CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission
Pedro Henrique Assis Silva
1 a
, Regina Braga
1 b
, Jos
e Maria David
1 c
Valdemar Vicente Graciano Neto
2 d
, Wagner Arbex
1,3 e
and Victor Stroele
1 f
Federal University of Juiz de Fora, Juiz de Fora, MG, Brazil
Federal University of Goias, Goiania, GO, Brazil
Embrapa Dairy Cattle, Juiz de Fora, MG, Brazil
Software Ecosystem, Support Decision, Enteric Fermentation, Carbon Emission.
Global concerns about agriculture’s impact, mainly related to the livestock enteric fermentation producing
methane Greenhouse Gas (GHG) emissions, demand solutions to mitigate these impacts. The CarbonSECO
platform, tailored for carbon credit generation in Brazilian rural areas, responds to the popularity of carbon
credits to offset GHG emissions. However, additional solutions related to GHG need to be conceived. This
article extends the CarbonSECO platform, focusing on quantifying, monitoring, and controlling carbon emis-
sions from livestock enteric fermentation. Intelligent techniques, including ontologies and machine learning,
provide emissions management solutions for assessing and managing the environmental impact of livestock
farming. These techniques address the research question of mitigating carbon emissions from Brazilian dairy
farming. The article explores strategies, reviews related works, and proposes platform extensions. A feasibil-
ity study using data from an intelligent farm showcases the platform’s ability to predict and assess changes
for carbon emission reduction. As a result, this work enhances the CarbonSECO platform, offering emissions
management for dairy farming. Integrating ontologies and machine learning can promote standardization,
aiding property owners in better planning.
Global warming and its possible consequences on life
on earth have gained more and more space and im-
portance in society, transforming the mitigation of
Greenhouse Gas (GHG) emissions into the main fo-
cus of environmental debates and policies. In the
Brazilian context, agriculture plays a significant role
in greenhouse gas (GHG) emissions, representing
33.6% of the total, with 19% originating from enteric
fermentation (Embrapa, 2023b).
Enteric fermentation is a natural process in ru-
minant herbivorous animals, the result of convert-
ing vegetable carbohydrates into fatty acids used as
a source of energy by the animal that ingested the
food. During this conversion, gases such as methane
are produced, which is exhaled into the atmosphere
during the rumen. These gases are responsible for
worsening the greenhouse effect, resulting in acceler-
ated climate change (Siqueira and Caviglioni, 2021).
Livestock farming is responsible for 97% of methane
emissions, divided between 86% from beef cattle and
11% dairy cattle. Notably, methane from enteric fer-
mentation is one of the main sources of emissions in
milk production, contributing 40 to 58% of the total,
followed by waste management (Embrapa, 2023a).
Carbon credits represent a way to measure and
offset greenhouse gas (GHG) emissions associated
with specific activities, like dairy farming. A carbon
credit is generated for each ton of carbon that is no
longer emitted or captured from the atmosphere. Us-
ing selected and validated rules and methodologies,
whether projects are reducing emissions is analyzed.
To generate credits, actions must be taken to replace
an activity that would generate greenhouse gas emis-
sions with another solution that would reduce or elim-
inate these emissions (WayCarbon, 2022).
The ”baseline” (Vale Fund, 2022) is a key aspect
to be addressed by a carbon project. Every project
needs to determine what its emissions would have
been if the project had not been implemented. These
Silva, P., Braga, R., David, J., Neto, V., Arbex, W. and Stroele, V.
CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission Decisions.
DOI: 10.5220/0012734300003690
Paper published under CC license (CC BY-NC-ND 4.0)
In Proceedings of the 26th International Conference on Enterprise Information Systems (ICEIS 2024) - Volume 2, pages 89-99
ISBN: 978-989-758-692-7; ISSN: 2184-4992
Proceedings Copyright © 2024 by SCITEPRESS Science and Technology Publications, Lda.
are called “baseline emissions” (Vale Fund, 2022).
The number of credits a project receives is then calcu-
lated by subtracting the project’s emissions from the
baseline emissions. The definition of the baseline will
depend on the project activity and the sector, to deter-
mine which elements and parameters should be con-
sidered for the project.
The Federal University of Juiz de Fora, in part-
nership with EMBRAPA (an acronym for Brazilian
Agricultural Research Corporation) and the Federal
University of Goi
as, develop projects in the domain
of dairy farming (Gomes et al., 2023; Ambr
osio et al.,
2021; Soares et al., 2021; Magaldi et al., 2017) to
seek computational solutions to support dairy pro-
duction activities. With this partnership, a software
ecosystem platform was proposed, called E-SECO
osio et al., 2021), an evolution of the collab-
orative E-Science project, to support the development
of solutions for agribusiness.
E-SECO encompasses several functionalities to
facilitate the development of this type of application,
considering the use of scientific workflows and the
process of scientific experimentation. In E-SECO,
data analysis is an important component (Gomes
et al., 2023). This platform offers a suitable envi-
ronment for carrying out various stages of experi-
ments involving data related to dairy farming, rang-
ing from problem investigation, prototyping, plan-
ning, conduction/analysis, packaging of results, and
preparation of reports (Ambr
osio et al., 2021).
