The United States Energy Consumption and Carbon Dioxide Emissions: A Comprehensive Forecast Using a Regression Model

Abstract

The Earth’s climate change, colloquially known as global warming, is detrimental to life across the globe. The most significant contributor to the greenhouse gas (GHG) effect is carbon dioxide (CO₂) emission. In the United States (US) economy, the major benefactor of CO₂ emissions is the energy sector, with the top contribution coming from fossil fuels. The estimated 2020 CO₂ emission was 5981 million metric tons, despite a dramatic reduction in the trendline compared to the year 2019. An ultimatum for energy consumption rises from fiscal development, growing population, and technological advancements. Energy use and GHG emissions are inclined upward, provoking an unwholesome nation. This paper studies (i) the principal sources of energy use and CO₂ emission, (ii) the inclination of such sources, (iii) trends and drivers of GHG emissions, (iv) low carbon development and carbon footprint, and (v) the diverse US projects for reducing GHG emissions and the challenges in deploying them. We have forecasted the emissions from fossil fuels from 2025 to 2050 and compared the results using MAPE to calculate the mean percentage error. The forecasted results of 2050 show high accuracy, suggesting probable approaches to reduce further CO₂ emissions, measures to reduce emissions through carbon capture and sequestration, and help in the development of improved GHG mitigations for the nation.

1. Introduction

Climate change has become a severe problem of the era. According to data from NASA and the National Oceanic and Atmospheric Administration (NOAA), each successive decade since the 1980s has been warmer than the one before it, and the 2010s were no exception [1]. In 2019, the average global temperature was about 18 °F higher than the average temperature from the late 1800s, in which more than 80% of the energy produced is from fossil fuels-related greenhouse gas (GHG) emissions [2]. The Intergovernmental Panel on Climate Change (IPCC) shows a trend line that likely eases heat to 35.6 °F quite likely by 2100, with over 39.2 °F warming feasible. Rising temperatures can disrupt the ecosystem, upend long-standing climate norms, and peril all life [3,4]. Eradicating such actions that result in catastrophic damage has been the work of habitats for a long while. Due to heat holding ability, carbon footprint has become a revolutionary cause of global warming and a threat to the environment [5]. Substantially, if a country urges to have a consistent produce avoiding the traumatic effects of climate change, sharper and deeper cuts in GHG emissions are a demand for all energy-related sectors [6].

The primary gases contributing to such devastations include carbon dioxide (CO₂), methane (CH₄), nitrous oxide (N₂O), and sulfur hexafluoride (SF₆) produced by man-made activities, such as burning fossil fuels to generate electricity, power buildings, and deforestation [7,8]. Precisely, midst the other gases, CO₂ reports >80% emissions, thus always manifested [9,10,11]. Globally, fossil fuel-related CO₂ emissions (FFCO₂) had a fall to 36 Gt in 2020, which is approximately 5.1% below that of 2019 emissions. With the 50-year global data comparison from 1970 to 2020, the huge reduction in FFCO₂ was from the transport sector with almost 12%, and the per capita FFCO₂ returned to 2005 levels of about 4.6 t CO₂/capita [12]. Especially in the US, FFCO₂ declined by 1.1% in 2017 but amplified by 3.1% in 2018 and again dropped to 2.4% in 2019, subsequently diminishing to 9.9% in 2020. On average, total FFCO₂ accounted for 4.5 Gt with almost 94% shares from the combustion sources. The 2020 fall owes to falls in oil with 12.6% and coal with 19.1%; the population increase was over 12% compared to 2005 peaks. This warning in the FFCO₂ emissions has roots in the reduction in coal consumption [13].

Considering all the emissions scenarios, the global temperature is said to continue uptrend for at least until mid-century [14]. According to IPCC-2021, unless deep reductions in CO₂ and other GHG happen, the 2 °C global warming is expected to overdo during the 21st century [15]. Subsequently, since the early 19th century, the main cause of the climate crisis has been energy systems. Of these, total CO₂ emissions have accounted for almost 90% of all the other gases, while total GHG is reported for about 75% [16]. Coal’s share of the global electric power output is the highest at 36.37%, making it the most used fossil fuel for power generation, followed by natural gas (23.32%) and oil (3.05%). Fossil fuels together account for 62.75% of the total electricity generated worldwide, as shown in Figure 1A,B [9]. The contribution of CO₂ in the energy sector of the US for 2001–2021 is depicted in Figure 1.

The United Nations has projected 151 countries’ commitment towards the net zero CO₂ pledge in 2050 [17]. Almost 90% of the global GDP and emissions are covered by the CO₂ neutrality vows and net zero pledges [18]. The top contributors, namely China, the US, the European Union, and Japan, accounted for approximately 56% of GHG emissions with a pertinent increase in GHG per capita and carbon intensity per GDP [19]. Considering this, several international agreements and initiatives, such as the familiar Paris Agreement, Kyoto Protocol, and the Conference of the Parties (COP), are slated worldwide [20]. Amongst which the latest and outspoken are the COP26 and COP27 of the United Nations Framework Convention on Climate Change (UNFCCC). Since the first COP in 1992, the climate change and severity of global warming have had no adverse change. Hence, the COP26 held in 2021 agreed to the “Glasgow Pact”, which is a set of commitments by developed countries to increase the level of their emissions reduction targets and to provide financial and technical assistance to developing countries [21]. Another important outcome of the conference was the launch of the “Race to Zero” campaign, which aims to mobilize a global coalition of businesses, governments, investors, and cities to take ambitious action to reach net-zero greenhouse gas emissions by 2050 at the latest [22,23].

