The Impact of Climate Change on Energy Consumption on Small Tropical Islands

Abstract

The anthropic causes of climate change are well known, but the influence of climate change on society needs to be better estimated. This study estimates the impact of climate change on energy consumption on small tropical islands using monthly temperatures and energy production/consumption statistics during the last decades. Here, we show, using energy, meteorological, demographic, and economic datasets, as well as statistical correlations, that energy consumption is sensitive to (i) cyclonic activity and (ii) temperature warming. On small tropical islands, increased electricity consumption correlates with temperatures rising above 26 °C in relation to air conditioner electricity consumption. On La Réunion Island, a +1 °C increase is expected to cause an electricity production of 1.5 MWh/inhabitant per year, representing a growth of 3.2%. Considering that non-renewable sources are primarily used to produce electricity, this feedback contributed significantly (i.e., 2000 to 4000 TWh) to the greenhouse gas increase caused by climate warming over the last decades on tropical islands. Demographic and wealth variations, as well as socio-economic crises, also have a significant impact on energy consumption (2 kWh for 1000 inhabitants, 0.008 GWh/inhabitant growth for a 10,000 GDP/inhabitant growth, and a 0.2 GWh/inhabitant decrease during COVID-19, for annual consumption, respectively) and must be taken into account for decadal variation analysis. The relationship between climate change and energy consumption in tropical areas should be better integrated into climatic scenarios to adapt building isolation and energy production.

1. Introduction

Multiple environmental [1], climatic [2,3], social, and economic crises [4,5] have been described during the last decades. The observed climate change is due to anthropogenic factors [6]. Greenhouse gases, primarily from the production and use of carbon-based energy sources (coal, gas, and oil), are the primary cause of climate change [6]. The consequences of increases in greenhouse gases are temperature increases, sea level rises, and more intense extreme events [7,8]. If global warming is a consequence of anthropic activities and energy consumption, global warming could also have increasing consequences on energy consumption and on our way of life [9]. However, the impact of climate warming and extreme events on energy production and consumption is poorly quantified. This study focuses on small tropical islands to investigate how climate change and cyclonic activity affect energy production and consumption. These territories are sensitive to climate change and are suitable for studying the historical impacts of extreme weather events on energy production and consumption. This approach is the first step toward anticipating the consequences of climate warming on energy production and consumption.

Energy production is significant and has increased over the last centuries [10]. Many social and economic activities require the use of energy, and energy production and consumption are useful to document human activities. Energy consumption variations have been monitored for decades, and observations show that variations depend on (i) the hour of the day, (ii) the day of the week, specifically between workday and weekend, and (iii) the season [11]. Energy consumption is closely linked to our lifestyles [12,13]. Energy is used for (i) production (food, industry, raw material extraction, etc.), (ii) transportation, (iii) office activities (light, computers, building heating and air conditioning), and (iv) homework and way of life [14,15].

According to Hall and Klitgaard (2011) [16], one indicator that may be used to describe the evolution of human activities is the amount of energy produced and consumed. For example, energy consumption is known to have changed significantly during the COVID-19 pandemic crisis in 2020 [17,18,19,20,21,22], subprime financial and economic crises, or in the case of natural disasters [22,23,24,25]. The Human Development Index (HDI) and the Night Light Development Index (NLDI) [26] are correlated.

The concentration of CO₂ in the atmosphere is increasing along with energy production (Figure 1) [27], particularly with regard to hydrocarbon use [6]. Energy production causes greenhouse gas emissions that are responsible for climate warming. Renewable energy has a lower impact on greenhouse gas emissions because little or no carbon dioxide emissions are directly produced. The production of energy is criticized for several reasons, including (i) pollution and greenhouse gas emissions, (ii) nuclear risks, (iii) geopolitical dependence, (iv) resource scarcity, (v) resource costs, and (vi) the production of socio-economic pathologies [4,28,29,30,31].

Small tropical islands are particularly sensitive and vulnerable to natural hazards [33]. The adaptation of energy demand while the effects of climate change are increasing is difficult. Reducing energy consumption may have significant implications for social and economic evolution, but sustainability is also a necessary objective. The challenge for small tropical islands is to balance climatic vulnerability with energy demand. Small tropical islands are hyper-reactive laboratories for global warming. The fundamental objective of this research is to better understand the complex connection between climate change and human activities. Studying small tropical island territories provides an opportunity to carry out a detailed analysis of the interactions between a territory and its environment. Energy has been shown to influence climate, but climate can also influence energy production and consumption [23,34]. More specifically, the small tropical islands of La Réunion, Guadeloupe, Martinique, Saint-Martin, Saint-Barthélemy, Mayotte, French Polynesia, and Nouvelle-Calédonie were studied using their energy consumption/production. The specific objective of this research is to assess the relationship between climatic variation (temperature and cyclone occurrence) and energy consumption/production in these territories. The comparison of these territories is useful to distinguish the impact of socio-economic factors (demographic growth, wealth, health crises, political evolution) and those related to climate (temperature change, extreme events) on energy production and consumption. This research contributes to the quantification of electricity production increases caused by climate warming in tropical areas during the last decades. The environmental energy sciences will advance by comparing the estimates of datasets from various regions and at the global level.

A better understanding of how climate affects energy consumption on these islands is used to assess and discuss their vulnerability to climate change. The main questions that were investigated are as follows: (i) How do extreme climatic events and temperature increases affect energy production? (ii) Do cyclones leave significant traces in energy production, and how? (iii) Can the effects of global warming on energy production be estimated and anticipated for the next decades in tropical areas?

2. Materials and Methods

2.1. Energy Consumption: Variation and Comparison

Energy production and consumption can be described using various parameters. The primary energy consumption and electricity production have been investigated to analyze the climatic impact. Primary energy consumption provides information on transport and industrial production, whereas electricity production is an efficient way to record energy consumed by buildings.

The variation in electric energy consumption highlights changes in the type of activity, including domestic and professional activities, or their intensity/frequency [11,12,16]. Individual and collective practices, such as the use of air conditioning, professional and domestic appliances, or an increase in the number of electric cars, can influence the evolution of electricity production.

The influence of different factors (i.e., number of inhabitants, GDP/inhabitant, temperature variation, cyclone occurrence) on energy production and consumption are presented (Section 3) to discriminate what variations are caused by climatic events from those triggered by other causes. Demographic changes and wealth evolution are expected to have an impact on energy production and consumption. These potential biases must be considered when analyzing the impact of climate change on energy.

Tropical areas are particularly sensitive to climate warming because their mean temperature is high, and extreme weather events, such as cyclones, occur. Small tropical islands, disconnected from other electrical networks, are vulnerable to climatic variation and their energy production must be adapted locally to their specific needs. More specifically, the expected energy consumption must be adapted to future temperature variations. Data about energy consumption, temperature variation, demography, and wealth are not always available and collected homogeneously in tropical areas. The small tropical islands selected are located on various longitudes and latitudes, but the datasets concerning their energy consumption, temperature, demography, and socio-economic characteristics were collected homogeneously by French institutes (INSEE, Météo France, IEDOM). Despite their differences, these islands are almost of the same size and with the same number of inhabitants (<1 million).