Based on the evolution of the group’s research
(Gomes et al., 2023), we extended the E-SECO plat-
form to CarbonSECO (Santos et al., 2023). Carbon-
SECO aims to enable the development of applications
related to generating carbon credits on rural proper-
ties in Brazil. CarbonSECO prioritizes mechanisms
that offer analysis, reliability, and traceability of data
and processes in carbon certification projects (Santos
et al., 2023).
Considering the importance of this topic today, the
relevance of Brazil in the global carbon emission sce-
nario, and specific decisions of COP 28
, this article
presents specific extensions made on CarbonSECO to
quantify, monitor, and control carbon emissions from
enteric fermentation. Specific services were devel-
oped, expanding the CarbonSECO support on car-
bon emissions in dairy farming, considering intelli-
gent techniques. These new services use ontologies
and ML techniques to provide intelligent solutions for
managing greenhouse gas emissions from livestock
farming. The main goal is to offer these services to
be used in developing applications for evaluation and
managing the environmental contributions of live-
stock farming. Specifically, these services can assist
in developing decision-making applications, based on
predictions about specific features related to the pro-
duction process of milk and its derivatives. In this
regard, a feasibility study was conducted, with data
from the smart farm maintained by EMBRAPA- Gado
de Leite. It was possible to predict planned changes to
reduce carbon emissions using ontological processing
and ML.
This work’s Main Research Question (RQ) is
”How to help mitigate carbon emissions from dairy
farming, with a focus on Brazilian farms?”. To ad-
dress this specific topic, this article is organized into
the following sections, including this introduction.
Section 2 discusses Brazilian strategies for dealing
with issues related to carbon emissions. In section
3, related work is discussed. Section 4 presents the
CarbonSECO extension proposal. In section 5 a fea-
sibility study is detailed. In section 6, conclusions and
future work are presented.
The global reference for quantifying greenhouse gas
(GHG) emissions, is the IPCC Guidelines for Na-
tional Greenhouse Gas Inventories (IPCC, 2021). The
IPCC established a standard bottom-up predictive
model, which has undergone refinements over time.
These models are stratified into different levels of
complexity. Tier 1 is the simplest, using standard
emission factors based on general literature. Tier
1 does not consider the characterization of regional
livestock systems, such as breed types, animal age,
and physiological issues (IPCC, 2021). Tier 2 in-
corporates detailed emission factors, adding specific
characterization of food and animals, using estimates
based on gross energy consumption (GEI) and CH
conversion factor (Ym, expressed as % of GEI con-
verted to CH
) (IPCC, 2021).
The Verified Carbon Standard (VCS) Program
stands out as a global leader in voluntary certification
of carbon offsets (Greenfield, 2023). With a back-
ground of more than a billion tons of carbon and
other greenhouse gas emissions reduced or removed
through VCS-certified projects, this program plays an
essential role in the ongoing global effort to protect
our shared environment (Verra, 2022).
Considering the VCS, the proposal presented in
this article is related to Tiers 1 and 2. Emissions in the
reference scenario are estimated as the sum of annual
emissions from enteric fermentation according to the
ICEIS 2024 - 26th International Conference on Enterprise Information Systems
following equation:
i, j
× GW P × 0.001
= Total CH
emissions in the reference
scenario from enteric fermentation of cattle on the
farm i tCO
i, j
= Enteric CH
emissions factor for an-
imal group j during the monitoring period kgCH
GW P = Global Warming Potential of Methane.
n = Total number of animal group on farm i.
To determine the enteric emission factor, it is pos-
sible to use two calculations, the choice depending on
the availability of data for each group of animals on
the farm. The first option provides procedures to cal-
culate the enteric emission factor for each group of
animals, applying an IPCC Tier 2 method, using the
following equation:
i, j
× Y
× N
i, j
× Days
i, j
100 × EC
i, j
= Enteric CH
emission factor for
group of animals j during the monitoring period
= Average gross energy consumption (GEI)
per animal in group j on farm i (MJhead
= Conversion factor indicating the proportion
of GEI converted into enteric CH
energy by animal
group, CH
energy as a percentage of gross (dimen-
sionless) energy.
i, j
= Number of days spent on farm i by each
animal in group j during the monitoring period (d).
i, j
= Average number of heads in group of ani-
mals j on farm i in the monitoring period (head).
EC = Energy content of methane (55.65MJkg
Average Gross Energy Intake (GEI) is calculated
by multiplying Average Dry Matter Intake (DMI) by
the energy density of the food, using the following
× ED (3)
= Average dry mass of food consumed by
the group of animals j on a given day (kg head
ED = Average dry matter energy density (MJ
). Standard values of 19.10 MJ kg
for diets
that include oils with a fat content in the range of 4 to
6%, or 18.45 MJ kg
for diets that include oils with
a fat content below 4%, can be utilized.