Şahin et al. have utilized the fractional nonlinear grey model to forecast CO₂ emissions in the US until 2025 [24]. Namboori et al. used ARIMA, SVM, SVM-PSO, and Prophet models to forecast emissions until 2022 [25]. Silva et al. predicted emissions for 12 months using SSA, ARIMA, Holt-Winters, and EIA-SSA [26]. Steinhauser et al. use the MSFE (Mean Square Forecast Error) to forecast CO₂ emissions until 2030 [27]. Bennedsen et al. have proposed forecasts comparing 2018 and 2021 using the RMSE, MAP, and SADFM (structural augmented dynamic factor model) [28]. Jena et al. created a model using Multilayer Artificial Neural Network, MAP, MAPE, and RMSE and forecasted for the years 2017 to 2019 [29]. Subsequently, to the best of our knowledge, there have been no studies that used regression models to develop models for the forecast of CO₂ emissions. There have not been any studies detailing the top 10 CO₂-emitting states, especially of fossil fuels in the US.

Hence, the objective of this paper is threefold. First, we understand, analyze and compare the total FFCO₂ emissions, total population, and GDP per capita using a regression model. Secondly, we forecast the emissions until 2050 in consequent 5 years intervals. Thirdly, we use MAPE to find the average error comparing the actual values.

2. Energy Sector in the US

The energy sector in the US is an imperative constituent of the country’s economy with a significant quota for the country’s GDP and employment. It is a multifaceted web for most of the CO₂ emissions to process steady energy supplies. Approximately 80% of the country’s energy consumption relies heavily on fossil fuels, including coal, oil, and natural gas.

2.1. GHG Emissions

GHG emissions in the US have been a significant issue for many years, as the country is one of the largest emitters in the world. The United States emitted 6558 million metric tons of carbon dioxide equivalent (MMTCO₂e) in 2019, representing about 13% of the world’s total emissions. The largest source of GHG emissions in the United States is the burning of fossil fuels for energy, which includes transportation, electricity generation, and industrial processes. In 2019, the transportation sector was responsible for the largest share of emissions (29%), followed by electricity generation (25%) and industrial processes (23%), as in Figure 2. Other significant sources of GHG emissions in the United States include agriculture, forestry, and other land use, which accounted for 10% of emissions in 2019 [24]. Although the US is globally the second largest emitter of greenhouse gases, considering the individual fossil fuels, it mostly ranks number one. As per the latest data for 2022 ranking the most polluted countries in the world, China ranks first with 9.9 billion CO₂e while US orders second with 4.4 CO₂e [30]. Of which emissions are more prominent from the usage and burning of fossil fuels. Independently the territorial emissions show a hike of approximately 6.5% for the decade of 2010 to 2021. Of which coal’s share is about 14.3%, oil with 9.3%, but gas shows a reduction of about −0.7%. The highest emissions come from the usage of oil and gas, with 2644 MTCO₂ and 1188 MTCO₂ and next for coal, with 2190 MTCO₂ [23,31,32,33]. In 2020, the topmost contribution is from oil with 2049 MTCO₂, and gas and coal with 1654 MTCO₂ and 879 MTCO₂, respectively. Least emissions from gas flaring and cement with 68 MTCO₂ and 41 MTCO₂. Whereas in 2021, the oil’s share is seen to be the highest with 2234 MTCO₂, with gas taking the next place with 1637 MTCO₂, and coal with 1002 MTCO₂. The least emissions are from gas flaring and cement, with 68 MTCO₂ and 41 MTCO₂ [34]. From 2019 to 2020, the decline in FFCO₂ emissions was 5.6%, but eventually recoiled to comparable levels of 2019 in 2021 [32].

As an important participant in COP, the US has recently re-joined the Paris Agreement, which had been withdrawn by the previous administration, and has announced an ambitious target of reducing its greenhouse gas emissions by 50–52% below 2005 levels by 2030 and set a goal of reaching net-zero emissions no later than 2050 [35]. Emissions from the energy sector sum up to almost three-quarters of total GHG emissions, accounting for about 75% of the country’s total emissions, which is 4554 MtCO₂eq exclusively in 2021 [36]. The sector wise emissions for the top 10 polluted states are shown as in Figure 2A. This includes emissions from electricity generation for both commercial and residential use, transportation, and industrial processes. The overall U.S. CO₂ emissions in 2021 grew by 7% when compared to 2020, but they were still 5%, less than in 2019. The CO₂ emissions of the transport sector increased by 11%, due to the increase in travel, the commercial sector had an increase of 7% owing to the change in the electricity fuel mix and other commercial activities, the residential sector’s emissions hiked by 4%, which was mostly from the fuel mix, and the industrial sector increased by 4% due to the high industrial emission [37].

2.2. Energy-Related Emissions

The fundamental bases of GHG emissions include energy-related emissions, emissions from transportation, electricity generation, industrial process, agriculture, and land use, which ultimately roots in the energy sector [38]. The estimations of state-level energy-related CO₂ emissions include direct fossil fuel use across all sectors, involving industrial, residential, commercial, and transportation, as well as fossil fuels consumed for electricity generation [39]. Energy-related emissions and electricity generation are the largest sources of FFCO₂ emissions which include the combustion of fossil fuels for electricity and heat production, transportation, industrial processes, and power plants that burn coal, natural gas, and other fossil fuels. Transportation is the next highest source involving emissions from cars, trucks, planes, ships, and other vehicles. Releases from the production of cement, steel, and other materials and chemical manufacturing comprise the industrial processes. Additionally, agriculture and land use rank next highest, emitting in the sector with emissions from livestock, soil management, and other agricultural practices, as well as from deforestation and other land use changes. According to EIA, in 2019, energy-related emissions accounted for about 5514 MMT CO₂e, electricity generation for 1745 MMT CO₂e, transportation for 1876 MMT CO₂e, industrial processes for 465 MMT CO₂e, and agriculture and land use for 663 MMT CO₂e. It is worth noting that these figures may have changed since 2019, and the EPA typically updates its GHG inventory on an annual basis [40].