Seasonal temperature variations and cyclonic impacts are analyzed for a 5-year period, from 2017 to 2022, using monthly electricity production data (the temperature of the years 2009 and 2010 are shown for comparison), whereas climate warming or socio-economic parameters are analyzed for a longer time using yearly electricity production data.

Similarly, the impact of climate warming and more intense temperatures is examined over a longer period, from 1981 to 2022, using a dataset with an annual step. First, a comparison of primary energy production and electricity consumption between small tropical islands (i.e., French Polynesia, Guadeloupe, La Réunion, Martinique, Mayotte, Nouvelle-Calédonie, Saint-Barthélemy, and Saint-Martin) is presented to decipher the influence of the number of inhabitants (Section 3.1). An estimation of the mean value for the period 2010–2020 for inter-island comparison is estimated for (i) primary energy consumption, (ii) electricity production, (iii) population, and (iv) GDP. Electricity production is used rather than electricity consumption because some consumption may be due to technical issues that are beyond the scope of this study and because electricity production is often the only data available.

Second, the comparison of the temporal evolution of electricity production and population growth is presented over a longer time period, from 1975 to 2020, to highlight the contemporaneous changes that occurred in La Réunion using yearly data. The temporal change in energy is used and analyzed.

Third, a comparison of the evolution of electricity consumption between the islands of La Réunion, Saint-Martin, and Mayotte is analyzed. Comparing changes in electricity production is a useful method to assess the impact of specific events (climatic, cultural, and political) on society. The impact of socio-economic change (political change, COVID-19) has been investigated over the last 20 years (from 2006 to present) with a yearly step. To compare the electricity production on the islands of La Réunion, Saint-Martin, and Mayotte, the energy consumption was normalized by the number of inhabitants for each island. The energy consumption was also normalized by the amount of electricity produced in 2010 to better analyze recent changes (Section 3.2). This second normalization enables a more accurate graphical representation and an easier comparison of the changes that occurred after 2010.

Fourth, the impact of rising temperatures on electricity production was estimated by comparing temperature variations to electricity production on La Réunion (Section 3.3). The data were obtained from Météo France (https://meteofrance.re/fr, accessed on 9 September 2024) for each month over a 6-year period. The mean monthly temperature on La Réunion from 2017 to 2022 was calculated. Temperatures above 26 °C (the coldest monthly temperature in June) on La Réunion Island were calculated for each month. Using data from the La Réunion Energy Observatory (Observatoire de l’énergie de La Réunion; https://oer.spl-horizonreunion.com/), the average amount of additional electrical energy for each month from 2017 to 2022 was estimated in comparison to the one in June. Electricity production is used instead of primary energy production because it is more sensitive to temperature fluctuations. For each month, the temperature excess is correlated with the excess of electricity production. The error represents the standard deviation of temperature and electricity production.

The impact of rising temperatures on electricity production was also estimated on a yearly scale. Deviations from the average yearly temperature were calculated by estimating the average temperature value from 1991 to 2020 (i.e., 30 years). The average decadal (30-year) value was then subtracted from the yearly temperature. The annual deviation from the average electricity production was calculated by estimating the mean decadal electricity production from 1991 to 2020. The difference between annual electricity production and the average decadal (30-year) value was calculated. Finally, these two deviations from the mean, for temperature and electricity production, were plotted on a graph spanning the years from 2000 to 2022.

2.2. Data Sources and Analysis

Comparing small tropical islands is useful to detect potential anomalies and mitigate bias. Cross-validation and the standardization of data (normalizing by the number of inhabitants or by a reference year) were performed before interpretation. Studying small tropical islands has some other benefits. There are few exports due to low internal production and energy is consumed for local needs. However, there is also a limitation: imports of primary energy are only counted for oil or gas, but not for the energy required to produce imported products, resulting in underestimates of the total energy consumption required to maintain the way of life on these islands. The total energy consumed in these territories, as a result of imported production (food, household appliances, automobiles, etc.), is not calculated using primary energy consumption.

GDP/inhabitant is considered a proxy of wealth. The GDP was estimated by French statistical institutes (INSEE, IEDOM, IEOM). An average value of GDP/inhabitant was used in this study for the period from 2010 to 2020 for comparison. When only one value is available for the GDP during this period, this value has been used, and the year of measurement is indicated (Table 1). An uncertainty of ±500 EUR/inhabitant was considered in this study for the GDP/inhabitant.

The data used in this study are mainly obtained from Electricite de France Systemes Energetiques Insulaire (EDF SEI) [35,36,37,38,39,40,41,42], the La Réunion Energy Observatory (Observatoire de l’Energie de La Réunion) [43,44,45,46,47], and data collected by IEDOM or IEOM (French institutes specializing in the synthesis of French overseas territories’ economies).

The number of inhabitants is estimated by the French Statistical Institute (INSEE). Some difficulties in the demographic estimation may arise during crisis events, such as the months following a cyclone or when a significant migration in or out occurs. However, several indicators, such as food and water consumption or number of children enrolled in schools, are used by INSEE to avoid overestimation or underestimations.

A correlation between the investigated parameters is established. Linear correlations f(x) = ax, where a is a constant, and x and f(x) are variables, are calculated to quantify the potential relationship between variables x and f(x). The fit’s accuracy is estimated by calculating the standard error, the variance of freedom, and the root mean square σ. The linear trends ax are presented ±2σ to highlight the expected values.

2.3. Small Tropical Islands Context

Isolated environments disconnected from the rest of the energy network were chosen in order to obtain the most comprehensive energy assessment possible. Characterizing variations in energy consumption and their correlation with parameters such as demographic growth, wealth variation, and climatic changes, all other things being equal, is easier and more controlled and accurate in small isolated territories than on a global scale. Furthermore, small tropical islands could be more vulnerable to climatic warming due to the occurrence of temperatures >30° as well as extreme events, such as cyclones or heavy rain.

This study focuses on (i) islands with independent electricity production that are disconnected from other territories, (ii) tropical environments with relatively high mean temperatures (>20 °C) and the possibility of extreme events such as cyclones, and (iii) territories administered by France, because they have a relatively similar data collection process.

More precisely, the territories studied are Saint-Martin, Saint-Barthélemy, Guadeloupe, Martinique, La Réunion, Mayotte, French Polynesia, and Nouvelle-Calédonie. Saint-Martin [48,49,50,51], Saint-Barthélemy [49,52,53,54,55,56], Guadeloupe [57,58], and Martinique [47,51] are located in the Caribbean. La Réunion [59,60,61,62] and Mayotte [63,64,65,66,67,68] are located in the Indian Ocean. French Polynesia [46,69,70,71] and Nouvelle-Calédonie Island [72,73,74,75,76] are located in the Pacific Ocean. An accurate presentation of these territories is beyond the scope of this study, and only basic information will be given here.