The second option is applicable in cases where
certain data required for the calculations of the first
option are unavailable. In this scenario, Tier 1 enteric
fermentation emission factors for cattle and buffalo
are employed. The enteric emission factor for each
group of animals is computed using the formula:
i, j
= EF
i, j
× N
i, j
× Days
i, j
i, j
= Enteric CH
emission factor for the
group of animals j during the monitoring period (kg
i, j
= Average enteric C H
emission fac-
tor for the group of animals j during the monitoring
period (specific national or regional factors) (kg CH
i, j
= Number of days spent on farm i by each
animal in group j during the monitoring period (d).
i, j
= Average number of heads in the group of
animals j on farm i in the monitoring period (head).
In this article, we apply these calculations to quan-
tify emissions from enteric fermentation based on the
type of information provided by farms. Section 5
presents examples of specific calculations using the
proposed equations.
The literature presents few works that deal with car-
bon emissions to mitigate and assist in decision-
making on emissions-related practices. Some works
deal with general context, but none effectively pro-
pose computational strategies, especially those in-
volving intelligent processing. In this context, we
seek to build an architecture with the objective of sup-
porting the development of a software ecosystem plat-
form, called CarbonSECO.
(Tedeschi et al., 2022) reviewed methods for
quantifying methane emissions from ruminants and
their manure. The study addresses various mea-
surement techniques from classical methodologies,
such as breathing chambers, to micrometeorological
approaches, and explores the challenges associated
with each method. The authors emphasize the need
for accurate quantification of greenhouse gas emis-
sions, mainly methane, to ensure adequate reporting
in greenhouse gas inventories and to design effective
methane emissions mitigation strategies. The review
shows the diversity of approaches, including aircraft,
drones, and satellites and highlights the importance
of addressing knowledge gaps and research require-
ments in this field. Among the methodologies covered
CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission Decisions
are those for calculating methane emissions using the
Tier 1 and Tier 2 approaches proposed by IPCC.
(Mazzetto et al., 2022) investigated the carbon
footprint of bovine milk production in several coun-
tries, identifying critical factors such as allocation
methods and functional units that significantly influ-
ence results. The 21 studies analyzed revealed consid-
erable variations in carbon footprint estimates, high-
lighting the process’s complexity. Additionally, the
study highlighted mitigation strategies, focusing on
the importance of waste management, livestock feed
production, and fertilization methods to reduce emis-
sions. The findings provide practical insights for the
industry, indicating specific areas for interventions
aimed at reducing the environmental impact of bovine
milk production, with the caveat that strategies must
be adapted to the agricultural conditions of each coun-
(Wang et al., 2023) conducted a review on the
impact of digital technologies, including the Internet
of Things, Big Data, cloud computing, blockchain
and AI, on sustainable supply chains. It highlights
the application of these technologies in specific prac-
tices, such as green procurement, environmentally
conscious production, sustainable consumption and
ecological logistics. The study emphasizes these
technologies’ interconnectedness and potential to re-
duce energy consumption, contributing to greener and
more efficient supply chains. It also includes the need
to explore pricing on sharing platforms, the design of
integrated digital systems, and continued technologi-
cal innovations.
(O’Brien et al., 2014) investigated the impact of
methodological decisions in Life Cycle Evaluation
(LCE) on the carbon footprint of milk in dairy produc-
tion systems. The authors highlighted the influence
of greenhouse gas emissions allocation between co-
products and the lack of consistency in international
LCE standards. Issues related to carbon sequestration
and land use change emissions were addressed. The
study highlights the need for more specificity in the
LCA methodology for accurate comparisons between
different dairy production systems.
(Pirlo and Car
e, 2013) also address milk produc-
tion’s greenhouse gas (GHG) emissions. They pro-
posed LatteGHG, a tool to estimate the carbon foot-
print of cow’s milk produced under typical conditions
in Italy. The study shows the effectiveness of Lat-
teGHG in the environmental evaluation of production
systems, revealing the model’s sensitivity to variables
such as milk production and manure management.
Despite the simplicity of some emission factors, the
research highlights the continued need to improve the
accuracy of estimates of emissions of gases such as
methane and nitrous oxide, considering diet composi-
tion and animal performance.
(Vogel and Beber, 2022) examine the carbon foot-
print and mitigation strategies on heterogeneous dairy
farms in Paran
a, Brazil. Grouping farms into four
categories through statistical analysis and life cycle
evaluation, the study highlights striking differences in
carbon footprint between these groups. Farms with
the largest carbon footprint predominate in the region,
characterized by less specialized herds and less tech-
nical support. To address climate change in the sector,
the study highlights the need to promote sustainable
practices, integrating them into a broader approach
to environmental governance and regional socioeco-
nomic development.