2.3. GDP and Emissions

In comparison to 2020, energy use per dollar of GDP was 1% lower in 2021. Although the growth of energy consumption and GDP are meticulously correlative, it is partly at equilibrium when energy-efficient changes in the economy end in less energy use per unit of economic output. Factors contributing to less energy consumption per capita also involve less energy consumption per dollar of GDP [41]. At low-income levels, CO₂ levels have risen in correlation with GDP per capita, and at higher-income levels, the concentrations have fallen, causing an inverted U shape connection between the income evolution and pollution concentrations [8]. Emissions per GDP had a dip for 2010–2021 with −3.1%, but a rise of 0.8% for 2021. Whilst considering the emissions per capita for 2010–2021, there has been a downfall of about −1.9%, yet a highly increased rate of 6.1% for 2020–2021. Thus, the world is not on track to reach the Paris Agreement goals so the temperatures can reach 2.8 °C by the end of the century. Hence, cutting down emissions by 45 percent to avoid a global catastrophe becomes urgent [42].

The energy infrastructure is broadly eminent as three segments, namely electricity, oil, and natural gas. The electricity segment contains 1075 gigawatts of installed generation with more than 6413 power plants. It uses 48% of coal combustion, 22% of natural gas combustion, and 20% of nuclear power plants to produce electricity, as shown in Figure 2B. The reason for the connection between transport and energy is the hefty dependence on pipelines to distribute products across the nation. Hence, the remaining generation is contributed by hydroelectric plants at 6%, oil at 1%, and renewable resources at 3% [43]. In 2015, the direct and indirect subsidies for oil, gas, and coal reached nearly $649 billion, which is ten times and $50 billion more than the educational and defense spending [44].

3. Fossil Fuel-Related Emissions

The FFCO₂ includes coal, crude oil, petroleum, and natural gas. The principles of such fuels, their trends, drivers, and Carbon Capture and Storage (CCS) are discussed below.

3.1. CO₂ Emissions from Fossil Fuel

CO₂ emissions from fossil fuels are more prominent in the US. The emissions are about 4.57 billion metric tons of CO₂ from energy consumption in 2019, with the vast majority coming from the combustion of fossil fuels. The following are some key fossil fuels that contribute to FFC_O2 emissions. Figure 3 depicts the fossil fuels and sector nexus of CO₂ emissions for 2021.

3.1.1. Coal

Coal combustion is one significant source that owes to FFCO₂ emissions. Coal-fired power plants accounted for approximately 34% of 2019, and 20% of 2020 of the country’s total emissions [45]. Although more than 300 coal-fired power plants retired or were slated for retirement in 2020, the emissions in 2021 were about 10.54 Btu, which is relatively high compared to 2020 emissions of 9.18 Btu [46]. Such plants are replaced with renewables, such as wind, solar, and natural gas. Despite the shift towards renewable energy, coal remains imperative in the energy mix and is sustained for electricity generation, industrial activities, and other purposes [47]. The burning of coal is prone to several health and environmental impacts. Nearly 52,000 premature deaths and $345 billion of damage to the economy were the results of coal-fired power plants, precisely in 2011 [48]. To further reduce the emissions of SF₆, NO₂, and CO₂ during the generation of electricity, the US has managed to transit from coal to natural gas [49]. Production of coal emits CO₂ over 60% from the electricity sector and 13% from the industrial sector. It is visible that there has been a declining trend in recent years. The production trend is depicted in Figure 4A. In comparison to 2005, coal combustion is seen to be reduced from 50% to 34% in 2019. Figure 4B shows the total coal consumption in the US. This is due to the shift from coal-fired power plants to renewable energy resources, the CCS is a technology that captures carbon emissions and stores them underground. Although it is not widely established in the US, it can lessen coal’s legacy of CO₂ emissions.

3.1.2. Crude Oil and Petroleum

Crude oil and petroleum are widely used in the US as primary energy sources for various industrial, transportation, and household purposes. In 2019, almost 2.24 billion metric tons of CO₂ was emitted from petroleum and other liquid fuels, about 49% of the country’s energy-related CO₂ emissions [50]. The transportation sector is the largest consumer responsible for approximately 68% of emissions from these fuels merely in 2020 [48]. Although there are many improvements in vehicle emission standards and fuel quality, transport-related air pollution always remains a hazard. It has adverse effects on respiratory and cardiovascular diseases [51]. The production of crude oil and petroleum is exhibited in Figure 5A. In 2019, the US accounted for almost 21% of petroleum consumption and 14% of CO₂ emissions from petroleum consumption as in Figure 5B [52]. Studies estimate that around $140 billion in annual health damages are caused by such fuels, of which premature death has the highest portion [53]. The Shale revolution and the global oil demand are the two most featured concerns in the oil industry. The boom in shale oil and gas production made possible by hydraulic fracturing (fracking) and horizontal drilling technologies eases a demand for crude oil and petroleum consumption estimated to increase by 1.4 million barrels per day in 2022 driven by economic growth and recovery from the COVID-19 pandemic [54,55,56,57]. Some states have already implemented one type of Zero Emission Vehicle (ZEV), which is a Battery Electric Vehicle (BEV) that uses electric energy [58]. Switching to renewable energy is a remedy to put off the FFCO₂ emissions from crude oil and petroleum and the US has implemented various policies, such as fuel efficiency standards, incentives for electric vehicles and renewable energy, and carbon pricing initiatives.

