Joint Life Cycle Assessment and Data Envelopment Analysis for the benchmarking of environmental impacts in rice paddy production

Journal of Cleaner Production 106 (2015) 521e532

Contents lists available at ScienceDirect

Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro

Joint Life Cycle Assessment and Data Envelopment Analysis for the benchmarking of environmental impacts in rice paddy production Ali Mohammadi a, b, *, Shahin Rafiee a, Ali Jafari a, Alireza Keyhani a, Tommy Dalgaard b, Marie Trydeman Knudsen b, Thu Lan T. Nguyen b, Robert Borek c, John E. Hermansen b a

Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran Aarhus University, Department of Agroecology, Blichers Allé 20, P.O. Box 50, DK-8830 Tjele, Denmark Department of Agrometeorology and Applied Informatics, Institute of Soil Science and Plant Cultivation e State Research Institute (IUNG-PIB), Czartoryskich 8 Str., 24-100 Puławy, Poland b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 November 2013 Received in revised form 15 April 2014 Accepted 3 May 2014 Available online 14 May 2014

The combined implementation of Life Cycle Assessment (LCA) and Data Envelopment Analysis (DEA) has been identified as a suitable tool for the evaluation of the environmental and economic performance of multiple similar entities. In this study, a total of 82 rice paddy fields for spring and summer growing seasons in north of Iran were assessed using a combined LCA and DEA methodology to estimate the technical efficiency of each farmer. Furthermore, the environmental consequences of operational inefficiencies were quantified and target performance values benchmarked for inefficient units so that ecoefficiency criteria were verified. Results showed average reduction levels of up to 20% and 25% per material input for spring and summer systems, leading to impact reductions which ranged from 8% to 11% for spring farms and 19% to 25% for summer farms depending on the chosen impact category. Additionally, the potential economic savings from efficient farming operations were also determined. The economic results indicate that an added annual gross margin of 0.045 $ per 1 kg rice paddy could be achieved if inefficient units converted to an efficient operation. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Life Cycle Assessment (LCA) Data Envelopment Analysis (DEA) Rice paddy Growing season Technical efficiency Economic savings

1. Introduction Rice is known as a staple human food source in many areas of Iran, where the average per capita consumption of rice is 100 g/day which ranks Iranian people as the 13th biggest rice consumers (Hormozi et al., 2012). Iran has to import 1.7 Mt rice per year to supply the domestic market which makes Iran the second largest rice importing country after Philippine, measured per capita (Pishgar-Komleh et al., 2011). Agricultural organizations in Iran have started the attempts toward self-sufficiency in rice production and statistics show that the annual production of this crop has increased in the last decade and achieved 2.3 Mt in 2010 (FAO, 2010). Thus, rice production plays a key role in food security in Iran but there are increasing concerns regarding Greenhouse Gas (GHG)

* Corresponding author. Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran. Tel.: þ98 2612801011; fax: þ98 2612808138. E-mail addresses: [email protected], [email protected] (A. Mohammadi), shahinrafi[email protected] (S. Rafiee). http://dx.doi.org/10.1016/j.jclepro.2014.05.008 0959-6526/Ó 2014 Elsevier Ltd. All rights reserved.

emissions related to cultivation of rice (Wenjun et al., 2006). Studies on environmental effects from flooded rice farming have particularly considered emissions of methane (CH4), nitrous oxide (N2O) and ammonia (NH3) (Linquist et al., 2012a; Hokazono and Hayashi, 2012). CH4 emissions in rice systems result in a high global warming impact relative to other crops, which contribute 10e13% of total methane emissions in the world (Neue, 1997). In some countries including Iran, besides the emissions, nonrenewable energy consumption and water use are also major concerns in rice farms mainly due to irrigation operation based on pumping systems. Furthermore, chemical fertilizers are among the most important inputs for increasing rice yield (Mohammadi et al., 2014), whereas it contributes to a wide range of soil and water pollution. Therefore efforts to identify pathways for mitigating environmental risks are required. Plenty of works is found on the environmental impact assessment for rice fields in countries such as Taiwan (Yang et al., 2009), China (Zhang et al., 2010), Japan (Koga and Tajima, 2011; Hokazono and Hayashi, 2012), Italy (Blengini and Busto, 2009) and USA (Linquist et al., 2012b), but in Iran very little attention has been paid to the GHG emissions from the

522

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

agricultural sector. Karimi et al. (2012) estimated that groundwater pumping for farm irrigation annually consumes 20.5 billion kWh electricity and 2 billion liters of diesel and contributes to 3.6% of the total CO2 emission of Iran. The improvement of irrigation practices can result in 40% reduction in energy use and consequently carbon emission of groundwater consumption. In another work, for wheat production, average GHG emissions of 1137 kg CO2 eq. ha1 and 291 kg CO2 eq. t1 were calculated (Soltani et al., 2013). Hence it is important to perform works to suggest options for reducing environmental costs in agricultural systems of Iran, particularly rice paddy since the harvested area has been increased in recent years (FAO, 2010). In this regard, Life Cycle Assessment (LCA) is one of the best methodologies for the environmental consequences of agri-food systems, by recognizing energy and inputs used as well as direct and indirect GHG emissions (Cherubini, 2010). LCA is commonly applied to assess all environmental impacts associated with a product, process or activity by accounting and evaluating the resource consumption and the emissions (Chang et al., 2012). LCA can be used to compare two or more systems or a system based on a reference. This methodology can also be combined with Data Envelopment Analysis (DEA) in order to face the assessment of multiple inputs and outputs for a large number of similar facilities. DEA is a performance measurement methodology used to quantify the comparative efficiency of a set of similar entities named Decision Making Units (DMUs) with multiple input/output data by some specific mathematical programming models (Samuel-Fitwi et al., 2012). In the current study, the results of the original life cycle impact assessment (LCIA) are compared to the computed target life cycle impact assessments (LCIA). This results in reduced environmental impacts for the calculated targets because for the same amount of output, a lower amount of inputs will be consumed (Samuel-Fitwi et al., 2012). Several recent studies applied the LCA þ DEA approach to improve the operational and environmental efficiency and enhance economic performance (Lozano et al., 2010; Iribarren et al., 2011; Vázquez-Rowe et al., 2012a; Mohammadi et al., 2013). In literature there are a few studies on LCA application on Iranian agriculture (Khoshnevisan et al., 2013a,b), but no work exists regarding the environmental impacts of rice paddy cultivation in this country. The purpose of this study was to assess the environmental impacts for rice paddy production in the different growing seasons and investigate improvement options by using an LCA and DEA approach. We use this methodology to achieve operational benchmarking and productive efficiency while evaluating the environmental performance of the rice paddy farms. The study also estimates the economic gains from the optimized values of physical inputs.

and economic information. Rice paddy farms in this region differed in growing season. Therefore, in order to work with homogenous groups, the farms were divided into two; spring and summer groups. Spring fields were cultivated rice paddy from the mid-April up to mid-September; and summer fields were sown it the first of June and harvested at end of October including seedling period. The length of seedling period for both the spring and summer fields varied from around 30e40 days.

