key: cord-0889991-hroglq5e authors: Ashkanani, Ahmad M.; Bahman, Ammar M.; Aljuwayhel, Nawaf F. title: Impact of COVID-19 interventions on electricity power production: An empirical investigation in Kuwait date: 2021-12-10 journal: Electric Power Systems Research DOI: 10.1016/j.epsr.2021.107718 sha: f8cf5de89fcf6a072954f86d0a17fb99990ad0e6 doc_id: 889991 cord_uid: hroglq5e This paper investigates the impact of various COVID-19 containment measures on electricity power production in the State of Kuwait. Using longitudinal archival data from 2015 to 2020 and advanced regression analysis, the paper compares post-pandemic electricity production levels to pre-pandemic levels under various post-pandemic lockdown scenarios (ranging from no lockdown to partial and full lockdowns). The results showed a decline in average hourly electricity load during the various post-pandemic periods (compared with pre-pandemic levels) where the magnitude of the decline varied across the hour-of-day and the type of lockdown policy. On average, there was a decline of 9.0% to 14.0% in average hourly load during the post-pandemic period with no curfew policies (compared with pre-pandemic levels). This decline was magnified when lockdown was enforced, leading to a net reduction of up to 16.4%, but only when curfew hours coincided with peak load hours (i.e., from noon to sunset). These findings could help policymakers understand the impact of various COVID-19 measures on the demand for electricity, thus equipping them for optimal decision-making. • Electricity production declined 9% to 14% during post-pandemic (no curfew) period. • Electricity production decreased by up to 16 .4% during full lockdown period. • Curfew effect on electrical load was most evident during peak load hours. The outbreak of Coronavirus Disease 2019 (COVID- 19) has impacted almost all aspects of life globally, adding pressure on the medical sector, leading to closures in the educational sector, and affecting the commercial sector's viability. The industrial sector is of no exception with experts raising concerns about the COVID-19 pandemic impact on electricity supply chain [1] [2] [3] [4] . With the enforcement of lockdown restrictions across several countries around the globe, electricity demand has been fluctuating, with early evidence suggesting a reduction in the production of electricity from the energy sector [5] [6] [7] . Therefore, governments around the world, including Middle East and North Africa (MENA) and Gulf Cooperation Council (GCC) countries, have implemented emergency action plans with the help of field experts and researchers to address concerns about how and when COVID-19 might impact electricity power production [8] [9] [10] [11] [12] [13] [14] [15] . Energy consumption changed amid the COVID-19 outbreak [16] [17] [18] . Jiang et al. [19] and Zhong et al. [20] overviewed COVID-19's impact on energy demand and consumption, where Abdeen et al. [21] , Rouleau and Gosselin [22] , Ruan et al. [23] and Zanocco et al. [24] reported high correlation between electricity consumption and the number of infected people, social distancing range, and residential and commercial activities level. Chen et al. [25] surveyed the utility power consumption bills during the month of February and reported approximately 50% of the residents experienced higher electricity use than average, whereas 40% estimated their electricity consumption to be about the same. Ruan et al. [26] found higher fluctuation in electric demand and prices compared to electricity power production. Cheshmehzangi [27] extended the survey to May 2020, where he observed an average increase of of 67% (February 2020), 95% (March 2020), 35% (April 2020), and 22% (May 2020) in electricity energy costs compared with similar time periods in 2019. These results matched the IEA [28] reports about the increase in residential energy consumption due to the pandemic lockdown measures and changes in social activities. Similarly, Akrofi and Antwi [15] and Snow et al. [29] exhibited an increase in electricity consumption in the residential sector in Africa and Australia, respectively, while the COVID-19 outbreak situation was further examined by Edomah and Ndulue [30] under three scenarios (i.e., no lockdown, partial lockdown, and full lockdown) in Lagos, Nigeria. Compared with the industrial and commercial sectors, the residential sector's electricity consumption increased by up to 7.8% relative to the no lockdown case. In addition, during the post-pandemic and no lockdown period, a relative increase in the total power consumption (about 3.86%) was observed, which was mainly driven by the increase in electricity demand in the commercial sector (increased by 9.3%). However, the full lockdown measures decreased the total consumption by up to 2.3%, while partial lockdown had no significant effect on the electricity demand. Despite the sector-level analysis, the study overlooked the impact of varying the number of curfew hours on electricity production. Bahmanyar [43] further extended the accuracy and stability of the electricity demand predictions using multi-objective optimization. However, the black-box model (i.e., artificial neural network) used in both studies [38, 39] include descriptions of the context, data and methodology, and results and discussion of the study, respectively. Section 5 summarizes the main conclusions and the policy implications withdrawn from this paper. The State of Kuwait is located in Western Asia, covering a total land area of nearly 18,000 km 2 . The climate in Kuwait is characterized by extremely hot and dry weather during most of the year. The average maximum monthly temperatures range between 37°C to 45°C (98°F to 113°F). Kuwait depends heavily on fossil fuels, including natural gas, heavy fuel oil, and crude oil, to generate electrical power through steam, gas, and combined power cycles. Around 60% of the power generated is by steam turbine cycles and the remaining 40% is generated by gas turbine cycles [44] . The gas turbines are used in emergencies and during times of peak load. There is only one renewable power plant (Shygaya station) that produces almost 0.02% of the total power in Kuwait. However, it is expected to generate 15% of total power production by 2030 (10 MW from solar modules, 10 MW from wind power, and 50 MW from solar collector power). Currently, nine power stations, including the renewable energy plant, generate a total electrical power of 7,980,224 MWh. About 10% of the total electrical power generated is utilized internally for water desalination and auxiliary power demand [45] . Table 1 lists the total capacity of each power station, while Table 2 shows the tariffs of electricity for each sector set by the government of Kuwait. of GDP) [44] . There are more than 60 production units of electric power distributed across the power stations, 26,000 electric transformers located across the country with 33,000 km of electrical medium voltage (MV) cables, and 36,000 km of electrical high voltage (HV) overhead power transmission lines. Figure 1 illustrates the power supply chain in Kuwait from the natural resources to the customer's household. To be noted that electricity consumption in Kuwait matches electricity production levels [46, 47] . This is mainly due to the use of a "Pull" system where electricity production in Kuwait is matched with forecasted demand in a given time period. This is achieved by diverting the hot steam in power plants to produce fresh water in the adjacent desalination plants (i.e., multi-stage flash (MSF) distillation process) [48] . In addition, electricity storage technology is unemployed currently in Kuwait. Thus, most of the produced electricity is generated to match the exact demand. Therefore, the data, methods and results of the electricity supply chain is only presented in term of electricity production. The COVID-19 pandemic impacted both the political and economical context in Kuwait. In terms of the political context, new mandates were issued to enforce partial and full lockdown policies. Unlike some countries that implemented soft lockdown measures (i.e., shelter-in-place orders meant that people were "advised" to stay at home and reduce their activities outside of home), Kuwait implemented hard lockdown measures where movement was restricted to various degrees depending on the enforced lockdown policy (e.g., partial vs. full lockdown). These lockdown policies were accompanied by laws that punished people who violated the lockdown orders with financial fines and/or imprisonment [49] . Law enforcement agents patrolled residential areas to enforce the lockdown policies and apprehend any violators. The economical situation was also influenced by the pandemic where the majority of businesses were forced to shutdown during various phases of the pandemic. This included private businesses (e.g., restaurants, cafes, gyms, shopping malls, etc.), governmental entities (e.g., ministries, public transportation, parks and beaches, etc.), and educational entities (e.g., private and public schools, colleges, and universities). Thus, these closures and lockdown policies were associated with an increase (a decline) of up to 46% (90%) in mobility within the residential (retail and recreation) sector, respectively, over the time period from February 24, 2020 to August 7, 2020 as shown in Figure 2 [50]. The average hourly electrical load data (in MW) were collected from the National Control in Kuwait (i.e., date of measurement ≥ February 24, 2020), zero otherwise. This factor was considered because it is expected to be associated with changes in electricity production levels, as suggested by Roidt et al. [54] . Table 3 summarizes the different time periods observed in this study. A curfew hour indicator (CURFEW ij ) was also created, which is a dummy variable that is equal to one if curfew was enforced during the hour that the electricity data was measured in (i.e., hour j of day i), zero otherwise. This variable allows us to examine the effect of enforcing curfew hours on the average hourly electrical load. Similar to POST i indicator, the pandemic-related news were tracked from the official daily announcements published by MOH and CGC to construct the CURFEW ij variable. The aforementioned factor was selected because it is expected to be associated with changes in power generation as suggested by López Prol et al. [39] and Hale et al. [42] . The average daily temperature (TAVG i ) in Kuwait was controlled for using climate data from Peterson and Vose [55] and EPA [56], which include daily temperature measurements (in°C) from two weather stations located in Kuwait international airport and Hawalli gov-ernance, respectively. These weather stations were approximately 10 km (6 miles) apart and shared similar elevation levels. The TAVG i was defined as the average of the two daily average temperature measurements obtained from the two stations. This factor was considered to control for potential climate effects where more electricity is expected to be consumed as the average temperature rises (i.e., electrical load is expected to increase during hotter days). The effect of the month of measurement (MONTH i ) was also controlled for using 2nd degree polynomial terms to account for the inverted U-shaped monthly curvilinear effect found in the raw data (see Figure 3 ). Finally, a set of dummy variables was constructed to control for non-observed temporal fixed effects (FE), which includes year FE (e.g., population growth) and day-of-week FE (e.g., weekday versus weekend effect). Additional controls for a 3rd degree polynomial effect of hour-of-day (HOUR ij ), which ranges from 0 to 23, was also considered to account for within-day curvilinear temporal effects (see Figure 4 ). Finally, the interactions between the curvilinear effects of hour-of-day (HOUR ij ) and the constructed post-pandemic (POST i ) and curfew hour (CURFEW ij ) indicators were controlled for to examine potential changes in the hourly effects