3.1.3. Natural Gas

The next significant contributor of CO₂ emissions is natural gas. Approximately 1580 million metric tons of CO₂ were found to be emitted, which is about 31% of the country’s energy-related emissions in 2020. This represents a decrease of 2.3% from 2019, owing to a combination of factors [59]. Although natural gas is touted cleaner fossil fuel than coal, crude oil, and petroleum, its carbon footprint is remarkable. Yet, during the pandemic, the production and consumption of natural gas decreased, eventually reducing CO₂ emissions [60]. Methane is the primary component and is a potent GHG, and over a 20-year time frame, it has a much greater warming potential than CO₂ [61]. Another trend that paved the way for CO₂ emission from natural gas is the shift away from coal-fired power generation. The replacement of natural gas-fired plants has increased the share of natural gas in electricity generation. While natural gas-fired power plants emit less CO₂ than coal-fired plants, they still contribute to GHG emissions [62]. Using natural gas for transportation has also been a driver of CO₂ emissions in the US. While natural gas vehicles emit less CO₂ than gasoline and diesel vehicles, they still emit CO₂ and other pollutants [63]. The production and consumption trend line of natural gas is shown in Figure 6A,B. The hike in such emissions is dedicated to the growth of hydraulic fracturing; hence leaks during the production, processing, and distribution of natural gas eventually increase the carbon footprint [64]. To reduce the CO₂ emissions from natural gas, the US has implemented strategies to improve its energy efficiency, increase the use of renewable natural gas produced from organic waste materials and offset traditional natural gas use, and develop CCS technology [65].

4. Modelling and Forecasting Techniques

Data and Methodology

As aforementioned, this study typically involves the data of the topmost CO₂-emitting states in the US. With Texas being the first-highest emitter and Michigan being the tenth-highest emitter, the data were collated from the information available on the United States Environmental Protection Agency webpage. In this regard, we have provided more prominence to the fossil fuels emissions from the Energy sector. Such chosen emitters include the burning of fossil fuel emissions from and using power plants, petroleum and natural gas, refineries, manufacturing industries, minerals, and metals. These emissions are due to the large facilities in the US for the year 2021. All the collected data are reported in million metric tons of CO₂e for all fuel combustions, which includes coal, natural gas, petroleum products, process emissions, and use of sorbent and carbonates.

The first ten highest CO₂ (fossil fuel related) emitting states of the US are as follows: Texas (TX), California (CA), Florida (FL), Louisiana (LS), Pennsylvania (PA), Ohio (OH), Illinois (IL), Indiana (IN), New York (NY), and Michigan (MI). The states’ total fossil fuel-related CO₂ emissions, population per state, Gross Domestic Product Per Capita, and Human Development Index are listed. This data are for 2021, exclusively on the filtered fossil fuel combustions as in

Table 1. National characteristics and emissions of the US states.

Rank	State	Total Emissions (MMT CO2e)	Population (Million)	GDP Per Capita (Thousand $)	HDI *	HDI Rank
1	TX	831	25.6	1831	0.917	32
2	CA	429	37.38	2871	0.936	15
3	FL	269	19.21	1009	0.916	33
4	LS	234	45.61	228	0.893	45
5	PA	298	12.66	714	0.928	26
6	OH	279	11.56	615	0.919	30
7	IL	270	12.7	775	0.934	22
8	IN	243	6.48	353	0.912	36
9	NY	213	19.42	1492	0.944	10
10	MI	206	9.97	473	0.918	31

* HDI ≥ 0.8 → very high; HDI = 0.7 → high; HDI = 0.55~ 0.69 → medium; HDI ≤ 0.54 → low.

.

The selected variables are the Total emissions from fossil fuel (EFF), Total population (TP), and Gross Domestic Product Per Capita (GDP_PC) for each state and are collected from the data freely available on the US official websites [66,67]. First, the main reason for choosing these variables is that their data and time series are exclusively available on the country’s official web pages. Second, they strongly correlate with the consumption and demand of energy resources. Thirdly, as the population rises, the demand and usage of fossil fuels tend to increase gradually, owing to the amplified human activities. Moreover, due to an upsurge in GDP per capita, fossil fuel emission increases intensifying associated energy demands [68]. The descriptive statistics of the variables were derived using the SPSS tool.

The descriptive statistics of the chosen variables are depicted in

Table 2. Descriptive statistics.