Except for Nouvelle-Calédonie Island, all of these islands are volcanic in origin [77,78,79,80]. Volcanic activity has been consistently active in the last centuries on several of these islands (Guadeloupe, Martinique, Mayotte, La Réunion). Present or recent volcanic activity could be favorable for producing geothermal energy. However, Guadeloupe is the only island in our set to produce electricity using geothermal energy. A non-negligible part of the electricity of La Réunion Island is produced using hydroelectric plants. The amount of renewable energy produced on these islands is increasing slowly. Despite the potential of renewable energy production, the energy dependence of these islands is significant (>80%) in relation to the primary energy consumed, based on gas and oil consumption [81,82] (INSEE, https://www.insee.fr/, accessed on 9 September 2024).

Comparing these small tropical islands is an interesting way to avoid bias caused by different contexts. Indeed, these islands are very similar in several characteristics because they (i) are tropical, (ii) have an economy based on tourism since the 1980s, (iii) have experienced a significant increase in their GDP (Gross Domestic Product) in recent decades, and (iv) have experienced significant demographic growth over the last decades; however, they (v) have very different GDP/inhabitant, (vi) do not have the same number of inhabitants (Table 1), and (vii) have no industrial production activity, except for Nouvelle-Calédonie Island [50,51,53,54,55,56,58,59,60,61,62,65,66,67,69,70,73,74,75,76]. Although they are administered by France and were once part of the French colonial empire, they do not currently have the same level of autonomy, at the present time. Political debates over increased autonomy or independence may be conflictual in these islands but are beyond the scope of this study.

Table 1. Socio-economic features of the studied islands and mean electricity production. E/yr/inhabitant = Electricity/year/inhabitant, NC = Nouvelle-Calédonie. GDP = Gross Domestic Product. Data on population and GDP come from INSEE (https://www.insee.fr/, accessed on 9 September 2024), but also from IEDOM [50,51,53,54,55,56,58,60,65,66,69,70] and IEOM [61,72,73,74,75,76]. Average values from 2010 to 2020 are presented. When there is only one value available for the period, the year of measurement is indicated. The uncertainty of the GDP/inhabitant is considered to be ±500 EUR/inhabitant. The demographic and socio-economic features specific to each island may influence electricity production and must be considered.

Island	GDP/Inhabitant (EUR)	Mean Population (2010–2020)	Electricity/yr (Mean 2010–2020, GWh)	E/yr/Inhab (GWh/yr/Inhab)	E/yr/GDP/Inhab
La Réunion	22,359 (in 2018)	846,062	2876.7	0.00340	0.13
Mayotte	10,600 (in 2021)	238,409	303.2	0.00127	0.029
Saint-Martin	16,572 (in 2014)	34,965	187.3	0.00535	0.011
Saint-Barthelemy	38,994 (in 2014)	9628	124.4	0.0129	0.003
French Polynesia	18,572	279,679	685.1	0.00245	0.037
Guadeloupe	23,449	396,286	1730.6	0.00436	0.074
NC (Metallurgy)	31,584	269,184	3062.9	0.01138	0.097
NC (No Metal.)	-	269,184	799	0.00297	0.025
Martinique	25,604	376,505	1562.8	0.00415	0.061

Table 1. Socio-economic features of the studied islands and mean electricity production. E/yr/inhabitant = Electricity/year/inhabitant, NC = Nouvelle-Calédonie. GDP = Gross Domestic Product. Data on population and GDP come from INSEE (https://www.insee.fr/, accessed on 9 September 2024), but also from IEDOM [50,51,53,54,55,56,58,60,65,66,69,70] and IEOM [61,72,73,74,75,76]. Average values from 2010 to 2020 are presented. When there is only one value available for the period, the year of measurement is indicated. The uncertainty of the GDP/inhabitant is considered to be ±500 EUR/inhabitant. The demographic and socio-economic features specific to each island may influence electricity production and must be considered.

Island	GDP/Inhabitant (EUR)	Mean Population (2010–2020)	Electricity/yr (Mean 2010–2020, GWh)	E/yr/Inhab (GWh/yr/Inhab)	E/yr/GDP/Inhab
La Réunion	22,359 (in 2018)	846,062	2876.7	0.00340	0.13
Mayotte	10,600 (in 2021)	238,409	303.2	0.00127	0.029
Saint-Martin	16,572 (in 2014)	34,965	187.3	0.00535	0.011
Saint-Barthelemy	38,994 (in 2014)	9628	124.4	0.0129	0.003
French Polynesia	18,572	279,679	685.1	0.00245	0.037
Guadeloupe	23,449	396,286	1730.6	0.00436	0.074
NC (Metallurgy)	31,584	269,184	3062.9	0.01138	0.097
NC (No Metal.)	-	269,184	799	0.00297	0.025
Martinique	25,604	376,505	1562.8	0.00415	0.061

The climate of the tropical islands investigated in this study is divided into two seasons: dry and wet. During the wet season, cyclones and storms may occur. The effect of climate warming on cyclones is complex. The number of cyclones is not expected to increase in the Caribbean. However, the intensity of major cyclones is expected to increase. The temperatures on these islands range from 20 °C to 35 °C. Weather data are provided by Météo France (https://meteofrance.re/fr, accessed on 9 September 2024). Climatic warming trends and rising sea levels indicate that these islands are becoming more vulnerable to extreme hydro-climatic events. The question of their adaptation and its modalities is controversial; tourism is considered a source of economic development, but is also a cause of natural resource consumption and waste increases.

3. Results and Interpretation

3.1. Influence of the Number of Inhabitants on Energy Consumption

3.1.1. Demographic Growth on La Réunion Island

Between 1975 and 2022, electricity production on La Réunion Island increased steadily (Figure 2A). Electricity production accounts for 15–18% of total energy consumption on the island of La Réunion. Electricity production growth accelerated from 1975 to 2000 but then slowed between 2000 and 2022 (Figure 2A). Electricity consumption has increased slightly in comparison to total energy consumption over the last few decades. Contemporaneously, there was an increase in the population, which appears to follow the same pattern (Figure 2B). From 1980 to 2000, the population growth rate accelerated. Following 2000, the variation in the number of inhabitants slowed down.

Renewable energies increased gradually from 1975 to 2022. More specifically, several growth pulses in renewable energy production can be observed in 1975, 2000, and 2011. However, the amount of electricity generated from carbon-based energy also increased (

Table 2. La Réunion Island energy consumption and production. Data source: Observatoire de l’énergie de La Réunion (https://oer.spl-horizonreunion.com/). Energy consumption and electricity production have increased during the last decades.

Year	Fossil Energy (ktep)	Renewable Energy (ktep)	Total Primary Energy (GWh)	Electricity Production (GWh)
2000	816.1	156.7	11,314	1758.1
2005	1043.8	149	13,872	2270.6
2010	1217	174.2	16,180	2698.8
2015	1214.8	196.5	16,415	2890.5
2017	1277.5	189.5	17,061	2986.0
2018	1256	186.7	16,779	2969.7
2020	1191	177.7	15,918	2878.5
2022	1227.4	203.2	16,638	3012.0

). Increasing consumption of carbon-based energy is observed despite the negative impact on the environment. Some annual variations in carbon-based energy have been decided to compensate for declining renewable energy production in 2021.