2. Materials and methods

2.3. Technical Efficiency (TE) and Super efficiency (SE)

2.1. Study area, field selection and farming practices

The basic feature of DEA is that the Technical Efficiency (TE) score of each DMU depends on the performance of the sample of which it is a part (Martínez and Silveira, 2012). TE is defined as the ability of a DMU to produce maximum output given a set of inputs and technology level, and its score calculated by the ratio of sum of weighted outputs to the sum of weighted inputs. The efficiency score (q) in the presence of multiple- input and -output factor is calculated as (Mousavi-Avval et al., 2011; Omid et al., 2011):

North of Iran has been recognized as the best area for rice cultivation in Iran and representing over 80% of the national rice production. The present study was conducted in the Gorgan region (36 830 N and 54 480 E) of the Golestan province, located in north of Iran. In the study year, the average annual rainfall was 528 mm, maximum and minimum temperatures were 32  C and 6  C, respectively. The data for the study were collected from 82 rice paddy farmers in 10 villages in 2010. Farmers were asked to fill out a questionnaire (Mobtaker et al., 2010; Mohammadi et al., 2014) to characterize their usual field operation (pesticide and fertilizer use, tillage and harvesting) growing season (spring or summer), yield

2.2. Selected DEA model DEA is a well-known mathematical procedure that uses a linear programming (LP) technique to assess the efficiency of Decision Making Units (DMUs) or units of assessment. DEA allows for the measurement of comparative efficiency for a group of DMUs that use inputs in the form of different scales because the model adjusts with the ratio of the weighted sum of outputs to the weighted sum of inputs. Therefore the results could clearly be presented and simply compared efficient and inefficient DMUs. An input-oriented slacksbased measure of efficiency model with constant returns to scale (CRS) was selected for the current study. The selection of an inputoriented model for a LCA þ DEA work is due to this approach focuses on reducing input utilization and its associated environmental costs as much as possible (Vázquez-Rowe and Tyedmers, 2013; Mohammadi et al., 2013). CRS were chosen since paddy farms operate within a competitive market. Before 2010, the agricultural sector of Iran was supported by the government and the farm inputs were subsidized for the producers. Over this period, agricultural produce market couldn’t be assumed as a free market, but in recent years all subsides allocated to energy and farm inputs have been removed gradually. Hence, it has been assumed that paddy farms work in a competitive market, therefore the CRS approach can be implemented for this case (Mohammadi et al., 2013). The units of assessment chosen for this study are rice paddy fields. The most relevant farm input and output data (DEA matrices) for both the spring and summer systems as the sample of 82 rice paddy farms is gathered in Table 1. All the selected operational items for analysis are assumed to be independent of each other. Hence, land used was not involved since its minimization would influence other inputs, like diesel or fertilizers. Likewise, direct emissions to the different compartments are not comprised in the DEA matrix, given their direct proportion to some of the inputs included in the matrix. Thus, these emissions are indirectly minimized through direct minimization of the associated inputs (Lozano et al., 2009; Vázquez-Rowe et al., 2010). In order to provide the study with a stronger socioeconomic dimension, other inputs such as labor are included as an additional DEA input in the matrix (Vázquez-Rowe et al., 2012a).

TE ¼

weighted sum of outputs weighted sum of inputs

or mathematically as:

(1a)

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

523

Table 1 The farm input and output data (quantity per ha-annual) for DEA matrix for 82 rice paddy farms. DMU

F1-1 F1-2 F1-3 F1-4 F1-5 F1-6 F1-7 F1-8 F1-9 F1-10 F1-11 F1-12 F1-13 F1-14 F1-15 F1-16 F1-17 F1-18 F1-19 F1-20 F1-21 F1-22 F1-23 F1-24 F1-25 F1-26 F1-27 F1-28 F1-29 F1-30 F1-31 F1-32 F1-33 F1-34 F1-35 F2-1 F2-2 F2-3 F2-4 F2-5 F2-6 F2-7 F2-8 F2-9 F2-10 F2-11 F2-12 F2-13 F2-14 F2-15 F2-16 F2-17 F2-18 F2-19 F2-20 F2-21 F2-22 F2-23 F2-24 F2-25 F2-26 F2-27 F2-28 F2-29 F2-30 F2-31 F2-32 F2-33 F2-34 F2-35 F2-36 F2-37

Input

Output

Labor (h)

Machinery (h)

Diesel (L)

Water (m3)

Electricity (kWh)

Chemicalsa (kg)

Fertilizerb (kg)

Seed (kg)

Rice paddy (kg)

607 681 430 686 730 576 472 658 758 610 522 643 828 603 431 392 480 514 730 524 518 791 419 390 544 629 712 530 490 361 892 655 640 543 710 951 881 950 1004 863 1027 1164 1079 886 912 1090 816 920 990 1150 1110 1144 861 674 560 870 1062 982 721 759 488 1052 1060 1166 924 845 693 1055 1042 672 850 856

25 15 31 30 31 21 20 21 23 19 12 18 14 29 28 24 14 30 37 23 23 25 12 26 20 11 15 32 22 11 12 19 13 16 12 15 25 33 16 10 14 15 17 16 39 37 35 15 28 33 35 34 16 17 24 16 23 15 16 20 18 20 22 11 22 18 19 29 18 35 16 22

375 106 341 390 153 102 122 438 500 97 78 103 583 420 122 136 79 398 570 360 440 584 288 161 96 457 545 183 129 75 580 524 370 91 535 122 674 569 708 590 675 96 770 623 196 840 580 609 900 940 179 779 89 590 618 610 722 680 591 100 538 121 720 65 641 574 101 765 640 580 105 670

8130 11,300 7600 8600 10,700 9100 8230 8715 11,440 9420 7850 11,680 12,600 9340 8510 8920 8800 7550 10,900 7060 8440 10,490 6300 7980 10,560 7880 10,450 8300 9190 8170 12,090 11,200 9500 8510 10,710 14,500 14,750 13,500 14,400 13,200 14,500 14,900 15,490 13,140 13,760 16,250 13,020 13,200 15,740 16,640 15,200 15,610 12,710 12,900 13,300 13,760 15,200 13,950 12,930 11,600 11,580 13,070 14,500 15,610 13,100 12,480 11,390 15,770 13,780 11,900 14,120 13,400

0 5020 0 0 4883 4250 3870 0 0 4622 3650 5490 0 0 3910 4081 4140 0 0 0 0 0 0 3945 4990 0 0 3883 4185 3636 0 0 0 3897 0 6830 0 0 0 0 0 6940 0 0 6420 0 0 0 0 0 7185 0 5820 0 0 0 0 0 0 5450 0 6128 0 7360 0 0 5350 0 0 0 6690 0

14.5 17.0 24.0 16.5 14.0 14.0 18.0 17.5 16.0 13.0 19.5 22.0 21.0 14.5 15.0 18.0 24.0 12.0 17.0 14.5 26.0 12.0 22.0 17.0 22.0 15.5 7.0 15.5 15.0 12.0 30.0 17.0 20.0 18.0 25.0 21.5 20.0 17.0 17.0 19.0 13.0 5.0 16.0 14.0 22.5 21.0 16.0 11.0 18.0 24.0 22.5 23.0 12.0 14.0 14.5 20.0 26.0 18.0 15.5 18.0 14.0 13.0 13.5 17.0 21.0 10.0 14.0 26.0 23.0 5.0 19.0 12.0

332 282 464 282 348 232 232 332 298 273 232 332 314 200 282 132 448 282 282 248 364 232 332 298 332 314 50 182 225 82 439 282 282 364 439 332 300 382 250 232 232 0 282 150 414 332 282 0 232 414 464 398 257 248 348 439 257 232 214 332 282 132 439 248 332 282 332 248 100 182 232 332

65 65 70 60 90 80 60 55 70 55 65 60 55 60 65 60 60 60 70 75 70 70 60 70 95 80 70 60 80 65 100 70 60 70 65 65 65 65 65 75 65 60 60 75 60 60 70 65 70 70 60 55 85 70 70 90 80 55 65 65 55 60 65 55 90 65 65 60 80 75 60 55

6160 5600 6650 7000 5950 7000 5600 6160 4900 5740 5740 4480 5250 4900 6510 7350 5600 5740 4480 3500 5250 5740 4900 5460 6160 7700 5740 7350 6300 4200 4550 7000 5250 7000 4900 4900 4550 3500 4200 3150 4200 4340 4900 3500 4340 6300 5040 3500 6160 5950 5950 3850 5600 4900 4480 3500 3850 5040 4900 4900 3360 5950 3850 3500 2800 5110 5250 4480 4200 5950 4200 4200 (continued on next page)

524

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

Table 1 (continued ) DMU

Input

F2-38 F2-39 F2-40 F2-41 F2-42 F2-43 F2-44 F2-45 F2-46 F2-47 F1-Mean F2-Mean

Output

Labor (h)

Machinery (h)

Diesel (L)

Water (m3)

Electricity (kWh)

Chemicalsa (kg)

Fertilizerb (kg)

810 660 1000 1072 1117 780 782 1030 880 925 591 919

14 35 10 17 18 29 25 27 16 24 21 22

630 569 70 95 846 141 522 138 103 610 301 491

13,270 12,690 12,540 14,280 16,630 13,070 12,180 13,600 14,050 12,510 9320 13,823

0 0 5840 6630 0 6090 0 6260 6510 0 1956 2032

11.0 7.0 19.5 23.0 30.0 16.0 13.0 15.0 14.0 19.0 18 17

332 339 314 282 332 114 348 382 314 150 287 278

Seed (kg) 60 70 80 50 55 70 55 60 60 65 68 66

Rice paddy (kg) 3500 3150 4340 3710 3850 5250 5950 5250 3150 5740 5766 4516

Note: F1 ¼ Planted in Spring; F2 ¼ Planted in Summer. The number following the dash in the DMU column represents the specific rice paddy fields within the system. DMU: Decision Making Unit. a Fertilizer comprises urea, P2O5 and K2O. b Chemicals includes herbicides and insecticides.