Variables	Statistics	TX	CA	FL	LS	PA	OH	IL	IN	NY	MI
** EFF	Mean	830.95	428.91	269.45	234.32	298.23	278.64	269.77	242.64	212.82	206.18
	Range	165.00	188.00	157.00	168.00	80.00	100.00	92.00	98.00	73.00	63.00
	Maximum	888.00	472.00	297.00	381.00	328.00	322.00	294.00	273.00	246.00	232.00
	Minimum	723.00	284.00	140.00	213.00	248.00	222.00	202.00	175.00	173.00	169.00
	Standard Deviation	37.38	40.60	31.28	34.35	24.54	34.50	25.27	29.22	20.58	16.68
	Variance	1397.19	1648.18	978.55	1179.66	602.18	1190.43	638.37	853.96	423.58	278.16
	Skewness	−1.29	−2.24	−3.59	4.03	−0.39	−0.12	−1.55	−0.91	0.15	−0.09
	Kurtosis	2.40	7.29	15.25	17.66	−1.07	−1.62	2.00	0.00	−1.08	−0.54
** TP	Mean	25.60	37.38	19.21	45.61	12.66	11.56	12.70	6.48	19.42	9.97
	Range	9.39	5.52	6.20	3.79	0.73	0.43	0.46	0.74	1.10	0.19
	Maximum	30.29	39.50	22.24	46.81	13.01	11.79	12.89	6.83	20.10	10.06
	Minimum	20.90	33.98	16.04	43.02	12.28	11.36	12.43	6.09	19.00	9.87
	Standard Deviation	2.93	1.80	1.85	1.02	0.22	0.12	0.14	0.23	0.29	0.06
	Variance	8.57	3.23	3.41	1.04	0.05	0.02	0.02	0.05	0.08	0.00
	Skewness	−0.08	−0.41	−0.03	−0.92	−0.31	0.26	−0.35	−0.21	0.44	−0.18
	Kurtosis	−1.31	−1.21	−1.04	0.39	−0.89	−0.80	−0.97	−1.06	−0.35	−1.20
** GDPPC	Mean	1386.58	2166.95	819.45	230.57	634.73	553.49	713.15	302.87	1281.48	439.26
	Range	835.70	1179.10	365.98	42.05	177.38	112.45	136.93	93.15	402.55	90.19
	Maximum	1831.36	2871.42	1008.69	247.77	716.17	615.42	777.65	352.62	1494.74	473.33
	Minimum	995.66	1692.32	642.71	205.72	538.79	502.97	640.72	259.47	1092.19	383.14
	Standard Deviation	269.50	352.47	97.38	10.99	55.53	34.39	42.81	25.13	132.91	21.68
	Variance	72,630.76	124,233.01	9482.55	120.77	3083.31	1182.45	1832.83	631.62	17,663.9	469.85
	Skewness	0.14	0.52	0.12	−0.75	−0.19	0.32	−0.21	0.07	0.12	−0.76
	Kurtosis	−1.28	−0.74	−0.38	0.56	−1.12	−1.06	−0.91	−0.52	−1.32	0.81

** EFF—Emissions from fossil fuel (MMT CO₂e); TP—Total Population (millions); GDP_PC—Gross Domestic Product Per Capita (Current USD ($)).

. The results show that each country’s mean is reasonably high compared to the standard deviations. For EFF the means of CA and TX are 428.9 and 830.95, while their standard deviations (SD) are 40.60 and 37.38, respectively. TP, TX, and FL show means of 25.60 and 19.21 with SD of 2.93 and 1.85, respectively. For GDP_PC, CA, and TX’s means are 2166.95 and 1386.58 with SD of 352.47 and 269.50. The difference is comparatively high, exhibiting most related variation defined by CV and a good scatter plot. It is also understood that TX and CA have high relativity in terms of EFF, TP, and GDP_PC. For almost every data set, we see that the mean is relatively higher than the standard deviation. This shows that the data are homogenous and consistent. The points when plotted, will be tightly clustered around the mean and will not exhibit much variability.

The general idea value for kurtosis and skewness to have a normal distribution for an observed series should be zero. Yet, some researchers have argued that it can be considered a normal distribution if both values are between ±1.5 [69,70,71]. Thus, the skewness and kurtosis values show a normal distribution. Specifically, EFF and TP show negative skewness which means that the values are skewed towards the left, indicating a longer tail on the left side of the distribution, whereas GDP_PC indicates positive skewness demonstrating a fatter tail on the right side of the distribution. Additionally, from kurtosis, we see all values less than the normal value (≤3), which suggests the kurtosis curve is platykurtic. Except for CA, FL, and LS in EFF, which advises the kurtosis curve to be leptokurtic as the values are >3.

Furthermore,

Table 3. Results of the correlation analysis.

State	TX			CA
Variables	EFF	GDPPC	TP	EFF	GDPPC	TP
EFF	1	0.82	0.98	1	0.63	0.69
GDPPC	0.82	1	0.99	0.63	1	0.93
TP	0.98	0.99	1	0.69	0.93	1
State	FL			LS
Variables	EFF	GDPPC	TP	EFF	GDPPC	TP
EFF	1	0.67	0.79	1	0.89	0.77
GDPPC	0.67	1	0.92	0.89	1	0.98
TP	0.79	0.92	1	0.77	0.98	1
State	PA			OH
Variables	EFF	GDPPC	TP	EFF	GDPPC	TP
EFF	1	−0.87	−0.86	1	−0.83	−0.93
GDPPC	−0.87	1	0.94	−0.83	1	0.91
TP	−0.86	0.94	1	−0.93	0.91	1
State	IL			IN
Variables	EFF	GDPPC	TP	EFF	GDPPC	TP
EFF	1	−0.66	0.88	1	−0.81	−0.87
GDPPC	−0.66	1	0.64	−0.81	1	0.94
TP	0.88	0.64	1	−0.87	0.94	1
State	NY			MI
Variables	EFF	GDPPC	TP	EFF	GDPPC	TP
EFF	1	−0.89	−0.89	1	0.99	0.98
GDPPC	−0.89	1.00	0.86	0.99	1	0.99
TP	−0.89	0.86	1	0.98	0.99	1