3.1.2. Demographic Impact on Small Tropical Islands

On small tropical islands, the larger the population, all other factors being equal (i.e., ceteris paribus), the higher the energy consumption (Figure 3). Including energy consumption associated with the production of imported products would significantly increase energy consumption on small tropical islands. Nonetheless, this bias is systematic, as no production occurs on the majority of the small tropical islands studied here. The case of Nouvelle-Calédonie Island shows that local ore extraction and production (extraction of nickel and metallurgical production) require a significant amount of energy. The inhabitants of Nouvelle-Calédonie Island consume as much energy as an “equivalent” island with over 800,000 inhabitants, nearly three times the actual population. The energy used for metallurgy purposes is included in the Nouvelle-Calédonie Island energy balance but not in the energy balances of the territories where metallurgic products will be used.

Primary energy consumption and electricity production on La Réunion Island are consistent with energy and electricity production and consumption on other small tropical islands. Each additional inhabitant contributes to energy consumption on the islands. Individual consumption and new activities associated with new residents appear to cause an increase in energy consumption. Some specific effects related to the urban or rural origin of the inhabitant may influence their energy consumption; however, this effect has not been investigated, and the focus is on the entire island’s energy consumption. There is apparently no scale reduction (i.e., no significant mutualization of equipment or infrastructures for energy reduction) in this case. The influence of industrial equipment, such as those on Nouvelle Calédonie, is significant and increases energy consumption. The role of infrastructure development in relation to wealth will be discussed later. Population growth may have an impact on economic growth [83]. The influence of the number of inhabitants and the increase in the number of inhabitants must be considered when interpreting variations in electricity consumption. To avoid misinterpretation of the variation in electricity consumption, the indicator of electricity consumption per capita will be used in the following sections.

3.2. Influence of Socio-Economic Features on Energy Consumption

3.2.1. Social and Economic Characteristics of La Réunion Island

This study will now focus on La Réunion, the island with the most inhabitants in our dataset, to analyze the influence of the other factors. When electricity production is normalized by the number of inhabitants, energy production per capita increases from 1975 to 2022 (Figure 4), as was the case when it was not normalized (Figure 2). The increase in energy consumption (Figure 2A) is not only due to population growth but is also caused by other factors.

New lifestyles and businesses may influence activities and consumption. Tourism has grown in many countries worldwide since the 1980s, including on tropical islands such as Saint-Barthélemy and Saint-Martin [49,52]. The development of this economic feature also modified the activities on La Réunion Island, favoring unemployment reduction. In 2022, there were around 500,000 tourist arrivals, whereas there were 400,000 tourists in 2005, 200,000 in 1990, and 3000 in 1963. However, employment in relation to tourism represents less than 5% of the employment on La Réunion in 2024. During the last decades, the GDP increased significantly in relation to economic growth on La Réunion as well as on the other small tropical islands in our dataset.

The general growth in electricity production was around 10% during the last period (2010–2022), which corresponds to a growth of around 1% per year. Nevertheless, this growth is smaller than previously. Let us investigate if the slowdown in energy consumption (even if the growth rate remains > 0) since 2010 is due to (i) an economic slowdown, (ii) a climatic effect, or (iii) increased efficiency in the use of energy.

The role of wealth and socio-economic crises on the long-term evolution of electricity consumption will be studied in Section 3.2.2 and Section 3.2.3, respectively. The impact of climate on electricity consumption will be analyzed in Section 3.3. The efficiency in the use of energy will be discussed in the Discussion section (Section 4).

3.2.2. Influence of Wealth on Energy Consumption on Small Tropical Islands

To better understand the evolution of electricity production per capita, it is necessary to consider the inhabitants’ wealth and evolution. The greater the GDP per capita, the higher the annual energy consumption per capita (Figure 5A). The causes of the wealth increase are beyond the aim of this study, but it must be taken into account that decadal GDP evolution could influence the decadal energy variation. In this study, wealth is described using the GDP per capita. This indicator has been criticized because it fails to characterize accurately the inequality within a territory (health, education, etc.) [84,85].

The greater the GDP per capita, the more the people have the possibility to have expensive social and cultural activities that increase electricity consumption. For example, the capacity to have a boat depends on income, and this will have an influence on primary energy consumption when the boat is used. The higher the GDP/inhabitant, the larger the houses and the more likely they have a swimming pool [49]. In this case, energy consumption could increase when people heat their swimming pool or use an air conditioner in larger houses. The increase in the GDP per capita could be associated with various activities and lifestyles. The use of new connected technologies (electric cars, computers, smartphones, air conditioners, etc.) increases electricity consumption. However, if the new inhabitants are very poor, the mean GDP/inhabitant will decrease, whereas the electricity consumption per capita could be stable. Higher incomes consume more energy, and their impact on sustainability is less virtuous. When income disparities are significant on an island, it is expected that energy consumption is not equally distributed on that island.

Electricity production is not always consumed directly by residents but could be consumed by other activities, such as metallurgy on Nouvelle-Calédonie or tourism on Saint-Martin, for example. A significant amount of energy was produced on Nouvelle-Calédonie Island for nickel extraction and metallurgical production during the period from 2010 to 2020. During this time, metallurgical production consumed more than 2200 GWh per year, accounting for more than 70% of the total of 3062 GWh (Table 1). Extraction and metallurgical production are exported rather than consumed for local use. The energy consumption of the mineral mining industry is significant [86]. The extraction of nickel and the transformation of the ore by the metallurgical industry enriched the population in a very inhomogeneous way. This industry employs approximately 5000 people, but an additional 15,000 people work indirectly in extraction and metallurgy out of a total of 270,000 inhabitants.

Wealth inequality is significant on Nouvelle-Calédonie. On Nouvelle-Calédonie, the 10% of the population with the highest incomes are 7.1 times richer than the 10% of the population with the lowest incomes, compared to 3.5 times in France [87]. When inequality is very high, the mean wealth obtained using the GDP per inhabitant is insufficient to characterize a territory accurately. The result presented in Figure 5A only indicates that the mean GDP/inhabitant causes an increase in energy consumption.

Excluding metallurgy on Nouvelle-Calédonie, the electricity consumption is relatively low compared to the GDP per capita. The GDP per capita without metallurgy is expected, in this case, to be less than EUR 25k per inhabitant, according to the graph in Figure 5B. This is comparable to the case of La Réunion Island. The extraction of raw materials (nickel) and production for export increases the GDP/inhabitant. However, the GDP/inhabitant on Nouvelle-Calédonie is not associated with electricity production comparable to the other small tropical islands with a similar GDP/inhabitant.

Wealth growth in these islands could be attributed to (i) new public investment, (ii) new economic activity development, such as tourism, (iii) new technological development, or (iv) new collective organization.

To avoid misinterpretation in relation to the GDP/inhabitant in these islands, it is possible to normalize electricity production (i) by the GDP/inhabitant and (ii) by the electricity production of a reference year.

3.2.3. Influence of Crises on Electricity Production

Small tropical islands may be vulnerable to natural hazards (cyclones, heavy precipitation, drought, marine submersion, landslides, erosion, and earthquakes), but they are also impacted by anthropogenic events (economic development, technological development, economic crises, demographic development, migration, health crises). Each island is impacted by various events that are not necessarily similar.

