Content uploaded by William Benjamin Putman
Author content
All content in this area was uploaded by William Benjamin Putman on Nov 12, 2021
Content may be subject to copyright.
Journal of Environmental Management 302 (2022) 114062
0301-4797/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Cradle-to-grave life cycle assessment of production and consumption of
pulses in the United States
Prathamesh A. Bandekar
a
,
*
, Ben Putman
a
, Greg Thoma
b
,
**
, Marty Matlock
a
a
Department of Biological and Agricultural Engineering, University of Arkansas, 203 Engineering Hall, Fayetteville, AR, 72701, United States
b
Ralph E Martin Department of Chemical Engineering, University of Arkansas, 3153 BELL Engineering, Fayetteville, AR, 72701, United States
ARTICLE INFO
Keywords:
Life cycle assessment
Pulses
Environmental impact
Dry beans
Pea
Lentil
ABSTRACT
Environmental impact associated with production and consumption of pulses in the United States was evaluated
using life cycle assessment (LCA). The system boundary was set to cradle-to-grave with a functional unit of 60 g
(dry basis) of pulses consumed in a US household. Varieties of pulses modeled in the study included eld pea
(Pisum sativum), lentil (Lens culinaris), chickpea (Cicer arietinum), and dry bean. Three methods of cooking
pulses at the consumer stage tested in the study were cooking in open vessel on electric cooking range (OVC),
cooking in stovetop pressure cooker on electric cooking range (SPC), and cooking in electric pressure cooker
(EPC). OVC formed the base scenario against which all other scenarios were compared. The environmental
impact of pulses varied with type of pulse crop, cooking method, and the batch size. Consumption of approxi-
mately 60 g of dry pulses resulted in the greatest environmental impact for OVC. The consumer stage contributed
at least 83, 81, 76, 75, and 87 percent for global warming potential (GWP), fossil resource scarcity (FRS), water
consumption (WC), freshwater eutrophication (FE), and marine eutrophication (ME), respectively for this sce-
nario. EPC resulted in the greatest decrease in the environmental impact, compared to OVC, for GWP, FRS, FE,
and ME for all pulse varieties, which was validated in the uncertainty analysis. SPC, on the other hand, decreased
the impact across these categories only for chickpea and dry bean. The uncertainty analysis suggested that the
differences associated with cooking methods in the mean land use and water consumption scores of pulses were
statistically non-signicant. The impact categories were also highly sensitive to the mass of pulses cooked in a
batch. Increasing the reference ow in OVC to 1 kg decreased the environmental impact of pulses by 49–87
percent for all impact categories, excluding land use. Overall, the study identied the consumer stage as the
hotspot for environmental impact in the supply chain of pulses in the United States. The large contribution of the
consumer stage to the overall environmental impact of pulses was attributed to electricity consumption for
cooking and associated upstream emissions.
1. Introduction
Growing population, dwindling resources, and changing climate
have increased the pressure on agriculture to improve production and
efciency while maintaining or improving sustainability of the sector.
The food sector contributes 19 to 29 percent of global anthropogenic
greenhouse gas (GHG) emissions and agriculture is the largest contrib-
utor of CH
4
and N
2
O emissions (MacWilliam et al., 2018). A few major
crops such as corn, rice, and wheat cover approximately 40 percent of
global arable land and satisfy 50 percent of caloric demand of global
population (Ebert, 2014). Overreliance on few major crops to meet the
demands of growing population could be agronomically, environmen-
tally, and economically perilous. These crops require substantial amount
of synthetic nitrogen (N) fertilizers which results in increased GHG
emissions from agriculture (MacWilliam et al., 2018). Monoculture also
increases pesticide demand of the sector and results in
pest-accumulation due to lack of crop diversity (MacWilliam et al.,
2015). Therefore, diversication in crop production is important to
improve pest and nutrient management, food production, and overall
sustainability of the agriculture sector.
Pulses, which include leguminous crops such as dry beans, eld peas,
chickpeas, and lentils, when included in crop rotation, can play a major
* Corresponding author.
** Corresponding author. 3149 BELL Engineering, Fayetteville, AR, 72701, United States.
E-mail addresses: pbandeka@uark.edu (P.A. Bandekar), wputman@uark.edu (B. Putman), gthoma@uark.edu (G. Thoma), mmatlock@uark.edu (M. Matlock).
Contents lists available at ScienceDirect
Journal of Environmental Management
journal homepage: www.elsevier.com/locate/jenvman
https://doi.org/10.1016/j.jenvman.2021.114062
Received 17 August 2021; Received in revised form 25 October 2021; Accepted 2 November 2021
Journal of Environmental Management 302 (2022) 114062
2
role in achieving these objectives by breaking disease and insect cycles
and improving soil fertility (MacWilliam et al., 2015). Pulses have an
ability to x atmospheric nitrogen to meet most of their nitrogen de-
mand. The synthetic N fertilizer demand of pulses ranges between 11
and 56 kg N/ha (Brouwer et al., 2015; Franzen, 1998; Kandel et al.,
2018; Schatz and Endres, 2009) while that of corn ranges between 110
and 280 kg N/ha (Halvorson and Bartolo, 2014; Kim et al., 2009; Kim
and Dale, 2008). This reduced reliance of pulses on synthetic N fertilizer
offer various environmental and agronomic benets. The production of
synthetic N fertilizers is energy intensive and their application to soil
results in GHG emissions, marine eutrophication, and atmospheric
acidication. These impacts can be mitigated by including pulses in crop
rotation, which also benets following cereal crop in terms of improved
yield and protein content (MacWilliam et al., 2015).
Pulses can be an excellent source of protein in human diets. Pulses
contain 18 to 36 percent protein and are rich in nutrients, vitamins, and
minerals (FAO, 2016). Furthermore, high levels of complex carbohy-
drates and ber can help stabilizing blood sugar levels, while also
providing a feeling of satiety. Chaudhary et al. (2018) reported that
when rened wheat our in pan bread, breakfast cereal, and pasta was
partially replaced by Canadian yellow pea our, the nutrient balance
score of these products improved by 11, 70, and 18 percent and
decreased GHG emission by 4, 11, and 13 percent, respectively.
Consuming pulses such as dry beans and peas was found to increase
ber, protein, folate, zinc, iron, and magnesium intake in human diet
while reducing intake of saturated fat and total fat (Mitchell et al.,
2009).
However, evaluation of potential benets and risks associated with
any changes made to the existing cropping system is important before
these changes are incorporated. Life cycle assessment (LCA), a measur-
able and quantiable framework for such assessment, can be valuable
for researchers, growers, and policy makers in making informed de-
cisions (ISO, 2006a). While LCA studies of pulse production are avail-
able for Canada and a few other parts of the world (Kulshreshtha et al.,
2013; MacWilliam et al., 2014a, 2015; Nemecek et al., 2008; Tidåker
et al., 2021), only one study exists specic to the US, which exported
11% of global pulse exports in 2017 (Bond, 2019). Gustafson (2017)
reported an LCA of US pulse production using survey data collected in
six states and covering ve pulse crops. The study estimated that GHG
emissions associated with pulse crop production were 0.26 and 0.31 kg
CO
2
e/kg for non-irrigated and irrigated crops, respectively. The irriga-
tion water use was 0.19 m
3
/kg, lower than many other row crops.
However, this study did not follow many of the commonly used and
internationally standardized methods for performing life cycle assess-
ment and included only two impact categories. The results for these two
impact categories were aggregated for all types of pulse crops and
differentiated only between irrigated and non-irrigated crops. Also, the
underlying survey data excluded North Dakota, one of the largest pulse
production states in the United States (USDA National Agricultural
Statistics Services, 2017). Furthermore, the study was ‘cradle to farm-
gate’ and did not consider post-farmgate processes, which is necessary to
provide a holistic sustainability picture of pulse crops. Assessment of
impacts associated with both ‘cradle to farmgate’ and ‘post-farmgate’
supply chains, including consumption stage, could be important in
evaluating and improving sustainability of agricultural sector in general
and of pulse production sector specically. The objective of this study
was to perform a ‘cradle to grave’ attributional LCA of pulse crop pro-
duction and consumption in the US using national average production
and consumption practices for the most commonly grown peas, lentils,
chickpeas, and dry beans.
2. Material and methods
Production and consumption of pulses was modeled in OpenLCA
(GreenDelta). The background processes involved in production, pro-
cessing, retail, and cooking of pulses were modeled using ‘EcoInvent 3.4
– allocation, cut-off by classication’ database (Wernet et al., 2016). The
model was divided into four stages: crop production, processing, retail,
and consumer stage. Process boundaries for each stage encompassed
gate-to-gate activities, except for crop production. For example, the
processing stage included all activities from transportation of harvested
pulses to the processing facility to loading packaged pulses into
tractor-trailer containers for distribution to retail. On the other hand, the
boundary for crop production stage was set to cradle-to-farmgate.
2.1. Goal and scope of study
The primary goal of this study was to evaluate impacts associated
with production and consumption of pulses in the United States using
attributional LCA. The impacts of pulses were evaluated in terms of
global warming potential (GWP) estimated over 100-year horizon, fossil
resource scarcity (FRS), land use (LU), water consumption (WC),
freshwater eutrophication (FE), and marine eutrophication (ME), using
ReCiPe 2016 (H) midpoint life cycle impact assessment (LCIA) method
(Huijbregts et al., 2017). These impact categories characterized sus-
tainability of pulse supply chain in the United States.
2.1.1. Functional unit
The functional unit (FU) quanties the product studied and denes
the reference ows for all the inputs and outputs. A functional unit of 60
g of pulses, cooked and consumed in the US household, was selected for
this study. The functional unit represented current average weekly
consumption of pulses in the United States (HHS and USDA, 2015). The
cooking methods evaluated in the study include boiling or
pressure-cooking pulses in water until they are cooked. Generally,
cooked pulses are used as an ingredient in recipes such as soups, salads,
spreads or can be consumed with rice. However, formulating and eval-
uating these recipes was out of scope for this study.