outlines the findings of the correlation analysis. The data of all the independent values show a high degree of correlation over 0.80 compared to the dependent variable. It is observed that EFF, GDP_PC, and TP have statistically significant correlations with the selected 10 states of the US. The data of all variables are between 2000 and 2021. The training and testing period uses the same range of data. The multiple regression tool from the SPSS statistical program application predicted the models. The suitable linear regression method was the backward multiple linear regression analysis methods, which consequently removes the values with the lowest partial F-value after including all variables for calculation. A model is created whenever the statistical contribution of the discarded variable is high. After the models are estimated, the goodness-to-fit is determined using various statistical methods. An adjusted coefficient of determination (adj. R2) was used to verify how well the model fits the data. F-test was carried out to test the presence of significance between the independent and dependent variables. Additionally, testing the influence of each coefficient of the model, the t-test opted. The actual and predicted data plots were used to study the linearity further. Finally, to check the accuracy of the forecasted values, mean absolute percentage error (MAPE) was adopted. The mathematical expression of MAPE is as defined:

$$\mathrm{MAPE}=\frac{1}{n}\sum _{i=1}^n\frac{\left|A_i-\left.F_i\right|\right.}{A_i}$$

$A_i$ is the actual value; $F_i$ is the forecast value; n is the total number of measurements.

5. Results

The developed regression models used to forecast the EFF are shown in

Table 4. Model development.

State	Models
TX	TP = (14.613) − (0.004) · EFF + (0.010) · GDPPC
CA	TP = (25.757) + (0.002) · EFF + (0.005) · GDPPC
FL	TP = (7.361) − (0.005) · EFF + (0.016) · GDPPC
LS	TP = (126.161) – (0.175) · EFF − (0.192) · GDPPC
PA	TP = (1539.156) − (98.143) · GDPPC
OH	TP = (11.219) − (0.002) · EFF + (0.002) · GDPPC
IL	TP = (11.236) + (0.002) · GDPPC
IN	TP = (5.258) − (0.002) · EFF + (0.006) · GDPPC
NY	TP = (20.828) − (0.01) ·EFF + (0.001) · GDPPC
MI	TP = (8.923) +(0.001) · EFF + (0.002) · GDPPC

. The contribution of each variable is also studied to evaluate the effectiveness of the developed models. Hence, the contribution rates are computed as shown in

Table 5. Contribution percentages of variables to the developed models.

Variables	EFF (%)	GDPPC (%)	TP (%)
TX	22.23	5.15	9.04
CA	10.57	6.15	20.72
FL	8.60	8.41	10.84
LS	6.82	20.98	43.84
PA	–	11.43	57.47
OH	8.06	16.09	96.26
IL	–	16.7	90.8
IN	8	12.05	29.39
NY	8	9.64	66.92
MI	10	20.26	166.17

. We see the highest contribution of about 22.23% of EFF from TX, 20.98% of GDP_PC from LS, and 166.17% of TP from MI. It is seen that GDP_PC and TP show high contributions among the variables, on average LS and MI show higher levels of contributions when compared state-wise. The results of the data testing using the F-test, t-test, and adj. R² are shown in

Table 6. Verification of the statistical results.

State	Independent Variables	Standard	tcalculated	ttable	Fcalculated	Ftable	R2
		Error of Estimation
TX	Constant	0.40738		2.093	489.434	3.52	0.98
	EFF		99.592
	GDPPC		23.937
	Constant	0.72275		2.093	55.771	3.52	0.85
CA	EFF		43.928
	GDPPC		28.474
	Constant	0.70982		2.093	55.385	3.52	0.85
FL	EFF		36.581
	GDPPC		39.205
	Constant	12.04626		2.093	18.21	3.52	0.98
LS	EFF		25.710
	GDPPC		79.265
	Constant	12.73524		2.086	57.971	3.52	0.99
PA	EFF		54.175
	GDPPC		52.739
	Constant	0.03269		2.093	129.239	3.52	0.92
OH	EFF		36.192
	GDPPC		74.156
	Constant	0.10978		2.086	13.649	3.52	0.99
IL	EFF		47.664
	GDPPC		76.899
	Constant	0.06591		2.093	108.245	3.52	0.92
IN	EFF		37.659
	GDPPC		55.778
	Constant	0.12201		2.093	48.846	3.52	0.84
NY	EFF		43.525
	GDPPC		44.623
	Constant	0.04222		2.093	11.256	3.52	0.99
MI	EFF		55.236
	GDPPC		93.064

. The tabulated F value is smaller than the calculated F value, indicating that all equations are significant at 95% confidence levels. The tabulated t values are higher than the calculated t values, which signify the coefficients used. The R² values are almost 0.98 and 0.99, which indicates high relativity.

The forecasted EFF, as shown in

Table 7. Forecasted CO₂ emission.

Projected Period	Forecasted EFF (MMT CO2)
State	TX	CA	FL	LS	PA	OH	IL	IN	NY	MI
2025	812.0	367.4	255.4	198.3	236.5	198.6	215.00	177.04	161.67	160.60
2030	819.4	339.4	238.9	193.2	207.8	162.8	186.00	144.79	138.05	141.12
2035	822.8	319.4	231.3	183.2	183.3	134.0	159.20	114.16	120.91	124.41
2040	809.8	292.1	221.8	175.6	160.5	106.7	134.66	90.07	100.56	105.40
2045	814.7	274.1	213.8	165.7	138.7	76.3	118.63	66.81	82.95	91.28
2050	814.5	248.3	202.1	158.3	112.5	45.4	89.47	36.02	62.35	72.26

, indicates that the emissions have eventually decreased and increased for every consecutive year. It is seen that the forecasted emissions for CA have reduced from 2025 to 2050. The average reduction is from 367.4 MMT to 248.3 MMT, which is a 119.1 MMT variance. More drastic change is seen in the states of OH and IN, where the difference is found to be 153.2 MMT and 141.02 MMT, respectively. An increase is found in TX with 2.5 MMT of emissions. However, other states show good decrease in the emissions of fossil fuels with an average of 60–70%. This indicates a positive measurement of the forecasted values and future emissions reduction in the US. As the major states’ emissions decrease, reaching net zero by 2050 is not a challenge. The MAPE model results shown in

Table 8. Performance methods for the proposed models.