To compare electricity production between different islands, such as La Réunion, Mayotte, and Saint-Martin, it is necessary to normalize the indicator by the number of inhabitants. Indeed, as previously demonstrated, the number of inhabitants has a significant influence on electricity and energy consumption. The number of inhabitants is used to normalize electricity production, as La Réunion has more people than Mayotte and Saint-Martin.

However, the normalized electricity production per capita on Saint-Martin is higher than on La Réunion and Mayotte (Figure 6A), making comparison difficult. To make a more detailed comparison of the evolution of electricity consumption on these islands, it may be useful to normalize the indicator by the electricity production of a reference year. Electricity production was normalized relative to consumption in 2010. The economic growth of small tropical islands was impacted in 2008, 2009, and 2010 by the consequences of the subprime crisis that began in the United States. This impact can be observed on Saint-Martin and Saint-Barthélemy [22]. Tourist arrivals, unemployment, and, more generally, economic growth were all impacted (82-INSEE, 2014; https://www.insee.fr/fr/statistiques/1285278, accessed on 9 September 2024). The Chikungunya epidemic on La Réunion Island in 2006 had a significant economic impact but had no significant impact on electricity production because, during the epidemic crisis, the residents stayed at home to avoid mosquito risk and consumed electricity as they did during all the other wet seasons.

The effect of COVID-19 can also be observed on the annual electricity production of La Réunion Island (Figure 6B). On Mayotte, the annual electricity production decreased in 2020 and after. This observation is in agreement with observations in Europe [18] and elsewhere [20,21]. Even if it is not clearly observed in the annual electricity production of Saint-Martin, COVID-19 had a significant impact on the island’s economic and social activities in 2020 [22,88].

Hurricane Irma (2017) [89,90] had a significant impact on annual electricity production on Saint-Martin, not only in 2017 but also subsequently (Figure 6B). The impacts of this cyclone on the economy of the island were still observable in 2020 when the pandemic crisis occurred. Consequently, in 2020, the ongoing recovery of Saint-Martin Island following Irma’s destruction superimposed the impact of COVID-19 on social and economic activity. COVID-19 had a significant negative impact on tourism, the main economic resource of Saint-Martin Island [88], but its impact was more difficult to distinguish from that of Hurricane Irma on electricity consumption. In 2017, on Saint-Martin, approximately 20% of the population left the island after Hurricane Irma, and 10% did not return after two years [22,23]. Tourist arrivals at the airports and harbor were significantly reduced. The impact of COVID-19 on electricity production on Saint-Martin was partially hidden by the contemporaneous recovery of economic activity following Irma.

Social and economic activity evolutions could also explain the decrease in electricity production observed on Mayotte between 2014 and 2018 (Figure 6B). On Mayotte, there was a change in 2014 regarding (i) tax laws and (ii) the code of entry and stay for foreigners and the right to asylum. Concerning the first point, Mayotte was designated as an “outermost region” of the European Union in 2014. As a result, new rules were applied. French laws, including new financial taxation, were applied in ways that they had not previously. Concerning the second point, it was suggested by the French Home Minister in 2021 (gendarmerie.interieur.gouv.fr, consulted in 2024) that migrants had arrived on Mayotte and that the number of inhabitants had increased to 400,000. Nevertheless, experts from the French Statistics Institute considered this to be an overestimation (blog.insee.fr/mayotte-census-adapted-to-non-standard-population, accessed on 9 September 2024).

3.3. Climatic Impact on Electricity Production

3.3.1. Seasonal Influence and Extreme Hydro-Meteorological Events

From 2017 to 2022, except in 2020, during the COVID-19 pandemic, La Réunion Island was not impacted by exceptional socio-economic events. Consequently, the data could be used to identify the role of climatic events on electricity production. On La Réunion Island, monthly data show that electricity production varies seasonally (Figure 7). The variations are caused by cyclical activities, specifically socio-economic activities. In particular, electricity production is greater during the southern hemisphere summer (November, December, January, February, and March) than during the southern winter (May, June, July, August, and September). The mean monthly electricity production increase from winter to summer is 15–20%, from 230–240 MWh to 270–280 MWh. Every year, there are two peaks, centered in January and March.

As previously mentioned, extreme events can cause significant damages [34,91,92,93,94,95,96,97] and significant changes in energy consumption, as seen in the case of Hurricane Irma on Saint-Martin (Figure 6B) and on Saint-Barthélemy. One of the climate-related effects is a decrease in electricity production. The primary causes of decreased energy consumption during and after extreme events are (i) destruction or damage to power plants, (ii) destruction or damage to electricity networks, (iii) destruction or damage to infrastructures, (iv) migration and population decrease, and (v) economic and social activity slowdown [11,22,23,26,34].

Initially, there is the destruction of electricity production plants, as well as a reduction in energy distribution when there is complete or partial destruction of the electricity network [11,26,34]. Blackouts can have severe health consequences [34]. The destruction of other infrastructures, such as buildings, water distribution, roads, and transportation vehicles, has a significant economic impact. Territorial recovery can take years to return to a situation similar to the original one when the damages are severe [23]. Hurricane Irma caused a population departure on Saint-Martin (7000–8000 people left), and the island’s population is still approximately 3000–4000 lower than it was prior to the hurricane. The population decrease may explain part of the electricity production decrease on Saint-Martin after 2017 [22]. Whereas it took 1.5 years for Saint-Barthelemy to return to electricity production levels similar to those observed prior to the hurricane, electricity production on Saint-Martin was lower in 2022 than in 2016.

In 2018, La Réunion Island was impacted by three cyclones in 11 days (Figure 7). In 2022, La Réunion Island was impacted by cyclone Batsirai for 4 days (

Table 3. Main climatic events during 2017–2022. Data source: Météo France (https://meteofrance.re/fr, accessed on 9 September 2024). Temperature, precipitation, and natural hazards are not constant during the last years and may influence electricity production.

Year	Extreme Events	Precipitation Maximum (mm)	Mean Annual Precipitation Variation (Normalized by the Mean Value Obtained Between 1981 and 2010)	Mean Annual Temperature Variation (Normalized by the Mean Value Obtained Between 1981 and 2010)
2022	Cyclone Batsirai	584 mm in 4 d, February	+5%	+0.1 °C
2021			−5%	+0.65 °C
2020			−25%	+0.2 °C
2019			−21%	+1.2 °C
2018	Cyclone Ava Cycl. Dumazile	558 mm in 6 d, January 541 mm in 4 d, March 176 mm in 1 h, April	+40%	+0.65 °C
2017			−8%	+0.9 °C

). Heavy rainfall during these cyclones was recorded (Table 3, Météo France data source; https://meteofrance.re/fr, accessed on 9 September 2024). Nevertheless, during the same events, winds did not significantly damage infrastructures or buildings. Consequently, electricity consumption decreased, but not as significantly as it had during Hurricane Irma on Saint-Martin and Saint-Barthelemy in 2017. When there are no significant cyclone-related destructions, the heavy rain alone does not result in a significant decrease in electricity production. On the contrary, when cyclones are associated with high temperatures and do not cause damage, significant electricity production may occur as a result of the temperature. Cyclones frequently occur when the sea temperature is high, and this coincides with a period of high air temperature on La Réunion.