(Desjardins et al., 2012)) analyzed the carbon
footprint of cattle in countries such as Canada, the
United States, the European Union, Australia, and
Brazil, highlighting variations from 8 to 22 kg of CO
per kg of live weight. Over the past few decades,
improvements in management practices have resulted
in significant reductions in greenhouse gas emissions
per unit of product. The study addresses the influ-
ence of different factors, such as breeding systems,
locations and management practices, and discusses
the competition for higher-quality land in beef pro-
duction. It also shows the importance of considering
changes in soil carbon due to land management. In
general, it considers that the carbon footprint is just
one aspect of beef production, and it is essential to
consider the impact on services and biodiversity for
a complete evaluation of the product’s environmental
(Singh et al., 2015) propose an integrated cloud-
based system to measure and reduce the beef supply
chain’s carbon footprint. Aimed especially at small
and medium-sized farms, the system allows access
to carbon calculators through a private cloud, provid-
ing crucial information to optimize emissions. Fur-
thermore, the collaborative approach aims to improve
stakeholder coordination, identify sustainable prac-
tices, and address the consumer demand for trans-
parency and traceability in beef production. The study
highlights the relevance of CCT in mitigating carbon
emissions, contributing to environmental efficiency
and government carbon reduction targets.
(Liu et al., 2023) investigated the impact of imple-
menting digital technology on carbon emission effi-
ciency on dairy farms in China using empirical meth-
ods and farm data. The results highlighted a posi-
tive influence of digital technology on improving car-
bon emission efficiency, emphasizing precision feed-
ing technology as the biggest contributor. Farmers’
educational experience, technical training, and gov-
ICEIS 2024 - 26th International Conference on Enterprise Information Systems
Table 1: Comparison of related works.
Authors Architecture Intelligent Enteric Methane Decision
and SECO Processing Emissions Support
(O’Brien et al., 2014) X
(Pirlo and Car
e, 2013) X
(Vogel and Beber, 2022) X X
(Desjardins et al., 2012) X
(Singh et al., 2015) X X X
(Liu et al., 2023) X X
(Berhe et al., 2020) X X
(Santos et al., 2023) X X X
(Gomes et al., 2023) X X X
ernment incentives contributed to carbon efficiency.
It was observed that the application of digital tech-
nology had a more pronounced effect on farms with
lower initial emission efficiency, highlighting the im-
portance of environmental regulation as a positive
moderator in this process. The findings provide in-
sights for strategies for transitioning to low-carbon,
digital dairy production systems.
(Berhe et al., 2020) analyzed the impacts of differ-
ent interventions on greenhouse gas (GHG) emissions
in livestock production systems in Ethiopia. A re-
duction in total GHG emissions was observed by im-
proving livestock feed, implementing improved ma-
nure management practices, and optimizing herd pro-
duction and reproductive parameters. Dietary supple-
mentation with corn grains resulted in a decrease in
enteric methane emissions. Furthermore, the joint im-
plementation of these interventions demonstrated an
overall reduction in emissions, highlighting the im-
portance of integrated management to promote envi-
ronmental efficiency in livestock production systems.
These results contribute to understanding the fac-
tors influencing GHG emissions in different produc-
tion systems, providing insights for environmentally
sustainable mitigation strategies. (Gomes et al., 2023)
present the evolution of the E-SECO platform to ad-
dress automation in livestock farming, highlighting
the e-Livestock architecture to improve animal pro-
duction. Two case studies at the Compost Barn sys-
tem demonstrate the platform’s effectiveness. The im-
portance of self-adaptation, AI, and IoT is highlighted
to face dynamic challenges in agriculture, promoting
efficiency and preventing losses in animal production.
The article points to the rise of Livestock IoT (LIoT)
as a new area of research, emphasizing the need to
study specific requirements and technologies for live-
stock farming with the widespread adoption of IoT.
(Santos et al., 2023) propose the CarbonSECO plat-
form for the Control of Carbon Credits, addressing
the growing concern about climate change. Focus-
ing on changing land use as Brazil’s main source of
emissions in 2021, the CarbonSECO platform is pre-
sented as support for developing applications related
to controlling emissions/carbon credits on rural prop-
erties. The aim is to generate knowledge to offer al-
ternatives for cultivating land and mitigating green-
house gas emissions. The platform incorporates in-
telligent data analysis, highlighting its potential pos-
itive impact on the Brazilian agricultural sector. Pre-
liminary results demonstrate the platform’s viability,
with an analysis of carbon stocks on a rural prop-
erty, highlighting its potential to provide insights into
sustainable agricultural practices and reducing green-
house gas emissions. The contribution of our work
in the group lies in introducing a service that deals
with managing carbon emissions while incorporating
intelligent data analysis to provide alternative solu-
tions in the livestock sector. Our approach seeks to
reduce emissions and promote sustainable livestock
Table 1 highlights the main features of each work
presented in this section. These studies reflect the
complexity and diversity of approaches needed to ad-
dress greenhouse gas emissions in livestock farming,
indicating the importance of quantification methods,
mitigation strategies, and technological solutions. As
we can see, no works propose a suite of intelligent
services to support decision-making related to green-
house gas emissions using intelligent processing as
we do in our work. In this regard, our work presents
a proposal encompassing services for quantifying,
monitoring, and controlling emissions from enteric
fermentation while at the same time offering support
for evaluating management in livestock farming, ex-
tending the functionalities of the CarbonSECO plat-
form proposed in (Santos et al., 2023). The focus
CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission Decisions
Figure 1: CarbonSECO Main Services with Enhanced Livestock Carbon Emissions and Control Services (Santos et al., 2023).
is to provide specific services that allow rural pro-
ducers to quantify methane emissions from ruminants
and make decisions based on strategic analyses pro-
vided by the platform, such as estimates of the impact
of a new diet strategy. These services reinforce the
support of the CarbonSECO platform in controlling
greenhouse gas emissions in agriculture in general,
adding specific services for evaluating and managing
environmental contributions in ruminant farming.