Ps u1 y1j þ u2 y2j þ / þ us ysj ur yrj q ¼ ¼ Pr¼1 m v1 x1j þ v2 x2j þ / þ um xmj i¼1 vi xij

l0 (1b)

TE of DMUj can be estimated by solving the following function (Mohammadi et al., 2011; Mousavi-Avval et al., 2012):

Ps

Maximize q ¼ Pr¼1 m

ur yrj

i¼1 vi xij

(2)

Subject to

Ps Pr¼1 m

ur yrj

i¼1 vi xij

1

j ¼ 1; 2; .; n

ur ; vi  0 where, q is the TE of the DMU under consideration; x and y are the amount of input and output, v and u are the weights of input and output, respectively; s is the number of outputs (r ¼ 1, 2, ., s), m is number of inputs (i ¼ 1, 2, ., m) and j denotes jth DMUs (j ¼ 1, 2, ., n). Eq. (2) can equivalently be written as a linear programming (LP) problem as follows (Omid et al., 2011; Mobtaker et al., 2012):

Maximize q ¼

Xs

uy r¼1 r ro

(3)

vx i¼1 i io uy r¼1 r rj



Xm

vx i¼1 i ij

In reality, the dual (envelopment) from of the DEA-LP problem is simpler to solve the above equation due to fewer constraints. Mathematically, the dual linear problem is written in vector-matrix notation as follows (Omid et al., 2011):

Subject to:

Y l  y0 X l  qx0  0

Xn

 yro  q0 xi0

1

lj  0

ur ; vi  0

Minimize q

ly j¼1 i rj

l j¼1 j

j ¼ 1; 2; .; n

(5)

Subject to

Xn

Xn

¼ 1

Xs

Minimize q0

lx j¼1 j ij

Subject to

Xm

where yo is the s  1 vector of the value of original outputs produced and xo is the m  1 vector of the value of original inputs used by the oth paddy farm. Y is the s  n matrix of outputs and X is the m  n matrix of inputs of all n units included in the sample, and l is a n  1 vector of weights. The measure of efficiency q, is bounded between 0 and 1. Concretely, if q ¼ 1, means that the DMU will be considered efficient whereas, 0  q < 1, symbolizes that the DMU is inefficient. A number of works (Mohammadi et al., 2011; Iribarren et al., 2011; Mousavi-Avval et al., 2012) have shown that it is possible to rank efficient DMUs on the basis of weights to input values assigned by all DMUs. Super Efficiency (SE) is an approach to achieve this purpose. In the SE method, a DMU under evaluation is excludes from the reference set so that efficient DMUs may have efficiency scores larger than or equal to 1. The input-oriented CRS superefficiency model can be defined as follows (Lee and Zhu, 2012):

(4)

where q0 is the SE of efficient DMUs. For the present study, two different DEA matrices were made up for both systems according to Table 1 and conducted with the use of the EMS e i.e. Efficiency Measurement System e software (Barr, 2004; Mousavi-Avval et al., 2011). 2.4. LCA þ DEA framework A common criticism that LCA studies receive is the difficulty in managing multiple input/output data, since the way data are dealt with may cause an important influence on the obtained results (Iribarren et al., 2010). The most extended way to solve this problem is an average inventory which comprises the average values for

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

the different inputs and outputs. However, the high degree of variability often associated with multiple data sets may create large standard deviations. The combined use of LCA with DEA as suggested by Lozano et al. (2009) and Vázquez-Rowe et al. (2010), could be an alternative approach to reduce the impact of increased standard deviations, while providing useful additional information in terms of result interpretation (Vázquez-Rowe et al., 2012b). In the current case study, the high number of DMUs to be handled leads us to consider the use of this joint method. As mentioned in the previous section, the unit of assessment or DMU refers to each rice paddy field. Fig. 1 depicts the items included in the LCA þ DEA framework of rice paddy farms. Since DEA just involves a selection of the inputs and outputs in an LCA study, DEA and LCA items are separated in Fig. 1. LCA and DEA inputs involved key practices of the rice paddy production, such as irrigation, fertilization, spraying and harvesting. LCA outputs embraced not only the product of rice paddy, but also unwanted generation as well as direct emissions associated with energy, water, fertilizers and chemicals, whereas rice paddy crop was the unique DEA output. Operational and environmental aspects for this study can be summarized into the five steps LCA þ DEA method (Lozano et al., 2010; Mohammadi et al., 2013). The first step (i) is life cycle inventory (LCI) for every DMUs which involves data collection regarding the input and output flow of the rice paddy fields. The second step (ii) is the life cycle impact assessment (LCIA) for each of the DMU. This step aims to estimate the characterization of the environmental profile of the current DMUs from the LCI developed in the previous step. The third step (iii) deals with DEA model from the LCIs of the first step and computation of the target DMUs. The operational and environmental impact efficiency of each DMU is estimated, and the target values are calculated for the inefficient DMUs. In this stage, for identification of a set of best-performing farms for environmental benchmarking, the super-efficiency model was also conducted to compute new efficiency scores for ranking efficient DMUs. In the fourth step (iv) LCIA of the target DMUs are performed according to the new LCI data arising from the previous step and finally, the potential environmental impacts

525

determined for the virtual DMUs. The last step (v) involves the interpretation of the results based on eco-efficiency criteria. Consequently, the current and target values for the environmental impact categories and economic issue are compared for the spring and summer rice paddy fields. 2.5. Field emissions and environmental impact categories Field emissions comprise direct air emissions of methane (CH4), nitrous oxide (N2O) and ammonia (NH3), as well as emissions of nitrates and phosphorus to water. These values were calculated according to Blengini and Busto (2009); Hokazono and Hayashi (2012). Nitrous oxide and methane emissions were based on (Dong et al., 2012) and the Tier 1 method, described in the 2006 IPCC Guidelines for National Greenhouse Gas inventories (IPCC, 2006). The above-ground residues were excluded from the calculations because of removing from the fields, but the below-ground residues considered for N2O emission according to (IPCC, 2006). Electricity generation from natural gas source was assumed to be used at the grid (Mohammadi et al., 2013). It was also assumed that land-use changes related to rice farming are not major contributors of environmental impacts. Since land-use change due to the rice area expansion does not often occur in the area. The SimaPro 7 (Pré, 2009) software application was utilized to implement the LCA model and perform the assessment. Global Warming (GWP), Acidification (AP), Eutrophication (EP), Cumulative non-renewable Energy Demand (CED) and Water Depletion (WD) were used for the impact assessment of the current study. These categories were chosen based on a thorough review of LCA references in the grain production (Blengini and Busto, 2009; Knudsen et al., 2010; Hokazono and Hayashi, 2012). GWP and CED are of main concern in LCAs all over the world, and agriculture has a significant contribution to acidification and eutrophication. Therefore these impacts can be important when comparing agriculture production systems (Hokazono and Hayashi, 2012). WD impact expresses the total amount of fresh water consumption (direct and indirect use) throughout the life cycle. Extracting and using water in excess can

LCA items Direct & indirect emissions

DEA items

Seedling bed

Tilling (Disc-leveler) Puddling Covering

(CO2, CH4, N2O, NH3, …)

SeedlingPlant in the field

Sowing

Irrigation

Irrigation

Weeding

Spraying (insecticide)

Fertilization (N, P, K)

Transplanting

Spraying (herbicide& insecticide)

Standing rice paddy

Harvested grain (product)

Harvested straw (byproduct)

Threshing

Energy (Diesel & electricity)

Transportation

Fig. 1. LCA and DEA items for each Decision Making Unit (DMU, rice paddy field).