State	MAPE (%)	Forecast Accuracy for the MAPE
		MAPE ≤ 10 ⇒Highly accurate; 11 ≤ MAPE ≤ 50 ⇒Mostly Accurate; 21 ≤ MAPE ≤ 51 ⇒Reasonable; MAPE > 51 ⇒Inaccurate [65]
TX	3.99	Highly accurate	MAPE ≤ 10
CA	6.52
FL	17.67
LS	5.94
PA	12.17	Mostly accurate	11 ≤ MAPE ≤ 50
OH	14.04
IL	4.13	Highly accurate	MAPE ≤ 10
IN	8.81
NY	9.61
MI	11.48

suggest that the forecasted EFF is either highly or mostly accurate, as shown in Figure 7. TX, IL, LS, and CA show high accuracy, while PA and OH show good accuracy. This is an excellent fit model to suggest accuracy.

6. Current and Future Mitigations

Overall, the US is taking a multi-faceted approach to mitigating CO₂ emissions, with a focus on transitioning to a more sustainable, low-carbon economy. There have been several measures to mitigate CO₂ emissions and combat climate change in the country.

• Increased Deployment of Renewable Energy Sources—The deployment of renewable energy sources, such as wind and solar power, has been an initiative to reduce emissions. According to a study conducted by the National Renewable Energy Laboratory, renewable energy has the potential to meet more than 80% of the country’s electricity demand by 2050 [72]. Such a strategy is effective in other countries, such as Germany, which has rapidly increased its renewable energy capacity in recent years [73]. Policy measures, such as tax incentives and renewable energy mandates, can help to encourage the adoption of renewable energy technologies.
• Stricter Emissions Standards for Vehicles and Industries—The US has implemented more stringent emissions standards for vehicles and industries. In particular, the Environmental Protection Agency (EPA) has set targets for reducing emissions from power plants. According to a study by the American Lung Association, these regulations have resulted in significant reductions in air pollution and associated health benefits [66]. Corporate Average Fuel Economy (CAFE) standards for vehicles are another approach to reducing emissions. These standards require automakers to meet minimum fuel economy targets for their fleets, which has led to the development of more fuel-efficient vehicles that emit less CO₂ [74]. A study by the International Council on Clean Transportation found that the CAFE standards have led to a reduction of about 6% in CO₂ emissions from light-duty vehicles [75].
• Carbon Capture and Sequestration Technologies—The US is investing in CCS, which allows for capturing and storing CO₂ emissions from industrial processes. The Global CCS Institute studied that there are currently 28 large-scale carbon capture and storage projects operating or under construction worldwide, with the majority located in the US [76]. Among the 28 large-scale CCS projects, the United States has a renowned project named “The Petra Nova Carbon Capture Project”. Having begun operating in 2017, it has shown some notable results. Firstly, it has successfully captured approximately 90% of CO₂ emissions from the flue gas of the coal-fired power plant unit at the W.A. Parish Generating Station [77]. Secondly, it has shown that post-combustion carbon capture technology can be integrated into existing coal-fired power plants without significantly reducing their efficiency. Thirdly, the captured CO₂ from the Petra Nova project is being used for enhanced oil recovery (EOR) at an oil field located approximately 82 miles away from the W.A. Parish Generating Station. This process has been shown to increase the recovery of oil from depleted oil reservoirs, which can have economic benefits. Finally, it has provided valuable information about the costs and technical feasibility of post-combustion carbon capture technology [78]. Another potential strategy is implementing carbon capture and storage (CCS) technologies in fossil fuel power plants. This involves capturing CO₂ emissions from power plants and storing them underground or in other long-term storage facilities. CCS is effective in reducing CO₂ emissions in pilot projects around the world [79].
• Shift Towards Sustainable Transportation Options—The US is shifting towards more sustainable transportation options, such as electric vehicles and public transit systems. According to a study by the International Council on Clean Transportation, adopting electric vehicles (EVs) can significantly reduce CO₂ emissions from the transportation sector [80]. The International Energy Agency in 2020 reported that EVs accounted for around 2.4% of all passenger cars on the road in the US. The sales of EVs have been increasing in recent years, with 2020 seeing a slight uptick in sales despite the challenges posed by the COVID-19 pandemic, and there are predictions from industry analysts that the sales will continue to grow in the coming years [81]. Studies suggest that EVs emit 50–70% less GHG than conventional gasoline vehicles. This emission reduction is even more remarkable when the electricity used to power the EVs comes from renewable resources [82]. E-bike sharing programs, electric vehicle charging stations, and a public transit system that runs on renewable energy are a few steps towards a greener society by the Seattle government [83].
• Public Awareness and Education—The US also focuses on public awareness and education on the importance of reducing CO₂ emissions and combating climate change. According to a study by the Yale Program on Climate Change Communication, there has been a significant increase in public awareness and concern about climate change in recent years [84]. Finally, there is also a need to promote energy efficiency and conservation measures in industries and households. This could include policies, such as building codes that require energy-efficient buildings, incentives for using energy-efficient appliances, and education programs to promote energy-saving behaviors [85].