As stated before, the CarbonSECOfor Livestock is
an evolution of our previous platform CarbonSECO
(Santos et al., 2023) to provide a suite of services re-
lated to managing carbon emission/credits in the live-
stock domain. CarbonSECO emphasizes mechanisms
that bring intelligent analysis to Carbon certification
projects. Figure 1 presents CarbonSECO’s main ser-
vices enhanced with new services (left of the figure)
related to livestock carbon emissions and control, the
article’s focus.
It is not our objective to detail the CarbonSECO
previous services. They are detailed in (Santos et al.,
2023). The new service components are discussed be-
4.1 Smart Farm Dataset
In addition to the databases related to rural proper-
ties and crops, new data sources, primarily associated
with livestock, were added to CarbonSECO. The data
comes from various sources, whether data from sen-
sors attached to the animals, such as necklaces and
earrings, or platforms for weighing and feeders with
intake control. In addition, specific measurements of
milk production and animal registration data, includ-
ing their traceability, are part of the data from Smart
Datasets’ data sources. However, to guarantee the re-
liability of the carbon measurement process, it must
be based on international guidelines and methodolo-
gies, such as the IPCC, used by CarbonSECO (IPCC,
4.2 Ontology Processing
The EntericMeasureOnto ontology was developed to
identify semantic relationships between data, result-
ing in the discovery of specific features that must be
observed in the carbon measurement process. For ex-
ample, based on semantic relationships, we can iden-
tify a feeding protocol that explicitly does not present
an increase in carbon emissions but which, based on
the relationships identified in the ontology can be an
indirect source of increased carbon emissions. This
type of relationship can only be discovered by pro-
cessing inferences and semantic rules defined in the
ontology. Consequently, optimizing livestock man-
agement on a specific rural property can improve the
carbon credits associated with that property.
For this, classes and different types of relation-
ships were defined in EntericMeasureOnto Figure 2,
using the OWL language (World Wide Web Consor-
tium, 2004), SWRL rules (SWRL, 2004), and the Pel-
let reasoner (Sirin et al., 2007). The main ontolog-
ical classes are: RuralProperty, which represents a
rural property, BaselineEmissons, which represents
ICEIS 2024 - 26th International Conference on Enterprise Information Systems
Figure 2: Main ontological classes.
Table 2: SWRL rules.
Tag Rule
S1 Cattle(?cattle) ˆ milkProduction(?cattle, ?production) -> DairyCattle(?cattle)
S2 Cattle(?cattle) ˆ DairyCattle(?cattle) ˆ emissionFactorTier1(?cattle, ?e) ˆ daysOnFarm(?cattle, ?d) ˆ
multiply(?result, ?e, ?d) ˆ -> individualEntericEmissionFactorTier1(?cattle, ?result)
S3 Cattle(?cattle) ˆ individualEntericEmissionFactorTier1(?cattle, ?eeft1) ˆ gwp(?cattle, ?gwp) ˆ
multiply(?result, ?eeft1, ?gwp, 0.001) ˆ -> individualEntericEmissionTier1(?cattle, ?finalResult)
S4 DairyCattle(?cattle) ˆ energyDensity(?cattle, ?e) ˆ dryMatterIntake(?cattle, ?d) ˆ multiply(?result,
?e, ?d) ˆ -> grossEnergyIntake(?cattle, ?result)
S5 Cattle(?cattle) ˆ grossEnergyIntake(?cattle, ?gei) ˆ daysOnFarm(?cattle, ?d) ˆ
emissionFactorTier2(?cattle, ?ef) ˆ multiply(?result, ?gei, ?d, ?ef, 0.01) ˆ divide(?finalResult,
?result, 55.65) ˆ -> individualEntericEmissionFactorTier2(?cattle, ?finalResult)
S6 Cattle(?cattle) ˆ individualEntericEmissionFactorTier2(?cattle, ?eeft2) ˆ gwp(?cattle, ?gwp) ˆ
multiply(?result, ?eeft2, ?gwp, 0.001) ˆ -> individualEntericEmissionTier2(?cattle, ?finalResult)
the baseline methane emissions from the enteric fer-
mentation of animals, BaselineEmissonsTier1 and
BaselineEmissonsTier2, which represents the base-
line methane emissions from the enteric fermentation
of animals from the Tier 1 and 2 calculation methods
respectively, Cattle, which represents the cows for
which enteric emissions are estimated, DairyCattle
and OtherCattle, which represent the cows that are
producing milk or not, respectively, Measurement-
Data, which represents the data measurements asso-
ciated with a Cattle instance. ObjectProperties were
also specified to implement the relationships and spe-
cific rules in SWRL (SWRL, 2004) to process carbon
emissions. As stated before, the rules were specified
according to Tier 1 and 2 calculations.