2.1.2. System boundary
Dening system boundary is crucial in LCA (ISO, 2006a). The system
boundary determines the processes in the product life cycle that are
included or excluded from analysis. The system boundary for this study
was cradle (production of seeds and other agronomic inputs and crop
production) to grave (consumption of pulses at consumer’s home). The
processes included in the system boundary are illustrated in Fig. 1.
Resource use and wastage at each stage were fully accounted for each
process. Consumption of pulses away from home was excluded from the
study. The system boundary also excluded processing and consumption
of various nished products (hummus, canned beans, soups etc.) con-
taining pulses. The consumer stage of the analysis was restricted to
purchase, cooking, and consumption of dry pulses only. A cutoff crite-
rion of 1% was established for mass ows and/or environmental impact
categories. However, data were included regardless of cutoff criterion if
they were readily available.
2.1.3. Allocation methodology
Allocation of resources and burden is required for a process with
multiple outputs. An ISO 14044 allocation hierarchy (ISO, 2006b) was
followed in this study for allocation of inputs and emissions. The pri-
mary byproduct of harvesting at the farming stage is crop residue, which
is often left on the soil (USA Dry Pea & Lentil Council, 2019). Although
the crop residue may provide nutrients to the crops planted in the
following season (Bedard-Haughn et al., 2013; Miller et al., 2015), the
system boundary excluded recycling of soil nutrients and the burden of
material, resources, and emissions was allocated to the harvested pulses.
A single processing plant often processes several crops. Therefore, the
system specic to the pulses was separated from processing of other
crops at the processing facility. Processing pulses primarily produces
seed coat and sometimes broken and powdered pulses. Due to lack of
data regarding fate of these materials, they were treated as waste
disposed in the municipal landll. Therefore, inputs and emissions were
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
3
allocated to the packaged pulses. For multifunctional activities such as
retail allocation based on shelf space occupied was adopted. A
revenue-based approach was adopted at the consumer stage to attribute
transportation associated with grocery purchase as well as refrigeration
load and microwave usage to pulses when necessary.
2.2. Life cycle inventory
Data for life cycle inventory (LCI) was obtained from peer-reviewed
manuscripts, crop budgets, extension documents published by the uni-
versities, technical specications published by the manufacturers of
crop processing machineries, and various publicly available data re-
positories and sources. We also consulted experts from the universities
and the United States Department of Agriculture (USDA) for crop pro-
duction data such as seeding rate, fertilizer application rates, and tillage
practices.
2.2.1. Cradle-to-farmgate
Cradle-to-farmgate activities constitute the rst stage in the cradle-
to-grave system boundary. It includes seed production and transport,
production and transport of fertilizers and pesticides, and all other on-
farm activities associated with production of pulses.
2.2.1.1. Crop production. The pulse production methods and related
data were obtained from expert opinion and from crop budgets and
published extension documents. Crop yield was obtained from USDA-
NASS survey data. Based on expert opinion, lentils, eld peas, and
chickpeas were modeled as no-till, dry land crops. This represented the
general pulse production practices in Montana and North Dakota. Pro-
duction practices could vary in other pulse production states. However,
no state-specic data were available. Moreover, in 2018 Montana and
North Dakota together produced 86 and 81 percent of total national
production of eld pea and lentil, respectively (USDA National Agri-
cultural Statistics Services, 2017). Therefore, the production practices in
these two states were assumed to represent national average. Data
provided by experts included fertilizer application rates, seeding rates,
and information about types of chemicals used (Miller et al., Personal
Communication). Fertilizer application rates suggested by the experts
were similar to those used by MacWilliam et al. (2014a, 2014b) for dry
pea and lentil production. The dry bean production practices varied
from other pulse crops (Miller et al., Personal Communication). How-
ever, in absence of specic data, the dry bean production was modeled
as a conventionally tilled, dryland crop (Brouwer et al., 2015). Fertilizer
application rates for dry beans were considered as an average of the
upper and lower threshold provided by Brouwer et al. (2015). Produc-
tion data for pulse production is provided in Table 1.
Fig. 1. Conceptual model with system boundaries and processes in production and consumption of pulses in the United States.
Table 1
Life cycle inventory for cradle-to-retail stage for variety of pulses. Data in the
table are presented for the reference ow of each stage.
Parameter Chickpea Dry bean Field pea Lentil
Farming stage
Seeding rate, kg/ha 179.33 146.83 168.13 56.04
Yield, kg/ha (Reference ow) 1769.83 1922.60 2028.45 1342.20
Nitrogen fertilizer, kg/ha 5.60 96.13 5.60 5.60
Phosphorous fertilizer, kg/ha 28.02 33.63 28.02 28.02
Potassium fertilizer, kg/ha 8.41 8.41 8.41 8.41
Pendimethalin, kg a.i./ha 1.24 – 1.24 1.24
Metolachlor, kg a.i./ha 1.60 – 1.60 1.60
Paraquat, kg a.i./ha 0.54 0.54 0.54 0.54
Glyphosate, kg a.i./ha – 1.68 – –
Dimethenamid, kg a.i./ha – 0.86 – –
Processing stage
Reference ow, kg 1.00 1.00 1.00 1.00
De-stoning electricity
a
, kWh 4.46
E−04
4.46
E−04
4.46
E−04
4.46
E−04
Grading, electricity
b
, kWh 3.34
E−05
3.34
E−05
3.34
E−05
3.34
E−05
Decorticating, electricity
b
,
kWh
1.25
E−05
1.25
E−05
1.25
E−05
1.25
E−05
Optical sorting, electricity
b
,
kWh
1.79
E−04
1.79
E−04
1.79
E−04
1.79
E−04
Splitting, electricity
b,c
, kWh 1.25
E−03
1.25
E−03
1.25
E−03
1.25
E−03
LDPE lm, kg 6.06
E−03
6.06
E−03
6.06
E−03
6.06
E−03
Water, kg 0.13 0.13 0.13 0.13
Transportation, tkm 0.152 0.152 0.152 0.152
Pulses hauled from the farm,
kg
1.52 1.52 1.52 1.52
Pulses processed after de-
stoning, kg
1.33 1.33 1.33 1.33
Retail Stage
Pulses, reference ow, kg 1.00 1.00 1.00 1.00
Electricity, kWh 0.02 0.02 0.02 0.02
Transportation, tkm 0.48 0.48 0.48 0.48
Packaged pulses purchased
from processing plant
1.06 1.06 1.06 1.06
a
Electricity consumption estimated for 1.52 kg of pulses delivered from farm.
b
Electricity consumption estimated for 1.33 kg of pulses processed.
c
Electricity consumption was assumed equal to decorticating operation due to
lack of data.
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
4
The pulse production includes herbicide and fungicide applications,
and pre-harvest chemical desiccation using Paraquat (Miller et al.,
Personal Communication). Fungicide applications are particularly
important for chickpea production. Herbicides, pesticides, and respec-
tive application rates were selected based on data reported by Brouwer
et al. (2015), Kandel et al. (2013, 2018), and Schatz and Endres (2009)
and available background data in the EcoInvent database. The appli-
cation rate for Paraquat (a desiccant) was obtained from Syngenta’s
(2019) website. Application rates of all chemicals were modeled in
OpenLCA as mass of active ingredient (a.i.) per hectare (Table 1).
Pulses require rhizobium bacteria to facilitate nitrogen xation
(Kandel et al., 2013). These bacteria are introduced to the soil through
inoculants applied either directly to the soil or through seed treatment.
The exact composition of inoculants was unavailable. However, Mac-
William et al. (2014b) reported that inoculants usually contain peat
moss. Therefore, process for mining peat moss was used as a surrogate
process for inoculants.
Field application of fertilizers often leads to nutrient loss in the form
of denitrication, leaching, and ammonia volatilization. Considering
their contribution to GWP, FE, and ME, accounting for these nutrient
losses in LCA model is crucial for accurate analysis. Direct emissions
from nitrogen fertilizer application were estimated using IPCC tier-2
method while IPCC tier-1 method was used to estimate indirect emis-
sions (IPCC, 2006). The N
2
O emission factor of 0.21% estimated by
Dusenbury et al. (2008) for wheat-pea cropping system in the semiarid
northern Great Plains was used in the IPCC tier-2 method for direct
emissions. This emission factor was less than the default emission factor
of 1% suggested by IPCC (2006). Lower fertilizer induced N
2
O emissions
in the semiarid regions were also conrmed by Sainju et al. (2020, 2012)
and Thies et al. (2020). Phosphorus applications often result in loss of
soluble phosphorus through leaching and runoff. These pathways were
modeled using the method provided by Potter et al. (2006).
Post-application fate of crop protection chemicals as well as desiccants
used prior to harvest were modeled as emissions to soil.
2.2.1.2. Seed production and fertilizer transportation. In 1997, annual
seed expenditure by farmers in the United States had reached $7 billion,
making it the largest seed market in the world (Fernandez-Cornejo,
2004). This $6.5 billion increase in expenditure, compared to 1960, was
largely attributed to increase in the share of seed purchased from com-
mercial sources as a result of technological developments and plant
breeding techniques. This makes seed production, processing, and
transport a crucial process in terms of LCA.
Commercial seed production processes are proprietary and there-
fore, are not available in the public domain. In absence of these data, a
seed production process ‘Pea seed production, for sowing | pea seed, for
sowing | Cutoff, U’ available in EcoInvient 3.4 database was adapted for
this study. The unit process included processes such as pre-cleaning,
cleaning, drying, chemical dressing, bag lling, and storage. Four
distinct seed production processes were created, each for a specic pulse
crop modeled (dry beans, chickpeas, lentils, eld peas). The source of
seed production and electricity was replaced with relevant crop pro-
duction processes modeled in OpenLCA and US electricity generation
and distribution network, respectively. However, only the source of
these processes was changed. We did not change the life cycle inventory
data of any input processes.