3.3.2. Influence of Temperature on La Réunion Energy Production

Seasonal increases in electricity consumption occur during summer on La Réunion Island, when the precipitation and the temperatures are higher (Figure 8). In 2018, 2020, and 2022, a small decrease in electricity production can be observed in comparison to 2017, 2019, and 2021 (Figure 7). As previously described, COVID-19 had a significant impact on energy consumption in 2020. The occurrence of cyclones in 2018 and 2022 decreased energy production slightly, but only for a few days.

The hottest year was 2019 (Table 3), and it also had the highest electricity consumption (Figure 9A). The seasonal increase in temperature causes an increase in energy consumption on La Réunion Island (Figure 9,

Table 4. Influence of temperature variation of electricity production on La Réunion. Data in relation to Figure 10. Data sources: Météo France (https://meteofrance.re/fr, accessed on 9 September 2024), Observatoire de l’Energie de la Réunion (https://oer.spl-horizonreunion.com/). The impact of monthly temperature variation on monthly electricity production is quantified.

Month	% of Electricity Production E Increase	Temperature T Increase (in °C)	Uncertainty on E Variation (%)	Uncertainty on T
1—January	15	5	5	0.5
2—February	10	5	2.5	0.5
3—March	17.5	4.5	5	0.5
4—April	12.5	3.5	3	0.5
5—May	4	1.5	1	0.5
6—June	0.5	0.25	3	0.5
7—July	4	0.5	1	0.5
8—August	5	1	0.5	0.5
9—September	4	1.5	0.5	0.5
10—October	9	2.5	4	0.5
11—November	10	3.5	4	0.5
12—December	17.5	4.5	5	0.5

). A 1 °C increase in temperature above 26 °C (i.e., from 31 °C to 32 °C) between December and March could increase electricity energy consumption by 3 to 12% (Figure 9B). The growth rate is around 3.2%. On La Réunion, temperature was the primary climatic cause of electricity consumption fluctuations, rather than heavy rains or winds, from 1975 to 2022. The heavy rains that occurred in 2018 did not cause significant electricity production variation. In 2017 and 2019, average temperatures were higher than in 2018 and correlated with higher electricity production during summer. In 2021, the average temperature was higher than in 2022, and the electricity production was, also.

The correlation observed between monthly temperature peaks and monthly electricity production increase suggests a potential causal effect of temperature increase on electricity production. Climate warming has caused an increase in average temperatures on La Réunion Island over the last fifty years (Météo France; https://meteofrance.re/fr, accessed on 9 September 2024). Contemporaneously, electricity consumption per capita has increased on La Réunion Island over the last few decades. As previously stated, a part of the increase in energy consumption per capita was caused by the wealth increase. Nevertheless, the increase in yearly energy production per capita on La Réunion Island over the last few decades can also be attributed in part to rising temperatures, analyzing the evolution of temperature and electricity production (Figure 10).

4. Discussion

4.1. Climate and Electricity Production

Hot temperatures increasing over recent decades have been documented by many studies [6,98]. The present study shows that increasing temperatures correlate with an increase in electricity production on both a monthly and yearly scale (

Table 5. Relationship between natural or anthropic events and energy consumption.

Event	Energy Consumption	Example
Cyclone	Decrease	Irma, 2017, Saint-Martin Ava and Dumazile, 2018, La Réunion
Temperature increase > 26 °C	Increase	La Réunion
Migration out	Decrease	Saint-Martin, 2017
Population increase	Increase	Small tropical islands
Wealth increase	Increase	Small tropical islands
Epidemic crisis	Decrease	COVID-19, Chikungunya

, Figure 9 and Figure 10). The causal relationship between temperature increases and the growth of electricity production per capita on small tropical islands has been observed since the 1990s. This causal relationship was also suggested in other areas such as South and East Asia, the Middle East, the Pacific, North Africa, and the Sahara [99]. This effect was not clearly observed before the 1990s on small tropical islands because (i) it occurred during a period of strong economic and demographic growth that hid “small” fluctuations, (ii) hot events are more intense now, and (iii) the use of air conditioning increased significantly over the last two decades.

Using a more accurate time step, the peaks of electricity consumption on La Réunion Island depend on the hour of the day and are from 10 a.m. to 2 p.m. during Austral summer, when the temperature increases, whereas there are from 5 p.m. to 7 p.m. during Austral winter, when temperatures are hot and at the end of the working day, when people return home and use household appliances. The majority of electricity is consumed when household appliances are used at the same time during hot weather events. This could indicate that air conditioners are one of the primary causes of increased electricity production.

The alternative explanation, which involves the role of tourism in explaining the increase in electricity production, is not relevant. The electricity production increase occurs during the wet season, which corresponds to the warmer season. However, during the wet season, the number of tourists is lower. The increase in tourist arrivals during the dry season seems to have no significant influence on electricity consumption on La Réunion Island, even though tourism represents around 10% of the island’s businesses and approximately 450,000 arrivals per year, mainly between June and November [60]. The impact of tourist arrivals on electricity production is more significant when tourism represents a higher percentage of the activities, as on Saint-Martin Island.

4.2. Implications of Climate Warming on Tropical Areas’ Electricity Production

The energy consumption and electricity production of small tropical islands are small compared to the rest of the world. Nevertheless, they could be used to estimate the impact of increasing temperatures in tropical areas on energy consumption and electricity production. It is estimated that Earth’s tropical area population is 3.5 billion inhabitants and represents 48% of the population [100]. A broad estimation of electricity production in tropical territories is to consider that 4000 more inhabitants produce 4000 more energy. Based on the data from 1991 to 2022 on La Réunion, this study shows that a +1 °C increase results in an increase of 0.0012 GWh/inhabitant (Figure 10). Basically, 850,000 inhabitants on La Réunion produce approximately 1.0 TWh per year for a temperature rise of +1 °C, whereas 3.5 billion inhabitants in tropical areas should produce 4000 TWh per year under the same conditions.

Another method to estimate the impact of climate warming, independent of the previous one, is the results of monthly temperature rise during summer in the southern hemisphere (Figure 9). According to the monthly dataset of La Réunion Island from 2017 to 2022, a +1 °C increase causes a 3.2% increase in monthly electricity production. The average monthly electricity production on La Réunion Island is 0.265 TWh during the six warmer months (Figure 8). A 3.2% increase of 0.265 TWh is equivalent to an increase of 0.0085 TWh per month when the temperature rises by +1 °C during the six warmer months. This is equivalent to 0.5–1.0 TWh per year for La Réunion Island (i.e., 0.5 TWh < 2 × 0.265 TWh < 1 TWh) and 2000–4000 TWh per year for a population 4000 times higher.