As a result, using the declared model (explicit
knowledge) with the addition of specific SWRL rules
and inference mechanism (Pellet reasoner), the ontol-
ogy infers, from the instantiated data, new relation-
ships under the actions on the rural property. As spe-
cific examples, in Table 2, Rule S1 infers that if an an-
imal has milk production, it is a DairyCattle instance.
Rule S2 infers the enteric C H
emission factor for the
animal during the monitoring period according to Tier
1. Rule S3 infers the enteric CO
emission for the an-
imal during the monitoring period according to Tier
1. Rule S4 infers the average gross energy consump-
tion for the animal. Rule S5 infers the enteric CH
emission factor for the animal during the monitoring
period according to Tier 2. Rule S6 infers the enteric
emission for the animal during the monitoring
period according to Tier 2.
4.3 Decision Support
Data analysis on a rural property can potentially dis-
cover insights that help manage and plan emissions
related to livestock practices. Furthermore, they pro-
vide support in the adoption of sustainable practices.
To assist in decision support related to carbon emis-
sions, in addition to the ontology, which allows the
discovery of new relationships between data, we use
ML techniques to predict specific actions to improve
practices related to carbon emissions. To do this, we
CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission Decisions
Figure 3: Number of Dairy Cattle Over Time.
use linear regression to analyze and predict emissions
linked to practices carried out on the farm. A dataset
was used with data captured by sensors and specific
measurements on the farm, plus specific data gener-
ated from information delivered by experts, includ-
ing the average daily consumption of dry matter, the
energy density of the feed, and the average number
of cattle in a specific period. These variables were
used as training data for the linear regression model.
Throughout the training process, the algorithm ad-
justed the coefficients to optimize the model’s ability
to predict emissions based on changes in the variables
In this context, the use of ML techniques aims
to assist in predictive analysis of the carbon emis-
sions of a rural property, taking into account expecta-
tions of changes in diets and the number of animals
over a specific period. This predictability helps in
planning farm activities and anticipating specific ac-
tions related to carbon emissions by the property, such
as sharing knowledge and allowing more informed
choices about the best conditions for using animal diet
This data analysis is presented on a dashboard to
assist in decision-making. It uses graphs and spe-
cific variables, allowing an integrative view of car-
bon management on the property, offering insights
into the environmental performance of its practices,
and enabling strategic adjustments to achieve sustain-
able goals. For example, Figure 3 depicts a graph il-
lustrating the number of dairy cattle on the property
throughout 2021 in Embrapa’s Smart Farm.
This section details a feasibility study to evaluate Car-
bonSECO livestock’s new services. This evaluation
considers using CarbonSECO to deal with quanti-
fying, monitoring, and controlling carbon emissions
from enteric fermentation. The scope of the study was
defined as Analyze the use of the CarbonSECO Live-
stock services from the point of view of Farmers in the
context of data extracted from Dairy Smart Farm”.
From this scope, we derived the RQ: “How to
help mitigate carbon emissions from dairy farming,
with a focus on Brazilian farms?”. The data extracted
can be reached at GitHub
. This data set encom-
passes 162 animals, including sensor data and pro-
duction system-related details, with 3,381 observa-
tions collected from January 2021 to December 2021,
encompassing monthly data collection that occurred
during three milking sessions. The dataset includes
key features such as Tag (serving as a unique iden-
tifier), Weight (measured in kilograms), Milk Pro-
duction (indicating the quantity of milk produced),
and Date (recording the month of the measurement).
For this analysis, the monitored cow remained on the
property throughout the month. It’s important to note
that no measurements were recorded in June. To pro-
cess this data, we used the EntericMeasureOnto on-
To find out existing ontologies that can be reused
in this evaluation, a search was conducted. Unfortu-
nately, no one that fits our purposes was found, there-
fore, the domain ontology was specified, consider-
ing the steps specified in (Verra, 2020; Nicola and
Missikoff, 2016; Feilmayr and W
ob, 2016), and the
data extracted from the Smart Farm related to EM-
BRAPAs research projects.
Using the CarbonSECO new services, we seek to
make predictions on calculating emissions in a future
scenario. These predictions can help choose strate-
Enteric Fermentation Emissions/blob/main/data/
ICEIS 2024 - 26th International Conference on Enterprise Information Systems
Figure 4: Monthly Carbon Emissions using Tier 1 Methodology.
Figure 5: Projected Total Carbon Emissions using Carbon-
SECO Service.
gies for mitigating carbon emissions from livestock
farming. Therefore, based on the available data, it
was possible to carry out emission calculations using
the EntericMeasureOnto ontology, following the Tier
1 method. These calculations imply that the analysis
of carbon emissions in dairy farming benefited from
the ontology’s structuring and organization, improv-
ing the results’ precision and reliability. Figure 4 il-
lustrates the carbon emissions trends over the months,
emphasizing the total emissions for the period.
To verify the calculation module utilizing the Tier
2 method, we added values into the dataset related
to the diet of animals. In this context, the Aver-
age Energy Density and Average Dry Matter Intake
were 25 kg per head daily and 18 MJ/kg, respectively.