526

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

cause very significant damages to ecosystems and human health in Iran as a dry country. The characterized LCA results for the impact categories were based on the inventory data in Table 1. The functional unit of the systems under study was 1 kg of rice paddy produced and delivered to the rice mill. 3. Results and discussion 3.1. Inventory data Data from questionnaire showed that the type of inputs used between spring and summer rice paddy systems are almost same, but a significant difference was observed in the consumption level of some elements in the rice paddy farming. The average of fertilizer and chemicals amounts in both systems were nearly similar to each other (mean data not shown). However, the energy consumption was 37% and 33% larger for diesel and electricity, respectively, on summer compared to spring farms. In the area, some rice farms are irrigated by irrigation channel, but other rice operators use electric pumps to bring water to their field. The groundwater resources are largely used to provide the farms water requirements. The need for labor was higher on summer farms principally due to water and time spend on irrigation. The yields were also 22% higher for spring compared to summer farms. Straw was removed from the both spring and summer fields and sold for some building utilizations at relatively low prices. Since this study assesses the economy performance of rice paddy fields, an economic allocation was applied between rice paddy and straw and around 2.5% of impact was allocated to straw. It varied from 2.3% to 3.1% depending on the rice paddy yield in spring and summer systems. Blengini and Busto (2009) reported that the allocation value of rice straw is no environmental burdens, due to its negligible market price. However, in literature the economic allocation is one of the most common allocation procedures used for rice production (Hokazono and Hayashi, 2012; Williams et al., 2005). 3.2. Current life cycle impact assessment The statistical analysis showed no significant differences between farm types when the environmental impacts is measured per hectare, whereas the difference between the systems is significant when environmental impact is measured per kg rice paddy produced. This difference between farm types is mainly due to the

100

Percentage (%)

80

lower amount of some inputs and the higher yield in the spring farms. Fig. 2, illustrates that GWP, AP, EP, CED and WD in spring systems are 34%, 29%, 30%, 35% and 44% lower than the summer systems (measured per kg rice paddy produced). When the environmental impacts were measured per hectare, these values were 15%, 8%, 17%, 16% and 26%, respectively. Table 2 presents the impacts relevant to the LCA model for spring and summer rice paddy. As observed, the emissions gave a total GWP impact of 1263 and 1911 g CO2-eq for spring and summer fields, respectively. The difference is mainly due to higher irrigation in the summer fields which results in the higher direct emissions and energy for irrigation. As it is seen in Table 2, the summer fields have relatively higher values for all the categories. The contributions to the GWP for rice paddy were mainly caused by CH4 emissions (41% for both systems) and energy for irrigation (35% for spring and 39% for summer farms). In addition to enhancing water use efficiency for reducing CH4 emissions, the removal of rice straw has also a great potential to mitigate CH4 emissions and as well as total GHG emissions for a large range of rice cropping systems (Naser et al., 2007). Koga and Tajima (2011) concluded that CH4 emissions from straw-incorporated rice paddy fields represented 78% of total GHG emissions. In the case of AP and EP of rice paddy, NH3 emissions are the greatest contributor to these impacts, followed by energy for irrigation especially diesel. Likewise, the contributions to CED come largely from energy for irrigation (71% and 76% for spring and summer farms) and chemical fertilizer production (16% for spring and 13% for summer systems). WD is mainly influenced by water use for irrigation and other field operations around 70% and 77% for spring and summer fields. The direct use of water for irrigation appears to be particularly intense for the summer systems which approximately 3.2 m3 kg1 of rice paddy in comparison with 1.7 m3 kg1 of the spring rice paddy. However, if the indirect use of fresh water is also investigated, the WD indicator for summer paddies 4.8 m3 kg1 and for spring paddies 2.7 m3 kg1 is obtained. This number is close to Pishgar-Komleh et al. (2011) which have reported the irrigation water requirement of rice in Iran. The average rate of urea fertilizer and fuel consumption in rice paddy farms were observed to be 116 kg ha1 and 410 l ha1 which highly influence on the AP and EP impacts. Since direct field emissions (especially NH3) are the largest contributors to these impacts, which depend exclusively on applied N fertilizers (Wang et al., 2010), improving fertilizer use efficiency could reduce the AP and EP of rice paddy fields in the region. In this scenario, green manures could be a good option. Green manures are utilized in many farming systems as a source of N fertilizer because of its ability to fix atmospheric N2. In rice systems a green manure crop is grown before the rice and incorporated into the soil prior planting, which can have considerable effects on direct emissions (Adhya et al., 2000; Linquist et al., 2012a). 3.3. DEA performance

60

40

20

0 GWP/ha AP/ha EP/ha CED/ha WD/ha Summer

GWP/kg AP/kg EP/kg CED/kg WD/kg Spring

Fig. 2. Difference among the impacts for both systems; measured per hectare and per kg rice paddy produced. Showed for GWP ¼ Global Warming Potential, AP ¼ Acidification Potential, EP ¼ Eutrophication Potential, CED ¼ Cumulative nonrenewable Energy Demand and WD ¼ Water Depletion.

A DEA study was carried out based on the LCI data available to establish the DEA matrix consisting of the most relevant inputs and outputs. The DEA optimization model results in Technical Efficiency (TE) for each rice paddy field (Table 3) and a definition of operational targets for the considered inputs and outputs. The mean TE for the spring fields (80%) is found higher that summer fields (71%). In addition, target operational benchmarks were determined. These points that turn inefficient units into efficient are also shown in Table 3 as reduction percentages of the current operational values. Those facilities with technical efficiency TE ¼ 1 (i.e. the efficient farms) were excluded in this Table. 26 farms (32% of the sample) were found to operate efficiently and the share of spring farms (46%) was over two times higher than of summer farms (21%). The

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

527

Table 2 Environmental characterization for 1 kg of spring (spr.) and summer (sum.) rice paddy production, in the form of Global Warming Potential (GWP), Acidification Potential (AP), Eutrophication Potential (EP), Cumulative non-renewable Energy Demand (CED) and Water Depletion (WD), (Spr. ¼ Spring, Sum. ¼ Summer).

Indirect emissions Chemical fertilizers Pest control-biocides Diesel for farm operations Diesel for irrigation Electricity for irrigation Irrigation & farm operations Direct emissions N2O CH4 NH3 NOx NO3 Total

WD (m3)

GWP (g CO2 eq.)

AP (g SO2 eq.)

EP (g NO3 eq.)

CED (MJ)

Spr.

Sum.

Spr.

Sum.

Spr.

Sum.

Spr.

Sum.

Spr.

Sum.