The Clean Air Act, the Clean Power Plan, and the Paris Agreement are a few other initiatives of the government. However, the country’s GHG emissions have been relatively stable since 2005, with a slight decrease in recent years. To achieve significant reductions in GHG emissions, further action will be necessary, including transitioning to cleaner energy sources, increasing energy efficiency, and reducing emissions from transportation and industry. The conferences of the parties (COP) are held by the UNFCCC annually to discuss and negotiate global action on climate change. The 26th COP (COP26) was scheduled in Glasgow, UK, in November 2020 but was postponed due to the COVID-19 pandemic. The latest COP meeting was held in Sharm El-Sheikh, Africa, in 2022 under the central themes of “together for implementation” and “net zero”. The takeaways of the summit include the funds established to support “loss and damage” for countries suffering severe climate changes, the planned implementations discussed at COP26 were not achieved, and the chances of achieving them in 2050 are less fortunate, as there is a need to reform the broader public financial system, revisit and strengthen their 2030 targets, to accelerate renewable energy deployment, global stocktake, and nature-based solutions [86,87]. Overall, the COP conferences are important forums for global action on climate change, and COP26 and COP27 are crucial conferences for advancing progress toward the Paris Agreement goals.

7. Discussion and Future Direction

The results of the study indicate that the carbon emissions from fossil fuels gradually decrease over time. As far as we know, there has been no detailed study on fossil fuel CO₂ emissions in the US. As per the IPCC forecasts, we understand that the temperature will hike until 2030 if it continues in the same trend and after severe mitigations, the emissions might fall, taking a downward drift [88]. Other studies have forecasted emissions for either 12 to 36 months or until 2030. Yet, from our study, we conclude that the emissions from fossil fuels can be reduced to a larger extent by switching to renewable energy or energy mix. This can be taken as a series of mitigation which will reduce carbon emissions, which is a serious threat to the environment. Despite what the other literature says on the forecast, our results indicate that the US CO₂ emissions from fossil fuels will eventually decrease with the increase in the use of renewable energy. We see that the emissions rate is gradually declining, which is a good sign that the country is working towards the net-zero policy. Adapting to green energy is a strong mitigation to see a drastic change in CO₂ emissions. While comparing our results with the literature, there are similar and more diverse results found.

Previous studies by Jacobson et al. discuss the potentiality of the US to transition to 100% renewable energy WWS (Wind, Water, and Sunlight) by 2050 and the implications for reducing CO₂ emissions [89]. Similarly, Ding et al. studied CO₂ emissions of the US as one of the countries using the Choquet fuzzy integral and grey multivariate delay model. Forecast results show that the emissions in the US will be reduced by 5.08% in 2025 using the Green Economic Development Strategy [90]. Wang et al. projected the emissions for 2020 and 2021, which showed a rise. The proposal of the study is if the Zero Carbon Action Plan adopted by the country in 2020 works well, the US emissions and the results of the global carbon reduction will be lower and better [91]. Şahin et al. use ROFANGBM to predict emissions in the US and show an increase in 2025 compared to 2020 [24]. Bennedsen et al. used a structural augmented dynamic factor model to compare 2018 and 2019 CO₂ emissions in the US and projected a reduction of about 2.6% per capita CO₂ emissions compared to 2018 [28]. Our study uses the regression model to analyze and create forecast models. In the future, other methodologies can be used or even changing the independent variables might result in different results. Different countries or other states of the US can be forecasted as our study details only the top 10 emitting states. Such forecasts might help the country to be cautious about the forthcoming.

8. Conclusions

The issue of CO₂ emissions in the United States is complex and multifaceted. While there have been some efforts to reduce emissions, such as the introduction of renewable energy sources and energy-efficient technologies, the US still heavily relies on fossil fuels for energy production, transportation, and manufacturing. Additionally, the US has a high consumption and waste level, contributing to carbon emissions. One of the main reasons for the US’s high level of carbon emissions is its economic and political structure [92,93,94,95]. The country has historically prioritized economic growth and development over environmental concerns, with many industries and politicians resistant to change. Additionally, the US has not always been willing to participate in international agreements to reduce greenhouse gas emissions, such as the Kyoto Protocol [96,97]. However, there has been some progress in recent years. Several states and cities in the US have set ambitious goals to reduce carbon emissions, and there has been an increase in renewable energy production.

The current study has discussed the emissions of fossil fuels in the US, especially for the top emitting states. We have forecasted and verified the error percentage to confirm the data further. Despite many studies on the forecasting techniques of CO₂ emissions, there has been no study on fossil fuel emissions in the US energy sector. Although this study discusses EFF in the US, the major limitation is that it does not commit with all states and all sectors in the United States. In the future, concentrating on a broader angle to forecast and compare is a good option. Even though the emissions have increased in all sectors, they have never surpassed the 2019 emission rate. Thus, it is notable that the United States has a hard time meeting anticipated national energy efficiency, utilizing renewable energy sources, reducing carbon emissions, and achieving a decarbonized economy. Overall, the issue of carbon emissions in the US is a complex and ongoing challenge that requires sustained effort and collaboration from all sectors of society. Our study is a notable entity as it forecasts EFF and shows the ways to reduce the carbon footprint to set a goal of achieving a 100% clean energy economy by 2050.