The emissions were subsequently computed using the
EntericMeasureOnto ontology, following the Tier 2
method. Figure 6 presents the monthly progression
of carbon emissions, detailing both the individual
monthly values and the cumulative total for the spec-
ified period.
To evaluate the feasibility of the forecasting mod-
ule, we processed a specific dataset with a fixed dura-
tion of 335 days, providing a representative timeframe
for the analysis. Key input parameters included the
average daily dry matter intake (Average DMI), 21,
representing typical feeding conditions for cattle. The
energy density of the feed (Energy density of feed)
wwas 18, as a measure of the nutritional characteris-
tics of the diet. The average number of cattle (Aver-
age number of heads) in the dataset was 200.
Using the CarbonSECO service in processing this
dataset, it became feasible to forecast an estimated
emissions total. The results indicate a projected to-
tal of 598.48 tons of CO2, offering insights into the
anticipated environmental impact within the specific
context of the provided conditions and practices. This
projection is graphically illustrated in Figure 5 , pro-
viding an overview of the evolution of emissions and
the total impact over the specified period.
In summary, the feasibility study of CarbonSECO
Livestock’s new services yielded insights into carbon
emissions management in dairy farming. These re-
sults help farmers in decisions related to carbon emis-
sions in the farm. Therefore, the CarbonSeco helps in
decision support related to caron emissions. Utilizing
Tier 1 and Tier 2 methods, and incorporating specific
data, we could answer the RQ, “How to help mitigate
carbon emissions from dairy farming, with a focus on
Brazilian farms?”. The Tier 1 method, using actual
data from Dairy Smart Farm showcased the efficacy
of the EntericMeasureOnto ontology in refining car-
CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission Decisions
Figure 6: Monthly Carbon Emissions using Tier 2 Methodology.
bon emission calculations, enhancing precision and
reliability. Illustrated trends in Figure 4 emphasized
monthly evolution and total impact, aiding informed
The Tier 2 method, employing specific values,
showed CarbonSECO’s calculation module capabil-
ity. The projection of 598.48 tons of CO2 over 335
days offers a predictive tool for farmers, providing
proactive mitigation strategies and fostering sustain-
able livestock management practices.
Therefore, CarbonSECO Livestock has the poten-
tial for carbon emission management in dairy farm-
ing, providing services for real-world and predic-
tive scenarios. The results and comparisons lay the
groundwork for further exploration, fostering envi-
ronmentally conscious decision-making in the agri-
cultural sector.
This paper presented services for quantifying, mon-
itoring, and controlling emissions from enteric fer-
mentation that will be integrated into CarbonSECO.
This addition expands the platform’s capacity to cover
specialized services focused on carbon emissions and
supports the development of applications related to
generating carbon credits on Brazilian rural proper-
The utilization of ontologies and ML algorithms
contributed to the standardization and estimation of
future emissions, these intelligent technologies em-
power property owners with better planning capabili-
The integration of these services supports the Car-
bonSECO platform with new functionalities, pro-
vides the user with social and environmental consid-
erations, promoting sustainable production practices
with reduced carbon generation. This, in turn, pro-
vides valuable decision support for stakeholders deal-
ing with carbon emissions in the agricultural sector.
In future work, it is essential improve the pro-
posed service with different methods for calculating
fermentation emissions arising from enteric fermen-
tation. Additionally, there is a need for continuous
improvement of the machine learning module to en-
hance the system’s predictive accuracy. The adap-
tation and incorporation of new types of input data
to enrich the analytical capabilities of the platform
further. Lastly, the addition of a new service dedi-
cated to quantifying, monitoring, and controlling an-
imal waste emissions will provide a comprehensive
solution, ensuring a more embracing approach to en-
vironmental sustainability in livestock farming.
The data used in this study and the code im-
plementation are available at
PedroAssisDev/Measuring Enteric Fermentation
This work was partially funded by UFJF/Brazil,
CAPES/Brazil, CNPq/Brazil (grant: 307194/2022-
1), and FAPEMIG/Brazil (grant: APQ-02685-17),
(grant: APQ-02194-18).
ICEIS 2024 - 26th International Conference on Enterprise Information Systems
osio, L., Linhares, H., David, J. M. N., Braga, R., Ar-
bex, W., Campos, M. M., and Capilla, R. (2021). En-
hancing the reuse of scientific experiments for agricul-
tural software ecosystems. Journal of Grid Comput-
ing, 19:1–24.
Berhe, A., Bariagabre, S. A., and Balehegn, M. (2020). Es-
timation of greenhouse gas emissions from three live-
stock production systems in ethiopia. International
Journal of Climate Change Strategies and Manage-
ment, 12(5):669–685.
Desjardins, R. L., Worth, D. E., Verg
e, X. P., Maxime, D.,
Dyer, J., and Cerkowniak, D. (2012). Carbon footprint
of beef cattle. Sustainability, 4(12):3279–3301.
Embrapa (2023a). Dairy livestock in brazil and global
warming. Technical report, Embrapa. Accessed:
Embrapa (2023b). Drinking milk helps livestock become
carbon neutral, says embrapa. Accessed: 2024-01-22.