77 13 80 131 307 e

97 16 107 332 412 e

0.3 0.3 0.7 1.1 0.5 e

0.4 0.4 0.9 2.8 0.6 e

1.0 1.0 1.2 2.0 0.5 e

1.3 1.2 1.6 4.9 0.7 e

1.6 0.3 1.0 1.7 5.4 e

2.0 0.3 1.4 4.3 7.2 e

0.4 0.1 e e 0.3 1.9

0.6 0.1 e e 0.4 3.7

137 519 e e e 1263

169 779 e e e 1911

e e 6.7 0.2 e 9.9

e e 8.4 0.3 e 13.8

e e 13.0 0.5 2.1 21.3

e e 16.2 0.6 4.0 30.5

e e e e e 9.9

e e e e e 15.2

e e e e e 2.7

e e e e e 4.8

application of a super efficiency DEA model to discriminate among the efficient units is discussed in the next section. The target operating inputs calculated through DEA in the third step led to calculate new LCIAs involving the modified LCI data for the inefficient farms. Thus, a new environmental characterization was computed for all of the inefficient ones, in order to estimate their potential environmental impacts if they worked under efficient operational conditions. Fig. 3 shows the environmental impacts per kg of output of the original inefficient DMUs in comparison with those associated with their virtual targets. Usually, the environmental impacts in the virtual targets are lower than the one of the original DMUs due to the optimization of the operational inputs. Fig. 4 graphically represents an important average reduction in the environmental impacts for the five categories included in the study, with average reductions between 8% (CED) and 11% (EP) for the spring systems, and 19% (GWP) and 25% (EP) for the summer systems. As expected, the reduction potential of the summer farms was considerably higher than the spring farms for all of impacts. A contribution analysis has also been carried out with reference to rice paddy and presented in Fig. 5. This figure can be helpful to consider the role of the different subsystems that contribute to the life cycle impact reductions of rice paddy production. The direct field emissions have a high potential in reducing the environmental impact. With regard to GWP, CH4 is the main contributor in the total reduction of this impact for spring (56%) and summer (37%) fields. NH3 had the highest share to the AP and EP declines, the largest contributor for CED reduction came from the electricity (53%) for both systems, also the contribution of irrigation and farm operations is the biggest for the WD reductions. With respect to the environmental improvement linked to operational benchmarking, important objectives were achieved in the two rice paddy systems of the region. Results confirmed that the link between operational efficiency and environmental impacts is possible by optimizing of material inputs usage in order to reduce potential environmental impacts, as it proved in earlier studies (Lozano et al., 2010; Vázquez-Rowe et al., 2011). In other words, the main core of the LCA þ DEA method is the eco-efficiency concept. Units with an inefficient operation should convert to an efficient performance so that environmental gains are obtained. The joint application of LCA and DEA provides a quantitative proof of efficiency scores and target operational and environmental benefits (Iribarren et al., 2011). Previous studies with regard to the LCA of rice cultivation have illustrated the important role of input resources and farm operations for greenhouse gas emissions (Breiling et al., 2005; Roy et al., 2007).

3.4. Super-efficiency analysis One of the possible reasons to undertake an LCA þ DEA study is to identify a set of best-performing units for environmental benchmarking. The application of DEA to the sample of 82 rice paddy fields distinguished 26 efficient farms. When the number of DMUs is large, a wide range of units is expected to be found efficient. Discrimination among efficient DMUs facilitates the identification of a smaller set of best performers, which can help to refine the detection of best operational practices. In this sense, Iribarren et al. (2010) proposed the use of super-efficiency analyses to rank best units by assigning efficiency score greater than 1. In the present study, an input-oriented slack-based measure of super efficiency model with CRS was selected for the discrimination among the efficient farms (Tone, 2002) and applied for the spring and summer systems separately. New efficiency scores calculated for the 16 spring efficient and 10 summer efficient DMUs are shown in Table 4. As observed, the super efficiency ranges from 1.0 to 1.62 and if a cut-off criterion of J > 1.2 is followed, just 9 rice paddy farms would be selected from the both systems. This cut-off criterion (J > 1.2) was chosen to show that the majority of these superefficient performers (6 out of 9) are rice paddy planted in the spring season. Consequently, these farms could be taken as a reference unit in order to suggest benchmarks for environmental decision making. Pollard et al. (2008) mentioned this kind of studies may be interesting and useful for policy makers and regulators in order to provide environmental reference values. 3.5. Economic savings The integration of environmental, economic and social indicators has been mentioned as a suitable solution for the limitations of LCA and improvement of sustainability assessment tools (Reap et al., 2008). Vázquez-Rowe et al. (2010) suggested that LCA þ DEA methodology can include an economic dimension to the environmental assessment by adding a consideration of the operational performance. Likewise, there are several studies which have used the five-step LCA þ DEA method to a set of agricultural systems from an economic view (Lozano et al., 2009; Iribarren et al., 2010; Vázquez-Rowe et al., 2011). Hence, this approach has been applied for the economic information of the current study. Table 5 illustrates the economic savings that could be attained if the inefficient rice paddy farms were to improve their performance in order to score efficient operation. To estimate this economic analysis, the average prices of energy (diesel fuel and electricity), chemicals, fertilizers and seed inputs in 2010 were

528

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

Table 3 Technical efficiency (TE) and operational reduction (%) of inputs for the inefficient rice paddy fields. DMU

F1-1 F1-2 F1-5 F1-7 F1-10 F1-11 F1-15 F1-17 F1-18 F1-21 F1-23 F1-24 F1-25 F1-27 F1-29 F1-30 F1-32 F1-33 F1-35 F2-1 F2-2 F2-3 F2-4 F2-5 F2-6 F2-8 F2-9 F2-10 F2-11 F2-12 F2-14 F2-15 F2-16 F2-17 F2-19 F2-20 F2-21 F2-22 F2-23 F2-25 F2-26 F2-28 F2-29 F2-31 F2-32 F2-33 F2-34 F2-36 F2-37 F2-38 F2-39 F2-40 F2-41 F2-42 F2-45 F2-46 F1-Mean F2-Mean

TE

0.97 0.94 0.84 0.67 0.81 0.92 0.98 0.60 0.57 0.77 0.60 0.86 0.99 0.67 0.96 0.57 0.80 0.94 0.77 0.79 0.80 0.69 0.78 0.58 0.69 0.80 0.57 0.59 0.90 0.68 0.85 0.74 0.81 0.60 0.75 0.68 0.39 0.50 0.94 0.79 0.77 0.61 0.71 0.33 0.92 0.83 0.68 0.67 0.69 0.75 0.58 0.95 0.70 0.65 0.81 0.54 0.80 0.71

Operational reduction (%) Labor

Machinery

Diesel

Water

Electricity

Chemicals

Fertilizer

Seed

0.0 26.0 10.4 3.9 22.6 32.3 2.7 0.0 0.0 0.0 20.2 8.0 0.0 21.8 0.0 14.9 0.0 17.1 16.4 36.8 0.0 20.0 22.4 23.1 29.2 36.7 8.2 32.9 33.4 5.0 19.4 21.7 42.6 27.0 0.7 0.0 5.0 12.7 43.2 12.3 29.1 0.0 46.5 2.3 27.8 23.0 24.9 12.6 32.6 14.2 1.5 49.2 42.0 30.3 29.7 14.6 10.3 22.0

0.0 0.0 0.0 0.0 2.3 0.0 3.2 13.0 24.4 6.7 0.0 0.0 1.9 0.0 0.0 0.0 19.8 0.0 0.0 0.0 7.7 37.7 0.0 0.0 0.0 0.0 0.0 21.6 14.6 3.3 0.0 0.0 7.5 9.9 0.0 14.8 0.0 0.0 0.0 0.0 8.3 0.0 0.0 0.0 0.0 0.0 4.3 0.0 0.0 0.0 3.8 0.0 0.0 0.0 9.4 0.0 3.8 3.9

0.0 0.0 0.0 4.5 0.0 33.4 0.0 9.5 0.0 10.7 18.2 10.6 0.0 22.3 0.0 10.1 0.0 0.0 19.7 0.0 9.7 7.3 19.9 20.2 25.0 34.2 3.8 16.9 48.2 11.6 39.0 35.4 0.0 32.5 12.7 21.9 4.4 12.4 40.2 21.7 0.0 20.8 0.0 0.0 17.6 0.0 32.0 0.9 0.0 22.2 2.2 0.0 0.0 33.1 0.0 0.0 7.3 14.8

4.1 29.5 0.0 11.6 22.0 41.3 15.7 4.5 3.9 9.2 16.2 0.0 17.3 26.9 3.4 18.1 4.0 24.4 25.9 33.9 14.7 18.5 22.3 27.4 30.3 33.2 8.7 21.0 42.4 14.3 30.2 27.7 36.3 29.7 19.6 29.4 11.9 15.0 40.8 29.4 22.6 22.2 42.7 1.2 26.3 22.4 30.7 7.4 30.5 16.6 6.2 42.0 32.4 32.6 18.7 19.8 14.6 24.6

0.0 30.3 3.6 0.0 23.0 0.0 17.5 0.0 0.0 0.0 0.0 0.0 18.3 0.0 6.2 0.0 6.4 0.0 0.0 38.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 21.7 0.0 0.0 0.0 0.0 36.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 23.4 0.0 43.4 0.0 0.0 23.3 0.0 0.0 32.4 0.0 0.0 42.9 32.6 0.0 18.3 20.7 5.5 9.0