This increase in electricity production causes a greenhouse gas increase because electricity is mainly produced by non-renewable energies. Assuming that 90% of the electricity is of non-renewable energy origin, around 0.9 TWh per year on La Réunion and 3780 TWh per year in tropical areas will be produced in relation to a temperature increase of +1 °C.

In the future, a temperature rise of +1 °C may cause an electricity production increase of 1900–3800 TWh per year in tropical areas, representing an increase of 3.2%, in agreement with previous studies in Singapore and Hong Kong that estimate increases ranging from 3% to 4% [101]. Currently, the majority of electricity is produced using fossil resources. A 1 °C or 2 °C increase in mean temperature in tropical areas could significantly increase greenhouse gas emissions. In these regions, the feedback between global warming and electricity production is unsatisfactory and is expected to increase. The impact of electricity production increase on climate has been suggested in previous studies [102,103]. Climate change will have a significant impact on energy production and consumption if current behavior remains constant in tropical areas.

4.3. Literature Review on Temperature–Energy Variations

The effect of temperature increase on energy consumption increase has been estimated in urban [104,105] and rural areas [106] for daily [107], monthly [108], as well as annual [109] temperature variations. This effect has been measured in industrial and commercial enterprises, in the hospitality sector, and in people at the individual and the collective level [109,110]. The measurement of electricity consumption for buildings for a given temperature increase is necessary to improve building isolation and architecture [111]. Electricity costs may be impacted when there is a significant rise in electricity demand [112].

Temperature increases above a certain threshold have been shown to increase electricity consumption in Europe [113], China [106], South America [114], the USA [115], Arabia [116], Singapore, and Hong Kong [101]. Energy consumption is independent of the tropical nature of climate but depends on the absolute temperature, which must be above a certain threshold, on the number of inhabitants, and on wealth. The average summer temperature in each nation may influence the use of air conditioners, but wealth also plays a role [117].

One of the present effects of climate change is highlighted by the measurement of the impact of rising temperatures on electricity production. These data are necessary to estimate the future evolution of energy consumption. In China, it has been observed that the temperature increases in summer cause more impact on electricity consumption than wintertime temperature decreases. A one-degree temperature increase may lead to 0.015% more electricity consumption per capita [106]. In Hong Kong, a +1 °C ambient temperature rise causes an electricity consumption increase of 9.2%, 3.0%, and 2.4% in the domestic, commercial, and industrial sectors, respectively [104]. Ang et al. (2017) [101] estimate that in Hong Kong and Singapore, the electricity consumption increase ranges from 3% to 4% under the same conditions. This demonstrates how significantly lowering the temperature through air cooling affects energy and greenhouse gas emissions. Potential technological solutions to reduce carbon dioxide emissions include new cooling techniques, improvements in energy efficiency, and an increase in the production of low-carbon electricity [117].

4.4. Limitation to the Anticipating Scenario

4.4.1. Socio-Economic Trends vs. Predictability

The potential consequence of expected climate warming on electricity production is partially speculative because it will be true only if the other parameter remains constant (i.e., ceteris paribus condition). First, the progressive increase in wealth in numerous countries, especially in tropical areas, may favor an increase in energy consumption. If there is no change in the trends demonstrated above, it is expected that energy consumption will increase in tropical areas.

Second, the fact that wealth correlates with an increase in electricity production on small tropical islands does not imply that these rules will always apply. Wealth causes human activities to use technologies that consume energy. The impact of wealth could also be relatively small on electricity production in the future. It has been shown that GDP per capita has increased significantly in recent years; despite this, the increases in electricity consumption were small. On the contrary, the development of mobility and electric vehicles is also expected to increase electricity consumption, even if in unequal amounts [99,118]. However, more sustainable technologies consume less energy than those that were not developed to be energetically efficient. Potential technological breakthroughs may reduce the present energy consumption.

Third, the present causal relationship between temperature increase and electrical production depends on the temperature absolute value: the effect is not the same if the temperature is <15 °C or >26 °C. In temperate areas, energy consumption increases significantly during the winter when the temperature is <15 °C. Increasing temperatures may have different consequences for energy consumption depending on the context, which must be investigated specifically.

Fourth, the amount of renewable energy in the mix may also increase in the future, limiting the growth of the negative effect of the carbon footprint [119], even if this point is controversial. Economic policy uncertainty and financial crises have a significant negative impact on renewable energy consumption [120]. For the latter, the uncertainty does not help to predict exactly the economic evolution.

Fifth, the estimate of the increase in electricity production in tropical areas due to temperature increases is broad and does not account for the potential improvement in air conditioner performance, as well as the increasing number of air conditioners that will be used over the next decades.

Socio-political decisions may have strong consequences on energy consumption. The increase in renewable energy production in the 1980s was a reaction to the oil crisis of the 1970s [28]. The fluctuations in energy production and consumption reflect interactions between the environment and society, but also geopolitical evolutions. Recent conflicts produced effects on oil and gas prices and energy consumption. Concerning more specific social impacts on energy consumption, it has been observed that economic crises (subprime), and also reduced economic activities (during COVID-19) caused energy consumption to decrease. In the future, behavior may change to reduce energy consumption.

4.4.2. Data Collection Efforts

To better adapt to or mitigate climate change in tropical territories, it is necessary to estimate the impact of climate events (cyclones, high temperatures, high precipitation, strong winds) on electricity production and consumption. More precisely, accurate climate change prediction depends on the feedback mechanism between rising temperatures and the production of electricity. Based on these projections, mitigation and adaptation policies could have a significant impact on the development of tropical territories.

These estimates are based on the availability and quality of climatic indicators (temperature, precipitation) and detailed energy production/consumption data. Data quality is determined not only by measurement accuracy but also by the number of data points used. Temperature varies significantly from one geographical area to another. This study uses data from three different sites on La Réunion Island. Increasing the number of data points can reduce uncertainty and improve data accuracy.

The reliability of similar effects on electricity production could be improved by increasing data collection efforts on other small tropical islands and tropical areas. Centralized data collection in each territory, as well as freely available and standardized data, could be a benefit not only for a better understanding of climate impact but also for energy production management and democratic debates about energy policy.

To improve the analysis of climatic events, data collection efforts must be conducted over long periods (decades) and with small steps (hours, days). If more intense climatic events could cause significant monthly electricity production, smaller events could be observed on a daily or hourly scale. Using appropriate data collection, it will be possible to integrate the impact of climatic events on electricity production at the regional or global levels.

4.5. Adaptation to Climate Warming

4.5.1. Development, Wealth, and Energy Consumption

Reducing energy consumption is a necessary solution to reducing the negative feedback of climate warming. The possibility of reducing negative retroaction between climate warming and energy consumption must be explored. First, a reduction in carbonic energy production is a necessary change. Economic activity should develop without increasing air pollution and inequality. The lifestyles in wealthy territories often generate an increased consumption of energy and environmental impacts, as suggested in previous studies [12,121,122,123]. A sustainable lifestyle should be promoted.

On small tropical islands, the energy consumption increase is positively correlated with GDP increase. The Human Development Index (HDI) also has a positive correlation with an index that measures electricity consumption, such as the Night Light Development Index (NLDI) [26], suggesting that an HDI increase may cause an energy consumption increase, even if in a reduced way. However, it has been suggested that the Human Development Index (HDI) or the GDP/inhabitant could grow significantly even if energy production and consumption do not increase when energy consumption reduction strategies are implemented [28].