The 2017 Commodity Flow Survey published by US Bureau of
Transportation Statistics was used to determine average transportation
distance and contribution from various modes of transportation. In the
United States, single mode transportation dominated the sector
contributing 92.1% of total mass moved and 81% of total value of
shipment. However, about 71% of mass (73% of value of shipment) was
moved by trucks in the United States. Therefore, transportation of seeds
was modeled as freight transport by road. The average transportation
distance of 196 km between seed production plants and the seed
distributor was used (Bureau of Transportation Statistics, 2018).
Commercial, conventional agriculture depends heavily on fertilizer
use. Production and application of fertilizers dominate the impacts
associated with fertilizer use in agriculture (Hasler et al., 2015). To
account for contribution of fertilizer production, unit processes in
EcoInvent 3.4 database for nitrogen, phosphorus, and potassium fertil-
izers were used. These processes included production of ammonium
nitrate phosphate, monoammonium phosphate, and production of po-
tassium fertilizers from various sources. The transportation distances
were modied to represent the United States transportation sector. The
transportation of fertilizers from production plant to the distributor was
modeled as freight by road to a distance of 214 km (Bureau of Trans-
portation Statistics, 2018).
2.2.2. Processing stage
The processing stage in the LCA model included transportation of
harvested crop to the processing plant, processing of pulses, and
bagging. Harvested pulses may need to be cleaned, dried, sorted, split,
milled, decorticated, and fractioned before they are bagged and shipped
to the retail markets (USA Dry Pea & Lentil Council, 2019). The pro-
cessing steps depend on intended use of pulses and sometimes, addi-
tional steps such as roasting, pufng, and grinding may be necessary.
The transportation distance between a farm and grain elevator varies
depending on proximity to the pulse processing plant. Data specic to
transportation distances of pulses are not available. However, O’Donnell
(2008) reported that wheat is usually grown within 100 km from pro-
cessing plants in northwest and central United States. Because pulses are
grown in northwest United States and most of the machinery that pro-
cesses pulses is also designed to handle wheat (Bühler, 2019a), a
transportation distance of 100 km was adopted for this study.
The output of the processing stage in this study was raw, processed
pulses, packed in 1 kg bags. Pulse processing steps included in the model
involved destoning, grading, decorticating, sorting using optical sorter,
and splitting (Wood and Malcomson, 2011). Electricity consumed for
each processing step was calculated using technical specications of
machinery obtained from Bühler (2019a, 2019b, 2019c, 2019d) and the
approach presented by St¨
ossel (2018) was used, when necessary, to ll
data gaps at processing stage. The resulting electricity consumption was
rst normalized for 1 kg of pulses processed using the throughput
specied in technical specications. When throughput was unavailable,
an average of available data was used. Technical specications were
unavailable for splitting operations. Therefore, electricity consumption
equal to decortication process was assumed for splitting because of
similarities in the processes.
The pulse processing results in considerable losses in the form of
husk, powder, broken, shriveled, and unprocessed pulses. These losses
can amount to up to 25% of total pulses processed (Patras et al., 2011).
However, stones and other debris collected during harvesting were not
considered in the losses estimated by Patras et al. (2011). In absence of
specic data, it was assumed that stones and debris accounted for 12.5%
(half of losses) of harvested pulses hauled from the farm. Therefore,
electricity consumption for destoning was estimated for 1.52 kg of
pulses brought in for processing while that for other operations was
adjusted to 1.33 kg of pulses processed (Table 1).
The decortication (also called dehulling) primarily removes seed
coat; however, small broken pulses and powder is also removed during
this process. Pulses can be decorticated using either wet or dry process.
The wet process is primarily used to produce decorticated and split
pulses, while dry decortication is used to produce both split and whole
pulses (Wood and Malcomson, 2011). Because splitting was modeled as
a separate process in the study, we assumed decortication by dry pro-
cess. The dry decortication process requires prior conditioning with
water or tempering with oil followed by drying to ease seed coat
removal and to avoid breakage, especially for chickpea and eld pea
that are hard to decorticate (Wood and Malcomson, 2011). Lentil and
dry bean varieties are easy to decorticate and are processed directly
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
5
without conditioning or tempering.
For chickpea and eld pea, conditioning prior to decortication was
modeled assuming addition of water at the rate of approximately 10%
(w:w), soaking for 4–8 h, and subsequent drying to 7–11% moisture
content (Wood and Malcomson, 2011). For 1.33 kg of chickpea and eld
pea processed, 0.133 kg of water was added. It was assumed that the
pulses were harvested at 12% moisture content (USA Dry Pea & Lentil
Council, 2019) and all water added during conditioning was absorbed.
The amount of water evaporated during drying (0.1613 kg) was esti-
mated by mass balance. The output of the processing stage included 1 kg
of pulses packed in a low-density polyethylene (LDPE) bag transported
to retails stores. Weights of empty packaging bags of pulses were
measured and modeled as 6.06 g of LDPE bag per kg of nal product.
2.2.3. Retail stage
The processes in retail stage included transportation of packaged
pulses from processing plants to retail stores, storage of these pulses at
the stores, food losses at the retail, electricity consumption by the
establishment, and land occupation. Input data used for the retail sector
are presented in Table 1. The output of retail stage model was 1 kg of
pulses stocked at the retail store. According to the USDA Economic
Research Service (2019) on an average 5.88 percent of legumes are lost
between and retail and consumer level. These losses were attributed to
the retail stage and therefore, input to the processing stage was set to
1.0625 kg of pulses.
The transportation distance for processed and packaged pulses de-
pends on locations of processing plants and retail stores, and regional
consumer demand for pulses. Transportation data specic to pulses were
not available. However, according to the Bureau of Transportation
Statistics (2018) food manufacturing industry transported the food
products to an average distance of 452 km. This transportation distance
was adopted for processed and packaged pulses, with trucks as the pri-
mary mode of transportation.
Data for electricity consumption by retail stores were obtained from
2017 Annual Retail Trade Survey (ARTS) (U.S. Census Bureau, 2017).
The total cost of electricity purchased by grocery stores was $5594
million in 2017. In the same year, average annual retail price of elec-
tricity for commercial sector was 10.66 cents per kWh (U.S. EIA, 2020a).
These data were used to estimate electricity consumption by grocery
stores in kWh in 2017. However, these estimates represented electricity
consumption by all grocery stores in the United States. The electricity
consumption was allocated to a kilogram of pulses stocked in the grocery
store using allocation based on shelf space occupied by a product and
per capita loss-adjusted availability of legumes at retail stores. Dry beans
occupy about 0.06% of consumer facing shelf space area at a super-
market (Willlard, 2016). In the absence of more granular data, this es-
timate was adopted to allocate retail stage burdens to all pulses. The
total mass of pulses sold by the retail sector was estimated using per
capita loss-adjusted availability of pulses at retail sector (5.40 kg/year)
and 2018 estimate of US population (327 million) (U.S. Census Bureau,
2018; USDA Economic Research Service, 2019). The average of land
occupation for superstore, neighborhood markets, and warehouse clubs
was 11,179 m
2
(Walmart Inc., 2019), which was allocated to a kilogram
of pulses using the same allocation factor estimated for electricity use.
2.2.4. Consumer stage
The consumer stage is the last stage in the cradle-to-grave LCA
model. It included purchase of pulses from retail stores, transportation
for grocery shopping, cooking, and consumption of pulses, and associ-
ated waste to landll. The reference ow of the consumer stage on the
dry basis was 56 g of dry bean, 58 g of chickpea, and 60 g each of eld
pea and lentil cooked and consumed at US household. The reference
ow represented average weekly consumption of pulses in the United
States (HHS and USDA, 2015). Accounting for an estimated 10% plate
wastage (USDA Economic Research Service, 2019) in the form of un-
eaten cooked pulses, the quantity purchased from retail was 62, 54, 66,
and 66 g for dry bean, chickpea, eld pea, and lentil, respectively.
Transportation at the consumer stage involved passenger car trans-
portation for grocery shopping. In the United States, the average dis-
tance to a grocery store in 2015 was 3.77 km (USDA Economic Research
Service, 2015). This included distances for average US households (3.45
km), SNAP recipients (3.16 km), and food insecure and WIC households
(4.70 km). However, it was reported in the same USDA study that
consumers often travelled to their preferred grocery store, often farther
than the closest one. Therefore, average distance of 5.52 km (average US
household- 6.10 km, SNAP participants- 5.41 km, food insecure house-
holds and WIC 5.07 km) was used in the model for grocery shopping.
Consumption of pulses at the consumer stage varied by the pulse
variety. The loss-adjusted per capita availability of dry beans and dry
peas and lentils at consumer level was 2.90 and 1.65 kg per year
respectively (USDA Economic Research Service, 2019). Per capita
loss-adjusted availability of dry peas and lentils was disaggregated into
chickpeas, lentils and eld peas based on proportion of these varieties in
total domestic availability of chickpea, lentil, and eld pea
(Tables S1–1). The burden of transportation was allocated to each pulse
variety (Tables S1–1) using percentage of total household expenditure
on chickpeas, lentils, eld peas, and dry beans estimated using average
2017 national average retail price for dry beans and household con-
sumption of each pulse variety (U.S. Bureau of Labor Statistics, 2018; U.
S. Census Bureau, 2019; USDA Economic Research Service, 2019). The
retail price was available only for dry bean, which was adopted for other
three pulse varieties.
For the base case scenario, it was assumed that cooking pulses
involved boiling and simmering pulses in an open vessel on electric
stove (OVC). The electricity consumption and water requirements for
cooking depend on pulse variety (USA Dry Pea & Lentil Council, 2019).
Dry beans and chickpeas require soaking which reduces the cooking
time and consequently electricity consumption. Pulses such as lentils
and eld peas can be cooked without soaking. The water requirement for
soaking and cooking and cooking time are provided in Tables S1–2. Data
provided by USA Dry Pea & Lentil Council (2019) included volumetric
measurements of pulses and water. These were converted to mass
measurements using density of pulses and water.