2.5 5.5 4.9 2.3 13.3 31.4 31.1 1.9 7.2 22.4 5.2 3.8 35.9 20.4 1.2 26.1 12.8 35.3 33.6 10.3 12.4 40.5 36.6 28.1 7.7 4.2 26.6 12.4 19.3 4.2 15.6 11.5 25.2 20.6 11.9 6.6 2.0 21.8 25.1 21.0 8.8 1.7 13.1 8.0 22.0 0.9 30.8 38.7 4.7 6.2 14.7 24.9 33.9 31.1 7.7 2.9 15.6 16.6

24.2 24.1 27.5 0.6 33.4 13.7 36.4 0.0 0.5 1.0 19.0 2.6 30.3 42.0 16.6 15.0 26.5 24.8 29.0 30.7 44.7 59.1 56.8 15.7 11.0 4.2 36.8 41.4 13.9 3.5 2.2 10.0 52.4 21.8 1.4 13.2 5.6 5.0 0.0 9.9 32.4 9.5 2.4 8.9 45.8 36.1 4.8 21.9 13.8 19.7 27.5 47.1 44.8 7.9 59.3 17.9 19.3 22.7

10.2 0.0 0.6 0.0 0.0 0.0 0.0 0.0 10.6 0.0 0.0 0.0 19.3 0.0 11.9 9.3 5.2 0.0 0.0 0.0 0.0 0.0 0.0 11.6 0.0 0.0 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.2 0.0 0.0 0.0 0.0 0.0 0.0 4.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 32.9 0.0 0.0 0.0 0.0 3.5 1.4

Note: F1 ¼ Planted in Spring; F2 ¼ Planted in Summer. The number following the dash in the DMU column represents the specific paddy rice fields within the system. DMU: Decision Making Unit.

gathered from the local agricultural organizations that sell these inputs to the farmers. These organizations work under control of agricultural ministry of Iran. As mentioned earlier, there is a significant difference in consumption level of some inputs between spring and summer growing seasons, which it affects the production costs of rice paddy. Hence the economic evaluation was separately carried out for each system. The economic results revealed that a total saving of near to 3600 $ was gained for the

spring sample and it ranged from 14 to 667 $ ha1 for inefficient farms. This value for the summer sample was obtained around 7400 $ that it varied from 12 to 1168 $ ha1 for inefficient ones. The average economic saving per 1 kg rice paddy for inefficient farms was found to be 0.041 $ for spring rice paddies and 0.046 $ for summer rice paddies (Table 5). According to mean values, the major contributions to the economic saving came from machinery and chemicals for both systems.

3500

GWP (g CO2 eq.)

3000

2500 2000 1500 1000 500

DMU

25

AP (g SO2 eq.)

20 15 10 5 0

DMU

50

EP (g NO3 eq.)

40 30 20 10 0

DMU 30

CED (MJ)

25

20 15 10 5 0

DMU

WD (m3)

8 6 4 2 0

DMU Current

Target

Fig. 3. Environmental impacts potentials of inefficient farms and virtual targets for 1 kg rice paddy production. (GWP ¼ Global Warming Potential, AP ¼ Acidification Potential, EP ¼ Eutrophication Potential, CED ¼ Cumulative non-renewable Energy Demand, WD ¼ Water Depletion, and DMU ¼ Decision Making Unit). 30

4. Conclusions Percentage (%)

25

In the current study, an environmental impact efficiency analysis of a set of 82 farms (35 spring farms and 47 summer farms) for rice paddy production was performed using Life Cycle Assessment combined with the Data Envelopment Analysis (LCA þ DEA methodology). The direct link between operational and environmental performance of the units was highlighted based on the utilization of physical inputs. The LCA results implied that the spring rice paddy has a lower environmental impact, with regard to global warming, acidification, eutrophication, non-renewable energy demand and water depletion per kg produced compared to the summer rice paddy. After optimization of physical inputs by DEA, the results indicated that the direct field emissions had the high potential in reducing the environmental effects for both rice paddy

20 15 10 5 0 GWP

AP

EP Spring

CED

WD

Summer

Fig. 4. Average reductions for GWP ¼ Global Warming Potential, AP ¼ Acidification Potential, EP ¼ Eutrophication Potential, CED ¼ Cumulative non-renewable Energy Demand and WD ¼ Water Depletion.

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

WD

CED

EP

AP

GWP

530

Spring Summer Spring Summer Spring Summer Spring Summer

Spring Summer 0%

20%

40%

60%

80%

Synthetic fertilizers

Biocides

Diesel-field operatios

Diesel-irrigation

Electricity

N2O

CH4

NH3

NOx

NO3

Irrigation & field operations

100%

Fig. 5. Contribution of subsystems to the environmental impact reductions. (GWP ¼ Global Warming Potential, AP ¼ Acidification Potential, EP ¼ Eutrophication Potential, CED ¼ Cumulative non-renewable Energy Demand and WD ¼ Water Depletion).

Table 4 Superior Efficient scores (SE) for the farms that were detected as efficient. DMU

J

DMU

J

DMU

J

F1-3 F1-4 F1-6 F1-8 F1-9 F1-12 F1-13 F1-14 F1-16

1.26 1.19 1.12 1.19 1.03 1.00 1.23 1.11 1.05

F1-19 F1-20 F1-22 F1-26 F1-28 F1-31 F1-34 F2-7 F2-13

1.29 1.02 1.03 1.40 1.38 1.57 1.09 1.16 1.05

F2-18 F2-24 F2-27 F2-30 F2-35 F2-43 F2-44 F2-47 e

1.11 1.62 1.42 1.01 1.05 1.33 1.03 1.08 e

Note: F1 ¼ Planted in Spring; F2 ¼ Planted in Summer. The number following the dash in the DMU column represents the specific paddy rice fields within the system. DMU: Decision Making Unit.

systems. According to the DEA model, the percentage of fields which were deemed efficient was lower for summer (21%) in comparison with spring systems (46%). Among the efficient units, a super-efficiency analysis was also carried out to distinguish the best performing units, which could be used as a reference value for environmental impacts. Finally, input reductions were evaluated from an economic dimension in order to estimate economic saving of rice paddy systems in target condition. Outcomes confirmed that significant extra benefits could be obtained if all the farms within the sample were operated in an efficient way. In summary, we believe that this study can offer important guidance for decision makers regarding the relative efficiency of Iranian agriculture. This is achieved by the ability of the

Table 5 Total Economic Saving (TES) linked to the accomplishment of operational targets. DMU

F1-1 F1-2 F1-5 F1-7 F1-10 F1-11 F1-15 F1-17 F1-18 F1-21 F1-23 F1-24 F1-25 F1-27 F1-29 F1-30 F1-32 F1-33 F1-35 F2-1 F2-2 F2-3 F2-4

Economic saving ($/ha/yr)

TES ($/kg paddy rice)

Labor

Machinery

Diesel

Electricity

Chemicals

Fertilizer

Seed

0.0 15.6 6.2 2.3 13.6 19.4 1.6 0.0 0.0 0.0 12.1 4.8 0.0 13.1 0.0 8.9 0.0 10.3 9.8 22.1 0.0 12.0 13.4

0.0 0.0 0.0 0.0 57.5 0.0 80.0 325.0 610.0 167.5 0.0 0.0 47.5 0.0 0.0 0.0 495.0 0.0 0.0 0.0 192.5 942.5 0.0

0.0 0.0 0.0 0.5 0.0 3.3 0.0 1.0 0.0 1.1 1.8 1.1 0.0 2.2 0.0 1.0 0.0 0.0 2.0 0.0 1.0 0.7 2.0

0.0 3.0 0.4 0.0 2.3 0.0 1.8 0.0 0.0 0.0 0.0 0.0 1.8 0.0 0.6 0.0 0.6 0.0 0.0 3.8 0.0 0.0 0.0

12.5 27.5 24.5 11.5 66.5 157.0 155.5 9.5 36.0 112.0 26.0 19.0 179.5 102.0 6.0 130.5 64.0 176.5 168.0 51.5 62.0 202.5 183.0