Feilmayr, C. and W
ob, W. (2016). An analysis of ontologies
and their success factors for application to business.
Data & Knowledge Engineering, 101:1–23.
Gomes, J., Esteves, I., Neto, V. V. G., David, J. M. N.,
Braga, R., Arbex, W., Kassab, M., and de Oliveira,
R. F. (2023). A scientific software ecosystem archi-
tecture for the livestock domain. Information and Soft-
ware Technology, 160:107240.
Greenfield, P. (2023). As carbon offsetting faces a ’cred-
ibility revolution’, shoppers should be wary. The
IPCC (2021). Climate change 2021: The physical science
basis - summary for policymakers. Technical report,
Intergovernmental Panel on Climate Change (IPCC).
Accessed: 2024-01-22.
Liu, C., Wang, X., Bai, Z., Wang, H., and Li, C. (2023).
Does digital technology application promote carbon
emission efficiency in dairy farms? evidence from
china. Agriculture, 13(4):904.
Magaldi, H., Braga, R., Arbex, W., Campos, M., Borges,
C., David, J., and Campos, F. (2017). Use of on-
tologies and visualization techniques to support re-
search in feed efficiency of dairy cattle. In: Brazilian
Congress of Agroinformatics, 11th, 2017, Campinas.
Science of...
Mazzetto, A. M., Falconer, S., and Ledgard, S. (2022).
Mapping the carbon footprint of milk production from
cattle: A systematic review. Journal of Dairy Science.
Nicola, A. D. and Missikoff, M. (2016). A lightweight
methodology for rapid ontology engineering. Com-
munications of the ACM, 59(3):79–86.
O’Brien, D., Capper, J., Garnsworthy, P., Grainger, C., and
Shalloo, L. (2014). A case study of the carbon foot-
print of milk from high-performing confinement and
grass-based dairy farms. Journal of dairy science,
Pirlo, G. and Car
e, S. (2013). A simplified tool for esti-
mating carbon footprint of dairy cattle milk. Italian
Journal of Animal Science, 12(4):e81.
Santos, L. F., Gomes, J., Braga, R., David, J. M. N.,
and Stroele, V. (2023). Towards a seco for carbon
credit control. In 2023 IEEE/ACM 11th International
Workshop on Software Engineering for Systems-of-
Systems and Software Ecosystems (SESoS), pages 13–
21. IEEE.
Singh, A., Mishra, N., Ali, S. I., Shukla, N., and Shankar,
R. (2015). Cloud computing technology: Reducing
carbon footprint in beef supply chain. International
Journal of Production Economics, 164:462–471.
Siqueira, F. and Caviglioni, M. (2021). Metagenomics, nan-
otechnology, and animal nutrition: alternatives to an-
tibiotic use and greenhouse gas mitigation.
Sirin, E., Parsia, B., Cuenca Grau, B., Kalyanpur, A., and
Katz, Y. (2007). Pellet: A practical OWL-DL rea-
soner. Journal of Web Semantics, 5(2):51–53.
Soares, N., Braga, R., Maria N. David, J., Beatriz Siqueira,
K., da Silva Nogueira, T., Wendelin Campos, E.,
Augusto Priamo Moares, E., and Vanessa Zabala
Capriles Goliatt, P. (2021). Redic: Recommenda-
tion of digital influencers of brazilian artisanal cheese:
Redic: Recomendac¸
ao de influenciadores digitais do
queijo artesanal brasileiro. In XVII Brazilian Sympo-
sium on Information Systems, pages 1–8.
SWRL, A. (2004). Semantic web rule language combining
owl and ruleml. W3C Member Submission (May 21,
2004), http://www. w3. org/Submission/SWRL/(last
visited March 2011).
Tedeschi, L. O., Abdalla, A. L.,
Alvarez, C., Anuga, S. W.,
Arango, J., Beauchemin, K. A., Becquet, P., Berndt,
A., Burns, R., De Camillis, C., et al. (2022). Quantifi-
cation of methane emitted by ruminants: a review of
methods. Journal of Animal Science, 100(7):skac197.
Vale Fund (2022). Carbon market report-fv ecosecurities.
Technical report, Vale Fund. Accessed: 2024-01-22.
Verra (2020). Methodology for the reduction of enteric
methane emissions from ruminants through the use
of feed ingredients. Methodology, Verra. Accessed:
Verra (2022). Verified carbon standard (vcs) program. https:
// Ac-
cessed: 2024-01-22.
Vogel, E. and Beber, C. L. (2022). Carbon footprint and mit-
igation strategies among heterogeneous dairy farms
in paran
a, brazil. Journal of Cleaner Production,
Wang, Y., Yang, Y., Qin, Z., Yang, Y., and Li, J. (2023).
A literature review on the application of digital tech-
nology in achieving green supply chain management.
Sustainability, 15(11):8564.
WayCarbon (2022). Opportunities for brazil in carbon mar-
kets. Accessed: 2024-01-22.
World Wide Web Consortium (2004). OWL Web Ontology
Language. World Wide Web Consortium Recommen-
CarbonSECO for Livestock: A Service Suite to Help in Carbon Emission Decisions