The base scenario in the study was open vessel cooking (section
2.2.4.1), which involved boiling and simmering pulses in an open vessel
on an electric stove. Two other methods of cooking pulses were evalu-
ated in this study, representing two alternative scenarios. These were
cooking in stovetop pressure cooker and in electric pressure cooker
(section 2.2.4.2). The LCI for the consumer stage, including the differ-
ences between study scenarios, is provided in Table 2.
2.2.4.1. Open vessel cooking (OVC). The total cooking time in OVC
included time required to bring the water to boil and simmering time
specic to the pulse variety. In OVC scenario, the energy required to
bring the water to boiling point was estimated using Eq. (1). On an
average the household electric stove draws between 1200 and 3000 W of
power (Direct Energy, 2019). An average power of 2100 W (2100 J/s)
was used to determine time required to bring the water to boiling from
an initial temperature of 25 ◦C. Electricity consumption (kWh) to fully
cook pulses was estimated assuming 20% of average cooking range
power requirement (simmering setting) and simmering times provided
in Tables S1–2. Cooking efciency of 39% for electric coil, estimated as
the ratio of energy transferred to water and energy input, was used to
account for specic heat capacities of water and vessel and radiative
energy losses (Karunanithy and Shafer, 2016).
Electricity and water consumption at the consumer stage also
included dishwashing. A typical dishwasher in a US household
consumed between 270 and 307 kWh of electricity per year and between
13 and 19 L of water per cycle (Appliance Standard Awareness Project,
2017). The dishwasher electricity consumption per cycle was estimated
assuming one cycle per day. This electricity and water consumption
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
6
were allocated to pulse varieties using the economic allocation provided
in Tables S1–1.
Q=m×c×(Tf−Ti)(1)
Where.
Q =energy required to raise temperature of water (J).
m =mass of water (g).
C =specic heat of water (J/g-◦C).
T
f
=nal temperature of water (◦C).
T
i
=initial temperature of water (◦C).
2.2.4.2. Pressure cooking (Stovetop, SPC and Electric, EPC). A pressure-
cooking scenario was evaluated to estimate the impact of cooking
method on sustainability metrics. Pressure cooking substantially reduces
cooking time, consequently reducing cooking energy use. However,
besides cooking time, energy savings also vary with the type of pressure
cooker (stovetop or electric) and related energy losses. The heating
components of electric pressure cookers are insulated making them
more energy efcient than stovetop pressure cookers (Reynolds et al.,
2018). These differences were captured by creating scenarios for sto-
vetop (SPC) and electric (EPC) pressure cookers. It was assumed that
temperature control on the cooking range was set to the medium heat
setting (50% of average power requirement of electric cooking rage) for
stovetop pressure cooker with the cooking efciency similar to OVC. For
electric pressure cooker, on the other hand, energy efciency of 95%
was assumed between heating element and wall power outlet with an
average power consumption of 1071 W. The power consumption of
electric pressure cooker was estimated from specications provided by
Instant Brands Inc (2020a). Data for cooking time of pulses were ob-
tained from FastCooking (2019) and Hawkins Ventura (2003) for SPC
(Tables S1–3) and from Instant Brands Inc (2020b) for EPC
(Tables S1–4). It was assumed that the ratio between volume of cooking
water and pulses reported by Hawkins Ventura (2003) was independent
of pressure cooker type. The cooking time varied with pulse variety and
is substantially reduced if pulses were soaked prior to cooking. Similar to
OVC, it was assumed that only chickpea and dry bean were soaked prior
to cooking, to ensure that only the inuence of cooking method was
evaluated. The amount of water required for soaking chickpeas and dry
beans was adopted from the OVC scenario. Electricity consumption for
the pressure-cooking scenario was estimated using the same method
used in the OVC scenario. However, pressure cooking did not require
bringing the water to a boil before adding the pulses. Therefore, cooking
time in this scenario reected time required to cook pulses that were
started with room temperature water.
2.3. Uncertainty analysis
Data used for life cycle impact analysis is based on mean estimates of
parameter values which carry uncertainty that could alter the conclu-
sions. Therefore, an uncertainty analysis was performed using Monte
Carlo Simulations (MCS) to increase condence in the interpretation of
results. Data for most parameters in the model included means and
range. Therefore, uncertainty for these foreground model parameters
was dened as a triangular distribution, with the exception of crop yield.
A normal distribution was dened for the crop yield using standard
deviation estimated using USDA-NASS data (USDA National Agricul-
tural Statistics Services, 2017). Background processes from EcoInvent
database were adopted in the model without changing their uncertainty
characteristics. Uncertainty in impact characterization factors is not
included in the evaluation, therefore this assessment represents a lower
bound on uncertainty of the results.
2.4. Sensitivity to the reference ow
It was discovered during the initial runs of the cradle-to-grave model
that the environmental impact categories were highly sensitive to the
mass of pulses cooked in a batch. In OVC and both pressure cooking
scenarios the reference ow of 60 g represented average weekly con-
sumption of pulses. The inuence of consumer stage reference ow on
environmental impact categories was assessed by changing this refer-
ence ow to 1 kg of pulses while maintaining cooking method to open
vessel cooking (OVC-RF1). This reference ow represented cooking one
large batch of pulses to be consumed over approximately 4 months at
current weekly consumption rate of 60 g. However, this required
freezing cooked pulses and reheating them before consumption, most
likely using a microwave. Annual household refrigerator and microwave
electricity consumption obtained from U.S. EIA (2015) was attributed to
each pulse variety using economic allocation factors used for passenger
travel for grocery (Tables S1–1). Safe storage period of 2–3 months
estimated for frozen soups and stews (FoodSafety.gov, 2021) was
adopted for pulses to estimate increased food wastage. Assuming that
four-month supply of cooked pulses can be safely stored only for
maximum of 3 months, food wastage of pulses was increased to 25% for
this scenario. However, it was assumed that pulses were stored for four
months before they were discarded. Therefore, refrigerator and micro-
wave electricity consumption were estimated assuming four-month
refrigerator use and 12 instances of microwave use (Table 3).
3. Results and discussion
Cradle-to-grave environmental impact of pulses was assessed in this
Table 2
Life cycle inventory for consumer stage for open vessel and pressure-cooking scenarios.
Pulse variety Inputs Reference Flow, kg
Mass of pulses, kg Water (L) Electricity, kWh Grocery Travel, km
Cooking Soaking Dishwasher Cooking Dishwasher
Open Vessel Cooking (OVC)
Chickpea 0.064 0.225 0.225 0.011 1.628 0.001 0.004 0.058
Dry bean 0.062 0.150 0.225 0.057 1.620 0.004 0.020 0.056
Field pea 0.066 0.150 – 0.014 0.685 0.001 0.005 0.060
Lentil 0.066 0.188 – 0.007 0.355 0.0004 0.002 0.060
Pressure Cooking, Stovetop Pressure Cooker (SPC)
Chickpea 0.064 0.225 0.225 0.011 0.588 0.001 0.004 0.058
Dry bean 0.062 0.225 0.225 0.057 0.303 0.004 0.020 0.056
Field pea 0.066 0.225 – 0.014 0.648 0.001 0.005 0.060
Lentil 0.066 0.225 – 0.007 0.327 0.0004 0.002 0.060
Pressure Cooking, Electric Pressure Cooker (EPC)
Chickpea 0.064 0.225 0.225 0.011 0.223 0.001 0.004 0.058
Dry bean 0.062 0.225 0.225 0.057 0.117 0.004 0.020 0.056
Field pea 0.066 0.225 – 0.014 0.366 0.001 0.005 0.060
Lentil 0.066 0.225 – 0.007 0.094 0.0004 0.002 0.060
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
7
study The base scenario, OVC, included cooking pulses on an electric
stove in an open vessel. SPC and EPC scenarios evaluated the impact of
cooking method while the OVC-RF1 scenario estimated the impact of
mass of pulses cooked per batch on environmental impact of pulses. We
also evaluated inter-varietal variability resulting from the differences in
crop production practices and time required to cook the pulses.
3.1. Open vessel cooking
3.1.1. Impact category scores
The GWP for 60 g (dry basis) of pulses consumed in a US household
was 1.26, 1.34, 0.53, and 0.31 kg CO
2
e for chickpeas, dry beans, eld
peas, and lentils, respectively (Fig. 2). Fossil fuel consumption in ReCiPe
2016 is reported as fossil fuel scarcity and expressed as kg oil eq. The
FRS ranged between 0.08 and 0.34 kg oil eq per 60 g of pulse crop
(chickpeas: 0.32, dry beans: 0.34, eld peas: 0.14, lentils: 0.08 kg oil eq).
The LU measured in m
2
a crop eq was 0.69 for chickpeas, 0.63 for dry
beans, 0.58 for eld peas, and 0.82 for lentils. Throughout the cradle-to-
grave processes, WC was estimated at 7.41, 7.75, 3.22, and 2.12 L for
chickpeas, dry beans, eld peas, and lentils, respectively. The FE,
resulting primarily from phosphorus fertilizer application, was 1.37,
1.43, 0.59, and 0.36 g P eq for chickpea, dry bean, eld pea, and lentil,
respectively. ME ranged between 0.021 g N eq for lentil and 0.092 g N eq
for dry bean. The ME for chickpea and eld pea was 0.088 and 0.037 g N
eq, respectively.
3.1.2. Inter-varietal variability and contribution analysis
Inter-varietal variability within environmental impact categories
was associated with factors such as crop management practices, fertil-
izer application rates, crop yield, and cooking time. With the exception
of LU, the greatest contribution to all other impact categories resulted
from the consumer stage, which involved purchasing and cooking pulses
and plate waste. The consumer stage contributed at least 83, 81, 76, 75,
and 87 percent of total impact for GWP, FRS, WC, FE, and ME,
respectively.