4.1 4.1 4.7 0.1 5.7 2.3 6.2 0.0 0.1 0.2 3.2 0.4 5.2 7.1 2.8 2.6 4.5 4.2 4.9 5.2 7.6 10.0 9.7

20.4 0.0 1.2 0.0 0.0 0.0 0.0 0.0 21.2 0.0 0.0 0.0 38.6 0.0 23.8 18.6 10.4 0.0 0.0 0.0 0.0 0.0 0.0

0.006 0.009 0.007 0.003 0.032 0.035 0.044 0.075 0.191 0.053 0.008 0.004 0.043 0.027 0.006 0.033 0.117 0.042 0.053 0.020 0.084 0.278 0.042

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

531

Table 5 (continued ) DMU

F2-5 F2-6 F2-8 F2-9 F2-10 F2-11 F2-12 F2-14 F2-15 F2-16 F2-17 F2-19 F2-20 F2-21 F2-22 F2-23 F2-25 F2-26 F2-28 F2-29 F2-31 F2-32 F2-33 F2-34 F2-36 F2-37 F2-38 F2-39 F2-40 F2-41 F2-42 F2-45 F2-46 F1-Mean F2-Mean

Economic saving ($/ha/yr)

TES ($/kg paddy rice)

Labor

Machinery

Diesel

Electricity

Chemicals

Fertilizer

Seed

13.9 17.5 22.0 4.9 19.7 20.0 3.0 11.6 13.0 25.6 16.2 0.4 0.0 3.0 7.6 25.9 7.4 17.5 0.0 27.9 1.4 16.7 13.8 14.9 7.6 19.6 8.5 0.9 29.5 25.2 18.2 17.8 8.8 6.2 13.2

0.0 0.0 0.0 0.0 540.0 365.0 82.5 0.0 0.0 187.5 247.5 0.0 370.0 0.0 0.0 0.0 0.0 207.5 0.0 0.0 0.0 0.0 0.0 107.5 0.0 0.0 0.0 95.0 0.0 0.0 0.0 235.0 0.0 93.8 96.6

2.0 2.5 3.4 0.4 1.7 4.8 1.2 3.9 3.5 0.0 3.3 1.3 2.2 0.4 1.2 4.0 2.2 0.0 2.1 0.0 0.0 1.8 0.0 3.2 0.1 0.0 2.2 0.2 0.0 0.0 3.3 0.0 0.0 0.7 1.5

0.0 0.0 0.0 0.0 2.2 0.0 0.0 0.0 0.0 3.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.3 0.0 4.3 0.0 0.0 2.3 0.0 0.0 3.2 0.0 0.0 4.3 3.3 0.0 1.8 2.1 0.6 0.9

140.5 38.5 21.0 133.0 62.0 96.5 21.0 78.0 57.5 126.0 103.0 59.5 33.0 10.0 109.0 125.5 105.0 44.0 8.5 65.5 40.0 110.0 4.5 154.0 193.5 23.5 31.0 73.5 124.5 169.5 155.5 38.5 14.5 78.1 82.9

2.7 1.9 0.7 6.3 7.0 2.4 0.6 0.4 1.7 8.9 3.7 0.2 2.2 1.0 0.9 0.0 1.7 5.5 1.6 0.4 1.5 7.8 6.1 0.8 3.7 2.3 3.3 4.7 8.0 7.6 1.3 10.1 3.0 3.3 3.9

23.2 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.4 0.0 0.0 0.0 0.0 0.0 0.0 9.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 65.8 0.0 0.0 0.0 0.0 7.1 2.7

0.052 0.014 0.007 0.029 0.103 0.082 0.018 0.024 0.015 0.078 0.107 0.016 0.081 0.003 0.024 0.046 0.030 0.079 0.004 0.021 0.008 0.030 0.006 0.067 0.049 0.014 0.014 0.040 0.063 0.053 0.034 0.058 0.009 0.041 0.046

Note: F1 ¼ Planted in Spring; F2 ¼ Planted in Summer. The number following the dash in the DMU column represents the specific paddy rice fields within the system. DMU: Decision Making Unit.

sustainability-oriented LCA þ DEA approach to deal with operational, environmental and economic aspects. However, this method has proven to be helpful to embrace different aspects when multiple inventory data are available for multiple similar entities, although more attempts are required to develop further uses of the LCA þ DEA methodology.

References Adhya, T.K., Bharati, K., Mohanty, S.R., Ramakrishnan, B., Rao, V.R., Sethunathan, N., Wassmann, R., 2000. Methane emission from rice fields at Cuttack, India. Nutri. Cycl. Agroecosys. 58, 95e105. Barr, R.S., 2004. DEA Software and Technologyda State-of-the-Art Survey. In: International Series in Operations Research and Management Science, vol. 71. Springer, pp. 539e566. Blengini, G.A., Busto, M., 2009. The life cycle of rice. LCA of alternative agri-food chain management systems in Vercelli (Italy). J. Environ. Manag. 90, 1512e1522. Breiling, M., Hashimoto, S., Sato, Y., Ahamer, G., 2005. Rice-related greenhouse gases in Japan, variations in scale and time and significance for the Kyoto protocol. Paddy Water Environ. 3, 39e46. Chang, N., Qi, C., Islam, K., Hossain, F., 2012. Comparisons between global warming potential and cost-benefit criteria for optimal planning of a municipal solid waste management system. J. Clean. Prod. 20, 1e13. Cherubini, F., 2010. GHG balances of bioenergy systems e overview of key steps in the production chain and methodological concerns. Renew. Energy 35, 1565e 1573. Dong, N.M., Brandt, K.K., Sørensen, J., Hung, N.N., Hach, C.V., Tan, P.S., Dalsgaard, T., 2012. Effects of alternating wetting and drying versus continuous flooding on fertilizer nitrogen fate in rice fields in the Mekong Delta, Vietnam. Soil Biol. Biochem. 47, 166e174. FAO, 2010. Available from: http://www.faostat.fao.org/. Hokazono, S., Hayashi, K., 2012. Variability in environmental impacts during conversion from conventional to organic farming: a comparison among three rice production systems in Japan. J. Clean. Prod. 28, 101e112.

Hormozi, M.A., Asoodar, M.A., Abdeshahi, A., 2012. Impact of mechanization on technical efficiency: a case study of rice farmers in Iran. Proc. Econ. Finan. 1, 176e185. IPCC, 2006. IPCC Guidelines for National Greenhouse Gas Inventories. Intergovernmental Panel of Climate Change (IPCC). National Greenhouse Gas Inventories Programme. Online at: http://www.ipcc-nggip.iges.or.jp/public/ 2006gl/index.html. Iribarren, D., Hospido, A., Moreira, M.T., Feijoo, G., 2011. Benchmarking environmental and operational parameters through eco-efficiency criteria for dairy farms. Sci. Total Environ. 409, 1786e1798. Iribarren, D., Vázquez-Rowe, I., Moreira, M.T., Feijoo, G., 2010. Further potentials in the joint implementation of life cycle assessment and data envelopment analysis. Sci. Total Environ. 408, 5265e5272. Karimi, P., Qureshi, A.S., Bahramloo, R., Molden, D., 2012. Reducing carbon emissions through improved irrigation and groundwater management: a case study from Iran. Agric. Water Manag. 108, 52e60. Khoshnevisan, B., Rafiee, S., Omid, M., Mousazadeh, H., Clark, S., 2013b. Environmental impact assessment of tomato and cucumber cultivation in greenhouses using life cycle assessment and adaptive neuro-fuzzy inference system. J. Clean. Prod. http://dx.doi.org/10.1016/j.jclepro.2013.09.057. Khoshnevisan, B., Rafiee, S., Omid, M., Mousazadeh, H., Sefeedpari, P., 2013a. Prognostication of environmental indices in potato production using artificial neural networks. J. Clean. Prod. 52, 402e409. Knudsen, M.T., Yu-Hui, Q., Yan, L., Halberg, N., 2010. Environmental assessment of organic soybean (Glycine max.) imported from China to Denmark: a case study. J. Clean. Prod. 18, 1431e1439. Koga, N., Tajima, R., 2011. Assessing energy efficiencies and greenhouse gas emissions under bioethanol-oriented paddy rice production in northern Japan. J. Environ. Manag. 92, 967e973. Lee, H.S., Zhu, J., 2012. Super-efficiency infeasibility and zero data in DEA. Eur. J. Operat. Res. 216, 429e433. http://www.scopus.com/authid/detail.url? authorId¼6602789530&eid¼2-s2.0-84863526982. Linquist, B.A., Adviento-Borbe, M.A., Pittelkow, C.M., van Kessel, C., van Groenigen, K.J., 2012a. Fertilizer management practices and greenhouse gas emissions from rice systems: a quantitative review and analysis. Field Crops Res. 135, 10e12. Linquist, B.A., van Groenigen, K.J., Adviento-Borbe, M.A., Pittelkow, C.M., van Kessel, C., 2012b. An agronomic assessment of greenhouse gas emissions from major cereal crops. Glob. Change Biol. 18, 194e209.