The Human Development Index is more adapted to characterize social and care activities than GDP, which is better suited to describing economic activities. Developing HDI rather than GDP/inhabitant may favor low-carbon activities if social and cultural activities are more local and environmental norms are considered. Inequality exists and is expanding in the carbon footprint [123]. The trends that correlate HDI with energy consumption could be partly modified by changing behavior and improving technologies.

4.5.2. Techno-Solution vs. Socio-Solution

Reducing energy production and consumption is sometimes considered a sustainable objective, not only to reduce carbon footprint but also to reduce human impact on the environment. Many of the suggested reduction strategies are based on innovation. Innovation could be based on techno-solutions or socio-solutions. Concerning techno-solutions, the main technologies that are considered relevant to solve the problem are (i) new low-carbon energy production such as those based on solar, eolian, nuclear, or hydro-based sources, (ii) new technologies favoring low-energy consumption when used in comparison to previous technologies, and (iii) low-tech materials favoring low energy consumption throughout the device’s life [119].

Strategies and policies influence energy consumption. On the small tropical islands investigated, there is no relationship, at the present time, between the territory’s GDP per capita and the development of renewable energies. The main renewable energies on small tropical islands are (1) hydroelectricity, (2) geothermal energy, and (3) solar energy. The percentage of renewable energy on small tropical islands depends on when renewable energy development began locally. In these territories, hydroelectricity was the first renewable energy, developed primarily in the 1980s, when conditions were favorable (water and relief). Solar energy has only recently developed on these islands and will not account for a significant portion of total energy production by 2024. Despite the volcanic origin of these islands, geothermal energy is not significantly developed, except in Guadeloupe.

The possibility of reducing energy consumption for a constant number of inhabitants should be considered as a possible objective. Energy consumption could be reduced by, for example, (i) using appropriate materials to reduce thermal effects in buildings [124,125], (ii) considering alternative modes of transportation behavior (train, bicycle, walking, etc.), (iii) considering more efficient networks and electrical devices, for example, more efficient air conditioning systems, and (iv) considering alternative economic and social activities that have a reduced impact on greenhouse gas emissions.

The activities impacted by new technologies may include both pre-existent activities and new activities generated by new technologies. New technologies may increase energy consumption if they generate more energy-intensive activities than before. Consequently, it may be necessary to implement specific socio-economic measures to limit or reduce energy production and consumption. In the specific case of socio-solutions for energy production and consumption reduction, current strategies are based on (i) individual sustainable strategies and (ii) collective sustainable strategies.

There are various lifestyles able to modify the carbon footprint and may decrease it [13,126]. Low-carbon lifestyles depend on low-carbon cognitions and propensity toward low-carbon behaviors [13]. Low-carbon cognitions could depend on education on the relevant behavior concerning waste management, sustainable mobility, and activities. Individual actions that could reduce the carbon footprint include (i) reducing mobility, (ii) eating locally, seasonally, and vegetarian, (iii) reducing home-building footprints, and (iv) reducing the footprints of home appliances [127].

Individual behaviors are necessary, but collective actions must also be considered when reducing the carbon footprint. The following collective actions could reduce the carbon footprint: (i) promoting collective mobility through appropriate infrastructures, (ii) promoting local–seasonal–vegetarian foods through appropriate infrastructures and cultural education, (iii) promoting well-designed buildings and urbanism (thermally efficient, reduced urban expansion), (iv) encourage low-carbon economic activities, (v) promote waste recycling and extending the life of consumer goods, (vi) promote knowledge improvement and education increase about consistent sustainable behavior, (vii) promote energy production and consumption reduction, and (viii) contemporaneously with the strategy of global energy reduction and reduction of the environmental impact of human activities, favor renewable energy rather than carbon energy [13,103,119,120,126,127].

The costs of strategies based on technological and socio-economic solutions are difficult to assess. Energy conservation campaigns and policy changes need time and dedicated staff to be efficient. First, if these solutions do not already exist and must be invented, their cost is highly speculative. Second, if these solutions exist but are not widely disseminated, the cost of their diffusion is not only financial. Social reactions to changes that affect the way of life could be negative. Community members may have their own road map that is compatible with socio-solutions or not. Reactions may vary depending on the trajectory of the territories. Behavioral transformation may be difficult to implement. On the contrary, adapted behaviors may be rapidly developing. Policy interventions may encourage more sustainable behaviors. Educational costs, as well as the democracy costs related to the organization of democracy (debates, legal processes, public processes of discussion), are beyond the aim of this study. Research costs concerning relevant socio-solutions and concepts could be underestimated when neglecting the time and the number of people necessary for the investigation. Techno-solutions are not independent of socio-solutions. The development and diffusion of efficient techno-solutions depend on the social context. Promoting research and education concerning socio-solutions for sustainability is a first step.

5. Conclusions and Perspectives

Energy is a relevant indicator for monitoring societal changes (wealth increase), as well as to record the impacts of climate on society. More precisely, the monitoring of climate impact on society could be measured from energy consumption. The impact of climate on electricity consumption is observed on small tropical islands when temperature increases or extreme hydro-climatic events, such as cyclones, occur. The effects of climate on energy consumption on small tropical islands are sometimes indirect but not negligible. Climate change causes variations in energy consumption through the following effects:

• (i) The destruction of electricity-producing infrastructures;
• (ii) The destruction of electricity-consuming infrastructures;
• (iii) The reduction of electricity-consuming activities;
• (iv) The destruction or damage to networks, such as roads or telecommunications;
• (v) The increased use of air conditioning;
• (vi) Lifestyle changes with air temperature and weather.

Socio-economic events (financial crises), socio-health events (COVID-19), socio-politic strategy modifications (geopolitical crises, new laws), and socio-demographic evolution (increased population, migration) influence electricity production on small tropical islands. Financial crises, socio-political crises, health crises, cyclonic events, and migration off islands cause electricity production to decrease. A temperature increase of +1 °C causes an electricity production growth of 3.2% and an increase of 1.5 MWh per year and per capita. The possibility to influence future energy consumption through socio-technical changes (desire to reduce waste and goods consumption, energy savings, better building insolation) could be efficient and avoid negative feedback between energy consumption and climate change.

Due to the negative feedback between temperature increase and energy production, reducing energy consumption/production is necessary to mitigate the negative impact of human activities on the environment. The research perspectives and policy improvements are as follows.

First, more data must be collected on energy and electricity production/consumption. These data should be published and open access.

Second, climate models should include feedback between temperature increase and energy consumption.

Third, more research into renewable energies and smart-grid technologies is required. The development of renewable energy must be contemporaneous with a diminution of hydrocarbon energy production.

Fourth, improving the efficiency of low-carbon and low-environmental-impact energy systems should be mandatory.

Fifth, research on socio-solutions must be conducted to assess the potential benefits and costs of these strategies.

Sixth, research and policy should also focus on territories’ resilience to climate extreme events.