3.1.2.1. Global warming potential and fossil resource scarcity. The
contribution of the consumer stage to GWP and FRS varied with pulse
variety. However, for both impact categories contribution from con-
sumer stage was the greatest for chickpea and the least for lentil (Fig. 3).
Greater contribution from the consumer stage to these impact categories
as well as inter-varietal variability in impact category scores could be
attributed to electricity consumed during cooking. Electricity was uti-
lized at the consumer stage primarily for cooking and for running the
dishwasher. However, cooking contributed to approximately 99% of
total electricity consumption at the consumer stage for which, the
driving factor was cooking time.
Cooking pulses in open vessels requires brining water to boil fol-
lowed by simmering until pulses are cooked through. The time required
to boil water (range: 57 s to 1 min 26 s) and consequently, associated
electricity consumption did not vary substantially. This was because
only small quantities of pulses were cooked, which required mass of
water that ranged between 150 g and 225 g. On the contrary, post-boil
simmering time varied between 19 min for lentil to 90 min for chickpea
and dry bean. This difference in cooking times resulted in the propor-
tional inter-varietal variability in electricity consumption, which was
reected in fossil fuel scarcity scores.
Electricity production in the United States relies heavily on fossil
fuels, primarily natural gas, and coal. In fact, about 63% of total elec-
tricity generated in the US in 2019 was produced using fossil fuels (U.S.
EIA, 2020b). Upstream emissions associated with electricity production
were responsible for increasing overall GWP impact scores of pulses and
contribution of consumer stage. For example, approximately 94% of
total GWP of chickpea was associated with electricity production from
all sources, while at least 78% of GWP resulted from electricity pro-
duction that relied on coal and natural gas.
The GWP and FRS scores of pulses followed a general trend similar to
electricity consumption at the consumer stage. However, a slight
Table 3
Electricity and water consumption at consumer stage for the reference ow of 1 kg.
Pulse variety Inputs
Mass of pulse, kg Water, kg Electricity kWh
Cooking Soaking Dishwasher Cooking Dishwasher Refrigerator Microwave
Chickpea 1.333 4.662 4.662 0.132 2.129 0.008 0.069 0.001
Dry bean 1.333 3.231 4.847 0.686 1.968 0.042 0.358 0.006
Field pea 1.333 3.019 – 0.171 1.009 0.010 0.089 0.001
Lentil 1.333 3.773 – 0.087 0.760 0.005 0.046 0.001
Fig. 2. Environmental impact of 60 g of pulses estimated for OVC scenario for following impact categories (a) GWP, (b) LU. Graphs for other impact categories are
presented in Figs. S1–1 in the Supplementary Material.
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
8
anomaly was observed in GWP and FRS scores of chickpeas and dry
bean. The GWP and FRS of dry bean was 6.4 and 6.3 percent greater
compared to chickpea when the electricity consumption at the consumer
stage for these were comparable. Greater GWP and FRS observed for dry
bean was attributed to marginally greater contribution from the cradle-
to-farm stage to these impact categories. Unlike other pulse varieties,
dry beans were grown using conventional farming methods. Increased
fossil fuel use required for conventional farming and related greenhouse
gas emissions marginally increased the contribution of farming stage to
these impact categories and overall impact scores.
3.1.2.2. Land use. In contrast to other impact categories, the primary
contributor to the LU was crop production. The factor responsible for
this contribution as well as for the inter-varietal variability in LU scores
was crop yield. The LU score was inversely related to the yield because
greater yield increased resource utilization efciency at the farm. The
crop yield varied with pulse variety ranging between 1342 kg ha
−1
for
lentil and 2029 kg ha
−1
for eld pea, respectively. Consequently, the LU
was the greatest for lentil and the least for eld pea (Fig. 2).
3.1.2.3. Water consumption. The greatest contribution to total WC came
from the consumer stage, amounting to 76% of total WC for lentil and
more than 88% of total WC for other pulse varieties. Similar to GWP and
FRS, electricity consumption at the consumer stage and associated up-
stream water use were responsible for the greater contribution from
consumer stage. For example, water use related to electricity con-
sumption at the consumer stage accounted for approximately 85% of
total water use. Only 7.3 and 7.7 percent of water use was associated
with cooking and dishwashing, and other upstream processes, respec-
tively. The electricity and water use at the consumer stage also inu-
enced inter-varietal variability in WC scores. Chickpea and dry bean
required longer cooking time and needed water for soaking which
increased their WC compared to eld pea and lentil.
3.1.2.4. Freshwater and marine eutrophication. The contribution of the
consumer stage to the total impact category scores ranged between 75
and 94 percent for FE and between 87 and 97 percent for ME. For both
impact categories, contribution from the consumer stage was the least
for lentil and the largest for chickpea. A greater contribution of con-
sumer stage was primarily because of electricity use at the consumer
stage and associated upstream emissions of NOx and phosphate
compounds.
However, phosphorus and nitrogen fertilizer application rates and
crop yield at the farming stage inuenced total eutrophication impact
scores as well as contributions from the farming stage. Dry bean, for
example, required more nitrogen fertilizers compared to other pulse
varieties. This resulted in the largest contribution to ME (0.0029 g N eq
per FU) scores for dry bean from the farming stage. In contrast, the ME
scores of other three pulse varieties at the farmgate were lower than the
dry bean because of lower nitrogen demand. However, despite identical
phosphorus and nitrogen application rates, lower crop yield of lentil
increased their FE (0.088 g P eq per FU) and ME (0.0026 g N eq per FU)
scores compared to chickpea (FE: 0.073 g P eq per FU, ME: 0.0022 g N eq
per FU) and eld pea (FE: 0.063 g P eq per FU, ME: 0.0022 g N eq per FU)
at the farmgate.
3.2. Pressure cooking
Switching cooking method from open vessel cooking to pressure
cooking reduced GWP of pulse varieties by 5–86 percent. FRS by 5–85
percent, WC by 1–78 percent, FE by 5–86 percent, and ME by 5–88
(Table 4). The lower impact scores observed for pressure cooking sce-
narios were attributed to shorter cooking times and associated energy
savings. However, shorter cooking times did not always result in pro-
portional decrease in the electricity consumption, especially for SPC
scenario. Despite 62% reduction in the cooking time for eld peas in SPC
(OVC- 39 min, SPC- 15 min), the electricity consumption decreased only
by 5%. This discrepancy was primarily because of assumptions made
Fig. 3. Results of contribution analysis for (a) chickpea, (b) dry bean, (c) eld pea, and (d) lentil, for OVC scenario.
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
9
regarding heat control setting on a cooking range, which was assumed to
use 20% (low heat) and 50% (medium heat) of available energy for OVC
and SPC, respectively. Therefore, more energy was required throughout
the cooking period to achieve shorter cooking times in SPC, which
decreased the magnitude of savings in electricity consumption. Never-
theless, SPC reduced the environmental impact scores of pulses by at
least 5% across all impact categories, excluding LU.
The EPC resulted in the lowest impact scores among all cooking
methods across all pulse varieties and impact categories, excluding LU.
This was attributed to lower energy demand and improved energy ef-
ciency of pressure cookers compared to OVC and SPC. The electric
pressure cookers required an average 1071 W power compared to 2100
W required for stovetop cooking. Moreover, the energy efciency of
electric pressure cookers was at least 95% resulting in more efcient use
of electricity. This lowered electricity consumption for cooking and
associated upstream emissions. Electricity consumption for dry beans
and chickpeas in EPC, for instance, was at least 61% lower compared to
SPC, whereas cooking time remained identical (Tables S1–3).
Overall, the greatest reduction in impact category scores, compared
to OVC, was observed for dry bean, followed by chickpea, lentil, and
eld pea. While the magnitude of this change was greater for EPC
compared to SPC, an identical trend was observed for both scenarios.
The magnitude of change in impact category scores compared to OVC
depended on the decrease in electricity consumption required for
cooking pulses. Compared to OVC, pressure cooking methods offered the
greatest savings in electricity consumption for dry bean (SPC- 81%, EPC-
93%), followed by chickpea (SPC- 64%, EPC- 85%), lentil (SPC- 8%,
EPC- 74%), and eld pea (SPC- 5%, EPC- 47%), which was also reected
in their environmental impact score across all impact categories,
excluding LU. The largest contributor to the LU was farming stage,
where crop yield was the primary driving factor. Because cooking
methods only inuenced electricity consumption, only a small to no
change in LU was observed for SPC and EPC (Table 4).
The pressure-cooking method expedited cooking of all varieties of
pulses, reduced environmental impact of pulses, and marginally
decreased the contribution of the consumer stage to the overall impact.
However, the contribution of the consumer stage still remained high
(Fig. S1-2 and S1-3). The consumer stage in pressure cooking scenario
contributed between 52 (EPC, dry bean) and 92 (SPC, eld pea) percent
of total GWP (compared to 83 to 95 percent for OVC) and between 52
(EPC, dry bean) and 91 (SPC, eld pea) percent of total FRS (compared
to 81 and 94 percent for OVC). This was primarily because in spite of
5–93 percent reduction in total cooking-related electricity consumption,
the upstream emissions associated with electricity production still
dominated total emissions from the pulse supply chain, increasing the
contribution of consumer stage for SPC and EPC.
3.3. Uncertainty analysis
Uncertainty analysis was performed to evaluate the robustness of
conclusions regarding differences in the environmental impact category
scores of pulse varieties and cooking methods. The results of MCS for
GWP and FRS are presented in Fig. 4. Results for other impact categories
are presented in (Figs. S1–4). Differences in GWP, FRS, FE, and ME
scores of pulse varieties were more prominent for OVC, compared to SPC
and EPC. For OVC scenario, there was more overlap of boxes and
whiskers for chickpea and dry bean compared to other two pulse vari-
eties suggesting a higher probability that impact scores of chickpea and
dry bean for these four impact categories were comparable to each other
but greater than eld pea and lentil. Within SPC and EPC, the overlap of
box and whiskers for chickpea, dry bean, and eld pea indicated that
only small to no differences in GWP, FRS, FE, and ME scores. This sug-
gested that inter-varietal variability between chickpea, dry bean, and
eld pea observed within each pressure-cooking scenario was statisti-
cally non-signicant. However, there existed a greater probability of
lentil having the lowest impact scores across these four impact cate-
gories for all three cooking scenarios. The uncertainty analysis also
suggested a probability that LU of chickpea, dry bean, and eld pea was
comparable to each other while that of lentil was marginally greater,
and a probability that the differences in water use scores of these pulse
Table 4
Environmental impact for 60 g of pulses cooked in stove top and electric pressure cooker.