532

A. Mohammadi et al. / Journal of Cleaner Production 106 (2015) 521e532

Lozano, S., Iribarren, D., Moreira, M.T., Feijoo, G., 2009. The link between operational efficiency and environmental impacts. A joint application of life cycle assessment and data envelopment analysis. Sci. Total Environ. 407, 1744e1754. Lozano, S., Iribarren, D., Moreira, M.T., Feijoo, G., 2010. Environmental impact efficiency in mussel cultivation. Resour. Conserv. Recycl. 54, 1269e1277. Martínez, P., Silveira, S., 2012. Analysis of energy use and CO2 emission in service industries: evidence from Sweden. Renew. Sustain. Energy Rev. 16, 5285e5294. Mobtaker, H.G., Keyhani, A., Mohammadi, A., Rafiee, S., Akram, A., 2010. Sensitivity analysis of energy inputs for barley production. Agricult. Ecosyst. Environ. 137, 367e372. Mobtaker, H.G., Akram, A., Keyhani, A., Mohammadi, A., 2012. Optimization of energy required for alfalfa production using data envelopment analysis approach. Energy Sustain. Dev. 16, 242e248. Mohammadi, A., Rafiee, S., Jafari, A., Dalgaard, T., Knudsen, M.T., Keyhani, A., Mousavi-Avval, S.H., Hermansen, J.E., 2013. Potential greenhouse gas emission reductions in soybean farming: a combined use of life cycle assessment and data envelopment analysis. J. Clean. Prod. 54, 89e100. Mohammadi, A., Rafiee, S., Jafari, A., Keyhani, A., Mousavi-Avval, S.H., Nonhebel, S., 2014. Energy use efficiency and greenhouse gas emissions of farming systems in north Iran. Renew. Sustain. Energy Rev. 30, 724e733. Mohammadi, A., Rafiee, S.H., Mohtasebi, S.S., Mousavi-Avval, S.H., Rafiee, H., 2011. Energy efficiency improvement and input cost saving in kiwifruit production using data envelopment analysis approach. Renew. Energy 36, 2573e2579. Mousavi-Avval, S.H., Mohammadi, A., Rafiee, S., Tabatabaeefar, A., 2012. Assessing the technical efficiency of energy use in different barberry production systems. J. Clean. Prod. 27, 126e132. Mousavi-Avval, S.H., Rafiee, S., Jafari, A., Mohammadi, A., 2011. Improving energy use efficiency of canola production using data envelopment analysis (DEA) approach. Energy 36, 2765e2772. Naser, H.M., Nagata, O., Tamura, S., Hatano, R., 2007. Methane emissions from five paddy fields with different amounts of rice straw application in central Hokkaido. Jpn. Soc. Soil Sci. Plant Nutrit. 53, 95e101. Neue, H.U., 1997. Fluxes of methane from rice fields and potential for mitigation. Soil Use Manag. 13, 258e267. Omid, M., Ghojabeige, F., Ahmadi, H., Delshad, M., 2011. Energy use pattern and benchmarking of selected greenhouses in Iran using data envelopment analysis. Energy Convers. Manag. 52, 153e162. Pishgar-Komleh, S.H., Sefeedpari, P., Rafiee, S., 2011. Energy and economic analysis of rice production under different farm levels in Guilan province of Iran. Energy 36, 5824e5831. Pollard, S.J.T., Davies, G.J., Coley, F., Lemon, M., 2008. Better environmental decision making e recent progress and future trends. Sci. Total Environ. 400, 20e31. Pré, 2009. SIMAPRO 7.1.8 PRé Consultants, Printerweg 18, 3821 AD Amersfoort, The Netherlands. www.pre.nl.

Reap, J., Roman, F., Duncan, S., Bras, B., 2008. A survey of unresolved problems in life cycle assessment e part 1: goal and scope and inventory analysis. Int. J. Life Cycle Assess. 13, 290e300. Roy, P., Shimizu, N., Okadome, H., Shiina, T., Kimura, T., 2007. Life cycle of rice: challenges and choices for Bangladesh. J. Food Eng. 79, 1250e1255. Samuel-Fitwi, B., Wuertz, S., Schroeder, J.P., Schulz, C., 2012. Sustainability assessment tools to support aquaculture development. J. Clean. Prod. 32, 183e192. Soltani, A., Rajabi, M.H., Zeinali, E., Soltani, E., 2013. Energy inputs and greenhouse gases emissions in wheat production in Gorgan, Iran. Energy 50, 54e61. Tone, K., 2002. A slacks-based measure of super-efficiency in data envelopment analysis. Eur. J. Operat. Res. 143, 32e41. Vázquez-Rowe, I., Iribarren, D., Hospido, A., Moreira, M.T., Feijoo, G., 2011. Computation of operational and environmental benchmarks within selected Galician fishing fleets (NW Spain). J. Indus. Ecol. 15, 776e795. Vázquez-Rowe, I., Iribarren, D., Moreira, M.T., Feijoo, G., 2010. Combined application of life cycle assessment and data envelopment analysis as a methodological approach for the assessment of fisheries. Int. J. Life Cycle Assess. 15, 272e283. Vázquez-Rowe, I., Moreira, M.T., Feijoo, G., 2012b. Review: best practices in life cycle assessment implementation in fisheries. Improving and broadening environmental assessment for seafood production systems. Trends Food Sci. Technol. 28, 116e131. Vázquez-Rowe, I., Tyedmers, P., 2013. Identifying the importance of the ‘‘skipper effect’’ within sources of measured inefficiency in fisheries through data envelopment analysis (DEA). Mar. Policy 38, 387e396. Vázquez-Rowe, I., Villanueva-Rey, P., Iribarren, D., Moreira, M.T., Feijoo, G., 2012a. Joint life cycle assessment and data envelopment analysis of grape production for vinification in the RíasBaixas appellation (NW Spain). J. Clean. Prod. 27, 92e102. Wang, M., Xia, X., Zhang, Q., Liu, J., 2010. Life cycle assessment of a rice production system in Taihu region, China. Int. J. Sustain. Dev. World Ecol. 17, 157e161. Wenjun, Z., Yanhong, Q., Zhiguo, Z., 2006. A long-term forecast analysis on worldwide land uses. Environ. Monit. Assess. 119, 609e620. Williams, A.G., Audsley, E., Sandars, D.L., 2005. Final Report to Defra on Project IS0205: Determining the Environmental Burdens and Resource Use in the Production of Agricultural and Horticultural Commodities. Department of Environment, Food, and Rural Affairs (Defra), London. Available from: http:// www2.defra.gov.uk/research/Project_Data/More.asp%3FI%3DIS0205%26M% 3DKWS%26V%3DERM. Yang, S.S., Lai, C.M., Chang, H.L., Chang, E.H., Wei, C.B., 2009. Estimation of methane and nitrous oxide emissions from paddy fields in Taiwan. Renew. Energy 34, 1916e1922. Zhang, A., Cui, L., Pan, G., Li, L., Hussain, Q., Zhang, X., Zheng, J., Crowley, D., 2010. Effect of biochar amendment on yield and methane and nitrous oxide emissions from a rice paddy from Tai Lake plain, China. Agricult. Ecosyst. Environ. 139, 469e475.