama H, Sudo S, Fumoto T, Cheng W, Yagi K (2011) Combined emission of CH4 and N2O from a paddy field was reduced by preceding upland
Zhao X, Min J, Wang S, Shi W, Xing G (2011) Further understanding of nitrous oxide emission from paddy fields under rice/wheat rotation in south China. Journal of Geophysical Research-Biogeosciences, 116, G02016. Zheng X, Mei B, Wang Y et al. (2008) Quantification of N2O fluxes from soil–plant systems may be biased by the applied gas chromatograph methodology. Plant and Soil, 311, 211–234. Zou JW, Huang Y, Jiang JY, Zheng XH, Sass RL (2005) A 3-year field measurement of
crop cultivation. Soil Science and Plant Nutrition, 57, 167–178. Pinoyrkb (2014) Pinoy Rice Knowledge Bank: FAQs on Philippine Seedboard (PSB)/ NSIC Rice Varieties. Available at: http://pinoyrkb.com/main/resources/rice-varieties (accessed 19 October 2014). PSA (2015) Philippine Statistics Authority - Bureau of Agricultural Statistics Homepage. Available at: http://www.bas.gov.ph (accessed 20 January 2015).
methane and nitrous oxide emissions from rice paddies in China: effects of water regime, crop residue, and fertilizer application. Global Biogeochemical Cycles, 19, GB2021. Zou J, Huang Y, Zheng X, Wang Y (2007) Quantifying direct N2O emissions in paddy fields during rice growing season in mainland China: dependence on water regime. Atmospheric Environment, 41, 8030–8042.
© 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448
448 S . W E L L E R et al.
Supporting Information Additional Supporting Information may be found in the online version of this article: Figure S1. Average daily CH4 (a) and N2O (b) flux rates of maize cultivation during the dry seasons of 2012, 2013 and 2014 in the RM system for each of the three fertilizer treatments. Figure S2. Average daily CH4 (a) and N2O (b) flux rates of aerobic rice cultivation during the dry seasons of 2012, 2013 and 2014 in the R-M system for each of the three fertilizer treatments. Figure S3. Average daily CH4 (a) and N2O (b) flux rates of paddy rice cultivation during the dry seasons of 2012, 2013 and 2014 in the R-M system for each of the three fertilizer treatments. Figure S4. Average daily CH4 (a) and N2O (b) flux rates of paddy rice cultivation during the wet seasons of 2012 and 2013 in the RM system for each of the three fertilizer treatments. Figure S5. Average daily CH4 (a) and N2O (b) flux rates of paddy rice cultivation during the wet seasons of 2012 and 2013 in the RA system for each of the three fertilizer treatments. Figure S6. Average daily CH4 (a) and N2O (b) flux rates of paddy rice cultivation during the wet seasons of 2012 and 2013 in the RR system for each of the three fertilizer treatments. Table S1. CH4 emissions (without land preparation and fallow period emissions) from each fertilizer treatment and rotation system (dry/wet season 2012–2014). Table S2. N2O emissions (without land preparation and fallow period emissions) from each fertilizer treatment and rotation system (dry/wet season 2012–2014). Table S3. Grain yields from each fertilizer treatment and rotation system (dry/wet season 2012–2014).
© 2015 John Wiley & Sons Ltd, Global Change Biology, 22, 432–448