Impact category SPC EPC
Chickpea Dry bean Field pea Lentil Chickpea Dry bean Field pea Lentil
Global warming potential, kg CO
2
eq 0.50 0.33 0.50 0.29 0.23 0.19 0.30 0.12
Fossil resource scarcity, kg oil eq 0.13 0.09 0.13 0.07 0.06 0.05 0.08 0.03
Land use, m
2
a 0.69 0.62 0.59 0.82 0.69 0.62 0.58 0.82
Water consumption, L 3.31 2.44 3.18 2.06 1.86 1.68 2.09 1.16
Freshwater eutrophication, g P eq 0.55 0.35 0.56 0.34 0.26 0.20 0.35 0.16
Marine Eutrophication, kg N eq 0.034 0.022 0.035 0.019 0.016 0.012 0.021 0.008
Fig. 4. Results of uncertainty analysis for (a) GWP and (b) fossil resource scarcity for OVC, SPC, and EPC. The results for other impact categories are presented in the
supplementary material (Figs. S1–4).
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
10
varieties were statistically non-signicant.
As expected, the absence of differences in LU in relation to cooking
method was conrmed by the uncertainty analysis. Similarly, the un-
certainty analysis indicated that the difference in mean WC scores in
relation to cooking method were not statistically signicant. For GWP,
FRS, FE, and ME (impact categories discussed hereafter), the uncertainty
analysis conrmed the mixed inuence of cooking method on environ-
mental impact. For chickpea and dry bean SPC decreased environmental
impact scores across these four impact categories compared to OVC.
However, the inuence of pressure cooker type on impact categories was
more pronounced for chickpea compared to dry bean. This suggested a
greater probability for chickpea (and a lower probability for dry bean)
that EPC signicantly decreased impact category scores, when
compared to SPC. The uncertainty analysis also indicated that for eld
pea and lentil, impact category scores of OVC and SPC scenarios were
comparable, but a greater probability existed that EPC lowered envi-
ronmental impact of these two pulse varieties.
3.4. Sensitivity to consumer stage reference ow (OVC-RF1)
Changing consumer stage reference ow to 1 kg of pulses substan-
tially reduced environmental impact of pulses across all impact cate-
gories (excluding land use) even after accounting for increased food
waste and electricity consumption. Estimated GWP in this scenario for
60 g of pulses was 0.18, 0.21, 0.10, and 0.10 kg CO
2
e for chickpea, dry
bean, eld pea, and lentil, respectively. The GWP in this scenario was
approximately, 86 (chickpea), 84 (dry bean), 82 (eld pea), and 68
(lentil) percent lower than OVC scenario (Fig. 5). Similar decrease in
scores was also observed for FRS (67–86 percent), WC (49–77 percent),
FE (60–85 percent), and ME (72–87 percent). The primary reason of this
decrease in environmental impact of pulses was lower electricity con-
sumption. Increasing the reference ow to 1 kg increased cooking
electricity consumption as more energy was necessary to boil larger
mass of water. However, the simmering time remained unaffected
resulting in very small change in electricity consumption that ranged
between 0.32 and 0.50 kWh. Moreover, because larger quantity of
pulses was cooked in a single batch, total electricity consumption,
normalized for mass of pulses cooked, remained between 0.61 and 1.78
kWh/kg of pulses for OVC-RF1 as opposed to 5.36 to 26.21 kWh/kg of
pulses for the OVC. This reduction in total electricity consumption also
reduced upstream emissions, resulting in lower environmental impact
scores across all impact categories, excluding land use. The land use in
OVC-RF increased by 18–20% compared to OVC. However, the Monte
Carlo Simulations indicated that this change in land use was not statis-
tically signicant (Fig. 5, Figs. S1–5).
A trade-off between the contribution from consumer and farming
stage was also observed for this scenario (Figs. S1–6). Cooking larger
quantity of pulses in a single batch decreased the contribution from
consumer stage to GWP by 42–48 percentage points compared to OVC. It
also increased the contribution from the farming stage to overall GWP,
which ranged between 28 and 57 percent for OVC-RF1 compared to 3 to
15 percent observed for OVC. A similar trend was also observed for FRS,
WC, FE, and ME. A small increase, compared to OVC, in contribution
from processing and retail stages to overall GWP (2–14 percentage
points) and FRS (2–13 percentage points) scores was also observed.
Similar to pulses, the inuence of batch size and cooking-related
energy demand on GWP and FRS was also observed for potatoes and
bread. Parajuli et al. (2021) reported that the contribution from the
consumer stage for at-home consumption of 1 kg of fresh potatoes was
47% of total cradle-to-grave GWP, primarily because of frying in vege-
table oil. For 1 kg of frozen potato fries this contribution was 38%. In
case of bread, electricity consumption for refrigerated storage and
toasting of bread at the consumer stage contributed as much as 25% of
total GWP in a cradle-to-grave analysis (Espinoza-Orias et al., 2011).
The most energy-intensive process in bread manufacturing was baking,
which accounted for an average of 64% of total energy consumption in
the bread supply chain (Braschkat et al., 2003). Moreover, the energy
consumption was three times greater for home baking compared to in-
dustrial baking, which also increased the GWP of home-baked bread
(Braschkat et al., 2003).
3.5. Cradle-to-farmgate impact analysis
The contribution of the farming stage to the most impact categories
was lower compared to the consumer stage. However, cradle-to-
farmgate impact assessment and contribution analysis can provide in-
sights into inuence of farming activities on sustainability of the pulses.
It can also facilitate easy comparison between pulses and other crops in
term of their environmental impact. The GWP of pulses at the farmgate
ranged between 0.32 and 0.61 kg CO
2
e per kg of harvested pulses
Fig. 5. Results of Monte Carlo Simulations indicating the inuence of consumer stage reference ow on GWP and LU of pulses. The results for other impact categories
are provided in Figs. S1–5.
Table 5
Environmental impact associated with production of pulses for 1 kg of harvested
pulses at the farmgate.
Impact Category Chickpea Dry bean Field pea Lentil
Global warming potential, kg CO
2
eq
0.39 0.61 0.32 0.45
Fossil resource scarcity, kg oil eq 0.10 0.16 0.08 0.12
Land use, m
2
a 6.31 5.66 5.40 7.80
Water consumption, L 3.65 4.99 3.09 4.40
Freshwater eutrophication, g P eq 0.69 0.79 0.59 0.84
Marine Eutrophication, g N eq 0.021 0.027 0.018 0.025
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
11
(Table 5), with the greatest GWP observed for dry beans, followed by
lentil, chickpea, and eld pea. A similar trend was also observed for FRS,
WC, FE, and FE (Figs. S1–7). The trend for LU was slightly different. The
greatest land use score was observed for lentil, followed by chickpea, dry
bean, and eld pea (Table 5). Primary contributors to the environmental
impact of pulses (excluding LU) were tillage operations, emissions
associated with fertilizer and pesticide manufacturing, production of
seeds and inoculant, and eld emissions related to fertilizer application
(Fig. 6.).
The environmental impact scores of pulses as well as the inter-
varietal variability were primarily inuenced by crop yield, tillage
practices, and fertilizer application rates. Conventional tillage and
higher nitrogen demand of dry beans increased fossil fuel consumption
as well as eld emissions. These factors increased the environmental
impact of dry bean even when the yield of dry bean was greater than
chickpea and lentil. Contrarily, despite identical tillage operations and
fertilizer and pesticide application rates, differences in yield resulted in
the lowest environmental score for eld pea compared to chickpea and
lentil. The inuence of crop yield was also evident in LU scores of pulses
which carried inverse relationship with the yield.
The GWP of pulses estimated in this study was somewhat greater
than the values reported by Gustafson (2017) for pulses grown in the
United States. For 1 kg of harvested pulses Gustafson (2017) estimated
the GWP of 0.31 and 0.26 kg CO
2
e for irrigated and dryland pulses,
respectively. Greater GWP observed in this study could be attributed to
the differences in yield, tillage practices, and use of synthetic fertilizers.
Crop yields used in this study ranged between 1,342 and 2,029 kg ha
−1
compared to 2,030 kg ha
−1
used by Gustafson (2017) for dryland pulses.
While the mean fertilizer application rate was not reported by Gustafson
(2017), an example data provided by the author reported that fertilizers
were not applied to dryland pulses. On the contrary, we assumed used of
nitrogen, phosphorus, and potassium fertilizers in the production of all
varieties of pulses studied. We also modeled dry beans with conven-
tional tillage practices and greater nitrogen fertilizer application rate
compared to other pulse varieties. These differences in crop manage-
ment practices between two studies may have contributed to the greater
GWP observed in this study.
4. Conclusion
The GWP of pulses ranged between 0.12 and 1.34 kg CO
2
e for 60 g of
pulses produced and consumed in the United States. Impact category
scores per functional unit for other impact categories was 0.03–0.34 kg
oil eq for FRS, 0.58–0.82 m
2
a for LU, 1.17–7.75 L for WC, 0.16–1.43 g P
eq for FE, and 0.007–0.092 kg N eq for ME. Overall, the environmental
impact of pulses varied with pulse variety, cooking method, and mass of
pulses cooked per batch. However, the consumer stage dominated the
environmental impacts of pulses for all pulse varieties and scenarios.
Electricity consumed during cooking was the principal driving factor for
cradle-to-grave impact of pulses and for contribution of consumer stage.
Overall, the study identied cooking time and energy use efciency as
two parameters that inuenced the electricity consumption at the con-
sumer stage. The direct proportionality of electricity consumption with
cooking time and inverse proportionality with energy use efciency
were evident from the results of three cooking method scenarios, where
OVC (longer cooking time and lower energy use efciency) resulted in
the greatest environmental impact and EPC (shorter cooking time and
greater energy use efciency) resulted in the least. The benets of
shorter cooking time in SPC were offset by lower energy use efciency
resulting in statistically non-signicant change in the environmental
impacts for eld pea and lentil as compared to OVC.
The study also identied the inuence of cooking mass per batch on
overall sustainability of the pulses. Even for the open vessel cooking
method, increasing the batch size signicantly decreased the environ-
mental impact of pulses across all impact categories, excluding LU,
despite increased food losses and added electricity demand for
refrigeration and microwave use. This was primarily because larger
batch size increased the resource utilization efciency, as larger mass of
pulses was cooked with only marginal increase in total cooking time.
This substantially decreased electricity consumption per kilogram of
pulses. However, the environmental impact of pulses in OVC-RF1 sce-
nario was comparable to EPC for most impact categories.
Overall, the consumer stage, specically electricity consumed during
cooking, was identied as the hotspot in the production and consump-
tion of pulses. Considering cooking pulses in electric pressure cooker or
cooking larger mass of pulses per batch resulted in statistically signi-
cant reductions in environmental impact category scores, these methods
can be adopted to ensure sustainable consumption of pulses.
Author credit
Prathamesh A. Bandekar: Formal analysis, Investigation, Method-
ology, Writing – Original Draft, Writing – Review & Editing; Ben Put-
man: Conceptualization, Methodology, Writing – Review & Editing;
Greg Thoma: Conceptualization, Methodology, Validation, Writing –
Review & Editing, Supervision, Funding acquisition; Marty Matlock:
Conceptualization, Methodology, Validation, Writing -Review & Edit-
ing, Supervision.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Acknowledgement
This work was supported by the Agricultural Research Service (ARS),
U.S. Department of Agriculture (USDA) [Grant number 58-3060-8-030]
under Non-Assistance Cooperative Agreement (NACA).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.jenvman.2021.114062.
References
Appliance Standard Awareness Project, 2017. Dishwasher, 8.7.19. https://appliance-stan
dards.org/product/dishwashers.
Bedard-Haughn, A., Comeau, L.-P., Sangster, A., 2013. Gross nitrogen mineralization in
pulse-crop rotations on the Northern Great Plains. Nutrient Cycl. Agroecosyst. 95,
159–174. https://doi.org/10.1007/s10705-013-9555-z.
Fig. 6. Contribution from various farm activities to the GWP associated with
production of pulses for 1 kg of pulses harvested at the farmgate. The graphs for
other impact categories are provided in Figs. S1–7.
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
12
Bond, J.K., 2019. Usda ERS - U.S. Exports of chickpeas, lentils, and dry peas have
dropped sharply due to a steep decline in shipments to India. URL 09.28.21. https
://www.ers.usda.gov/amber-waves/2019/september/us-exports-of-chickpeas-le
ntils-and-dry-peas-have-dropped-sharply-due-to-a-steep-decline-in-shipments-to-ind
ia/.
Braschkat, J., Patyk, A., Quirin, M.J., Reinhardt, G.A., 2003. Life cycle assessment of
bread production - a comparison of eight different scenarios. In: Proceedings from
4th International Conference. Presented at the Life Cycle Assessment in the Agri-
Food Sector, Ministry of Food, Agriculture and Fisheries. Danish Institute of
Agricultural Sciences, Bygholm, Denmark, pp. 9–16.
Brouwer, B., Atterberry, K., Miles, C.A., 2015. Commercial Dry Bean Production in
Western Washington State (No. EM092E). Washington State University Extension.
Bühler, 2019a. Destoner. Destoner, 6.20.19. https://www.buhlergroup.com/content
/buhlergroup/global/en/products/destoner.html.
Bühler, 2019b. PULSROLL (TM) Pulse Huller. PULSROLL TM, 6.20.19. https://www.buh
lergroup.com/content/buhlergroup/global/en/products/pulsroll_huller.html.
Bühler, 2019c. Rotosort (TM) Drum Grader. Drum Grader Rotosort TM, 6.20.19. http
s://www.buhlergroup.com/content/buhlergroup/global/en/products/rotosort_dru
m_grader.html.
Bühler, 2019d. SORTEX(R) A optical sorter. Opt. Sorting Solut. Pulse Sorting, 6.20.19.
https://www.buhlergroup.com/content/buhlergroup/global/en/process-technolog
ies/Optical-Sorting/Pulse-sorting.html.
Bureau of Transportation Statistics, 2018. 2017 CFS Preliminary Data.
Chaudhary, A., Marinangeli, C., Tremorin, D., Mathys, A., 2018. Nutritional combined
greenhouse gas life cycle analysis for incorporating Canadian yellow pea into cereal-
based food products. Nutrients 10, 490. https://doi.org/10.3390/nu10040490.
Direct Energy, 2019. How Much Energy Does an Electric Oven and Stove Use?, 8.7.19.
https://www.directenergy.com/learning-center/energy-efciency/how-much-ener
gy-does-oven-and-electric-stove-use.
Dusenbury, M.P., Engel, R.E., Miller, P.R., Lemke, R.L., Wallander, R., 2008. Nitrous
oxide emissions from a northern Great Plains soil as inuenced by nitrogen
management and cropping systems. J. Environ. Qual. 37, 542–550. https://doi.org/
10.2134/jeq2006.0395.
Ebert, A., 2014. Potential of underutilized traditional vegetables and legume crops to
contribute to food and nutritional security, income and more sustainable production
systems. Sustainability 6, 319–335. https://doi.org/10.3390/su6010319.
Espinoza-Orias, N., Stichnothe, H., Azapagic, A., 2011. The carbon footprint of bread.
Int. J. Life Cycle Assess. 16, 351–365. https://doi.org/10.1007/s11367-011-0271-0.
FAO, 2016. Pulses - Nutritious Seeds for a Sustainable Future. Food and Agriculture
Organization of the United Nations.
FastCooking, 2019. The Ultimate Pressure Cooker Cooking Time Chart.Pdf, 10.17.19.
https://fastcooking.ca/pressure_cookers/cooking_times_pressure_cooker.ph
p#pulses.
Fernandez-Cornejo, J., 2004. The seed industry in U.S. Agriculture: an exploration of
data and information on crop seed markets, regulation, industry structure, and
Research and development (No. 786). USDA-Economic Research Service. Agriculture
Information Bulletin.
FoodSafetygov, 2021. Cold Food Storage Chart, 1.21.21. https://www.foodsafety.
gov/food-safety-charts/cold-food-storage-charts.
Franzen, D.W., 1998. Fertilizing Field Pea and Lentil (No. SF-725). North Dakota State
University, Fargo, ND.
Gustafson, D.I., 2017. Greenhouse gas emissions and irrigation water use in the
production of pulse crops in the United States. Cogent Food Agric 3, 1334750.
https://doi.org/10.1080/23311932.2017.1334750.
Halvorson, A.D., Bartolo, M.E., 2014. Nitrogen source and rate effects on irrigated corn
yields and nitrogen-use efciency. Agron. J. 106, 681–693. https://doi.org/
10.2134/agronj2013.0001.
Hasler, K., Br¨
oring, S., Omta, S.W.F., Olfs, H.-W., 2015. Life cycle assessment (LCA) of
different fertilizer product types. Eur. J. Agron. 69, 41–51. https://doi.org/10.1016/
j.eja.2015.06.001.
Hawkins Ventura, 2003. Pressure Cooker Instruction Manual with 21 Tested Recipes
(accessed 10.15.2019). https://www.hawkinscookers.com/Cookbooks/Ventura%
20IM.pdf.
Huijbregts, M.A.J., Steinmann, Z.J.N., Elshout, P.M.F., Stam, G., Verones, F., Vieira, M.,
Zijp, M., Hollander, A., van Zelm, R., 2017. ReCiPe 2016: a harmonised life cycle
impact assessment method at midpoint and endpoint level. Int. J. Life Cycle Assess.
22, 138–147. https://doi.org/10.1007/s11367-016-1246-y.
Instant Brands Inc, 2020a. Pressure Multicookers. Instant Pot, 11.10.20. https://instantp
ot.com/pressure-multicookers/.
Instant Brands Inc, 2020b. Instant Pot Pressure Cooking Time Table, 11.10.20. https://in
stantpot.com/wp-content/uploads/2020/03/EPC-Time-Table-2020-03-31.pdf.
IPCC, 2006. IPCC Guidelines for National Greenhouse Gas Inventories. Chapter 11: N2O
Emissions from Managed Soils, and CO2 Emissions from Lime and Urea Application.
Intergovernmental Panel on Climate Change (IPCC), Geneva, Switzerland.
ISO, 2006a. ISO 14040 - Environmental Management - Life Cycle Assessment - Principles
and Framework. International Organization for Standardization, Geneva,
Switzerland.
ISO, 2006b. ISO 14044 - Environmental Management - Life Cycle Assessment
-Requirements and Guidelines. International Organization for Standardization,
Geneva, Switzerland.
Kandel, E.H., Zollinger, R., Jenks, B., Knodel, J., Markell, S., Wunsch, M., Pasche, J.,
Nowatzki, J., Hellevang, K., Olson, F., 2013. Growing Lentil in North Dakota (No.
A1636). North Dakota State University, Fargo, ND.
Kandel, H., Keene, C., Markell, S., Pasche, J., Franzen, D.W., Knodel, J., Stefaniak, T.,
Jenks, B., Forster, S., 2018. Growing Chickpea in North Dakota (Extension Document
No. A1236 (Revised)). North Dakota State University, Fargo, ND.
Karunanithy, C., Shafer, K., 2016. Heat transfer characteristics and cooking efciency of
different sauce pans on various cooktops. Appl. Therm. Eng. 93, 1202–1215. https://
doi.org/10.1016/j.applthermaleng.2015.10.061.
Kim, S., Dale, B.E., 2008. Effects of nitrogen fertilizer application on greenhouse gas
emissions and economics of corn production. Environ. Sci. Technol. 42, 6028–6033.
https://doi.org/10.1021/es800630d.
Kim, S., Dale, B.E., Jenkins, R., 2009. Life cycle assessment of corn grain and corn stover
in the United States. Int. J. Life Cycle Assess. 14, 160–174. https://doi.org/10.1007/
s11367-008-0054-4.
Kulshreshtha, S., MacWilliam, S., Wismer, M., 2013. Economic and environmental
assessment of pulse rotations in Canadian prairies. In: International Farm
Management Congress Proceedings. Presented at the 19th International Farm
Management Congress, p. 8. Warsaw, Poland.
MacWilliam, S., Parker, D., Marinangeli, C.P.F., Tr´
emorin, D., 2018. A meta-analysis
approach to examining the greenhouse gas implications of including dry peas (Pisum
sativum L.) and lentils (Lens culinaris M.) in crop rotations in western Canada. Agric.
Syst. 166, 101–110. https://doi.org/10.1016/j.agsy.2018.07.016.
MacWilliam, S., Wismer, M., Kulshreshtha, S., 2015. Life Cycle and Economic Assessment
of Western Canadian Pulse Systems: Dry Pea versus Soybean Meal as A Source of
Protein in Swine Diets 21.
MacWilliam, S., Wismer, M., Kulshreshtha, S., 2014a. Life cycle and economic
assessment of Western Canadian pulse systems: the inclusion of pulses in crop
rotations. Agric. Syst. 123, 43–53. https://doi.org/10.1016/j.agsy.2013.08.009.
MacWilliam, S., Wismer, M., Kulshreshtha, S., 2014b. Life cycle and economic
assessment of Western Canadian pulse systems: the inclusion of pulses in crop
rotations. Agric. Syst. 123, 43–53. https://doi.org/10.1016/j.agsy.2013.08.009.
Miller, P., Gusak, M., Jaconis, E., Jones, C., 2019. Personal Communication.
Miller, P.R., Bekkerman, A., Jones, C.A., Burgess, M.H., Holmes, J.A., Engel, R.E., 2015.
Pea in rotation with wheat reduced uncertainty of economic returns in Southwest
Montana. Agron. J. 107, 541–550. https://doi.org/10.2134/agronj14.0185.
Mitchell, D.C., Lawrence, F.R., Hartman, T.J., Curran, J.M., 2009. Consumption of dry
beans, peas, and lentils could improve diet quality in the US population. J. Am. Diet
Assoc. 109, 909–913. https://doi.org/10.1016/j.jada.2009.02.029.
Nemecek, T., von Richthofen, J.-S., Dubois, G., Casta, P., Charles, R., Pahl, H., 2008.
Environmental impacts of introducing grain legumes into European crop rotations.
Eur. J. Agron. 28, 380–393. https://doi.org/10.1016/j.eja.2007.11.004.
O’Donnell, B.G., 2008. Life Cycle Assessment of American Wheat: Analysis of Regional
Variations in Production and Transportation. University of Washington.
Parajuli, R., Matlock, M.D., Thoma, G., 2021. Cradle to grave environmental impact
evaluation of the consumption of potato and tomato products. Sci. Total Environ.
758, 143662. https://doi.org/10.1016/j.scitotenv.2020.143662.
Patras, A., Oomah, B.D., Gallagher, E., 2011. By-product utilization. In: Tiwari, B.K.,
Gowen, A., McKenna, B. (Eds.), Pulse Foods: Processing, Quality and Nutraceutical
Applications. Academic Press, pp. 325–362.
Potter, S.R., Andrews, S., Atwood, J.D., Kellogg, R.L., Lemunyon, J., Noreet, L.,
Oman, D., 2006. Model Simulation of Soil Loss, Nutrient Loss, and Change in Soil
Organic Carbon Associated with Crop Production. USDA-Natural Resources
Conservation Services.
Reynolds, T., Klisch, S., Ravalin, D., Soule, K.E., 2018. It’s Electric?! Breaking Down
Electric Pressure Cookers. https://ucanr.edu/blogs/blogcore/postdetail.cfm?postnu
m=28441.
Sainju, U.M., Ghimire, R., Mishra, U., Jagadamma, S., 2020. Reducing nitrous oxide
emissions and optimizing nitrogen-use efciency in dryland crop rotations with
different nitrogen rates. Nutrient Cycl. Agroecosyst. 116, 381–395. https://doi.org/
10.1007/s10705-020-10046-0.
Sainju, U.M., Stevens, W.B., Caesar-TonThat, T., Liebig, M.A., 2012. Soil greenhouse gas
emissions affected by irrigation, tillage, crop rotation, and nitrogen fertilization.
J. Environ. Qual. 41, 1774–1786.
Schatz, B., Endres, G., 2009. Field Pea Production (No. A-1166). North Dakota State
University, Fargo, ND.
St¨
ossel, F., 2018. Life Cycle Assessment of Agricultural Production: Inventories and
Impact of Land Use on Soil Degradation. ETH Zurich, Switzerland.
Syngenta, 2019. Gramoxone SL 2.0 Herbicide, 5.6.19. https://www.syngenta-us.com/h
erbicides/gramoxone-sl-2.0.
Thies, S., Joshi, D.R., Bruggeman, S.A., Clay, S.A., Mishra, U., Morile-Miller, J., Clay, D.
E., 2020. Fertilizer timing affects nitrous oxide, carbon dioxide, and ammonia
emissions from soil. Soil Sci. Soc. Am. J. 84, 115–130. https://doi.org/10.1002/
saj2.20010.
Tidåker, P., Karlsson Potter, H., Carlsson, G., R¨
o¨
os, E., 2021. Towards sustainable
consumption of legumes: how origin, processing and transport affect the
environmental impact of pulses. Sustain. Prod. Consum. 27, 496–508. https://doi.
org/10.1016/j.spc.2021.01.017.
U.S. Bureau of Labor Statistics, 2018. CPI Average Price Data, 8.25.19. https://data.bls.
gov/PDQWeb/ap.
U.S. Census Bureau, 2019. Households by Size: 1960 to Present. Hist. Househ. Tables,
8.5.19. https://www.census.gov/data/tables/time-series/demo/families/househol
ds.html.
U.S. Census Bureau, 2018. 2018 National and State Population Estimates, 8.5.19. https
://www.census.gov/newsroom/press-kits/2018/pop-estimates-national-state.html.
U.S. Census Bureau, 2017. Annual Retail Trade Survey (ARTS), 8.5.19. https://www.ce
nsus.gov/programs-surveys/arts/data/tables.2017.html.
U.S. Department of Health and Human Services (HHS), 2015. U.S. Department of
agriculture (USDA), 2015-2020 Dietary Guidelines for Americans. eighth ed.
U.S. EIA, 2020a. Electricity data browser - average retail price of electricity, 11.8.20. .
https://www.eia.gov/electricity/data/browser/#/topic/7?
agg=0,1&geo=g&endsec=vg&linechart=ELEC.PRICE.US-ALL.A~ELEC.PRICE.US-
P.A. Bandekar et al.
Journal of Environmental Management 302 (2022) 114062
13
RES.A~ELEC.PRICE.US-COM.A~ELEC.PRICE.US-IND.A&columnchart=ELEC.
PRICE.US-ALL.A~ELEC.PRICE.US-RES.A~ELEC.PRICE.US-COM.A~ELEC.PRICE.
US-IND.A&map=ELEC.PRICE.US-ALL.A&freq=A&start=2001&end=
2018&ctype=linechart<ype=pin&rtype=s&maptype=0&rse=0&pin=.
U.S. EIA, 2020b. Frequently Asked Questions (FAQs) - U.S. Energy Information
Administration (EIA), 11.18.20. https://www.eia.gov/tools/faqs/faq.php.
U.S. EIA, 2015. Residential Energy Consumption Survey (RECS) - Data - U.S. Energy
Information Administration (EIA), 5.15.20. https://www.eia.gov/consumption/re
sidential/data/2015/index.php?view=consumption.
USA Dry Pea & Lentil Council, 2019. USA Dry Peas, Lentils & Chickpeas: Processing
Information & Technical Manual, 1.16.19. https://www.usapulses.org/technical-
manual/introduction.
USDA Economic Research Service, 2019. Food Availability (Per Capita) Data System,
6.21.19. https://www.ers.usda.gov/data-products/food-availability-per-capita-data-
system/.
USDA Economic Research Service, 2015. Most U.S. Households Do Their Main Grocery
Shopping at Supermarkets and Supercenters Regardless of Income, 9.10.19. https
://www.ers.usda.gov/amber-waves/2015/august/most-us-households-do-their-
main-grocery-shopping-at-supermarkets-and-supercenters-regardless-of-income/.
USDA National Agricultural Statistics Services, 2017. NASS - Quick Stats.
Walmart Inc., 2019. Our Retail Division, 8.5.19. https://corporate.walmart.com/n
ewsroom/2005/01/06/our-retail-divisions.
Wernet, G., Bauer, C., Steubing, B., Reinhard, J., Moreno-Ruiz, E., Weidema, B., 2016.
The ecoinvent database version 3 (part I): overview and methodology. Int. J. Life
Cycle Assess. 21, 1218–1230. https://doi.org/10.1007/s11367-016-1087-8.
Willlard, Bishop, 2016. Total Store Super Study(TM).
Wood, J.A., Malcomson, L.J., 2011. Chapter 8: pulse milling technologies. In: Tiwari, B.
K., Gowen, A., McKenna, B. (Eds.), Pulse Foods: Processing, Quality and
Nutraceutical Applications. Academic Press, pp. 193–221.
P.A. Bandekar et al.