during post-pandemic periods with and without curfew enforcement. Observations with missing daily temperature measurements were dropped, resulting in a total of 48,312 hourly observations. Table 4 includes summary statistics of the main variables. A multilevel regression model with random intercepts [57, 58] was used to examine the statistical relationship between the average hourly electrical load (LOAD ij ), the main pandemic-related predictors (POST i and CURFEW ij ), and the control variables specified above. A multilevel approach allows for the examination of changes in the average hourly electrical load both within a day (level-1) and between days (level-2). The econometric specification for the level-1 component of the multilevel regression model is given by: where i and j denote the measurement day and hour of LOAD, respectively, and r ij is the within-day (level-1) random error where r ij ∼ N (0, σ 2 r ). The econometric specification for the level-2 component of the multilevel regression model is given by: π 0i = γ 00 + γ 01 POST i + X i Γ 02 + u 0i (2) π 1i = γ 10 where X i is a vector that includes the remaining control variables (Γ 02 is a vector of regression coefficients that are multiplied by the control variables in X i ), and u 0i is the between-day (level-2) random error (allows the intercept of LOAD to vary randomly across days) where u 0i ∼ N (0, σ 2 u ). Combining Eq. 1 and Eq. 2 yielded the following mixed model: An alternative variant of the above model was also considered by taking the natural logarithm of LOAD ij . This approach allows for the examination of the relationship between the explanatory variables and the percentage of change in the outcome variable and addresses potential skewness in the distribution of the outcome variable. The econometric specification of the mixed logged model is given by: These models were estimated using the PROC MIXED procedure in SAS software [59] , with the Maximum Likelihood (ML) estimation method [60] . The results of estimating regression models (3) and (4) are presented in Table 5 . These models were used to assess the post-pandemic and curfew effects on average hourly electrical load (henceforth hourly load). Figure 5 shows the parity plot based on the results of regression model (3), which suggests that the majority of predicted values were within ±10% of actual observed hourly load levels. Controlling for other variables, the results of model The results of regression model (4) mirrored the previous results as shown in Table 5 and Figure 6b . First, the hourly load declined by an average of 12.1% (p < 0.01) during the baseline hour of the post-pandemic period (compared with pre-pandemic baseline level). Differences in hourly load levels between the post-pandemic period and the pre-pandemic period ranged from a 9.0% post-pandemic decline (at 9 AM) to a 14.0% post-pandemic decline (at 11 PM). Second, the hourly load levels during the partial and full lockdown periods were consistently below pre-pandemic levels, as illustrated by the bottom row of Figure 6b . Third, differences in hourly load between the post-pandemic period and the full lockdown period were most evident during the time period from 12 PM to 7 PM, where a statistically significant curfew effect was found, leading to an additional 4.6% to 6.0% decline in hourly load (p < 0.05). Finally, differences in hourly load between the post-pandemic period and the partial lockdown period were statistically significant at the time period from 6 PM to 7 PM only. These findings support evidence that the demand for electricity declined post the COVID-19 outbreak (even when there was no curfew enforced) as found locally and globally by Alhajeri et al. [61] , Al-Abdullah et al. [62] , Abulibdeh [63] and Edomah and Ndulue [30] , Bahmanyar et al. [31] , Elavarasan et al. [35] , respectively. This can be explained by the postpandemic decline in electricity demand due to the closures of businesses, public sector agencies, and schools [30, 39, 63] . Moreover, the enforcement of partial and full curfew policies further reduced power consumption. When people stay at home, it is expected that electricity consumption would increase, especially for necessary appliances and lighting. However, air conditioning (which is the primary electricity consumer in the residential buildings in Kuwait) might not contribute to the expected increase in electricity consumption since it is always in operation even when people leave home [44] . In addition, more people in Kuwait stayed at home during the summer time (June to August 2020) due to travel bans, which might further increase the electricity demand in the residential sector for the same reason mentioned before. However, the increase in electricity consumption in the residential sector was compensated by the net decrease in electricity demand in the remaining sectors, mainly driven by the reduction in electricity consumption in the commercial and governmental sectors [30] . Unlike the context studied by Edomah and Ndulue [30] where they observed a relative increase of 3.86% in electricity consumption in the commercial sector post-pandemic (with no curfew), the immediate closure polices in Kuwait led to electricity demand reduction of up to 14.0% in the post-pandemic period (with no curfew) compared with pre-pandemic levels. This paper analyzes the impact of COVID-19 containment measures on electricity power production in the State of Kuwait. An empirical investigation was conducted to assess the impact of various lockdown measures on load production levels over one of the longest and most versatile COVID-19 lockdown periods in the world (lasting over 6 months and ranging from partial lockdown with varying curfew hours to full lockdown). The analysis used hourly longitudinal data and multilevel regression models to examine changes in electricity production levels, respectively, over the time period from 2015 to 2020. The research design used in this study has multiple advantages. First, the use of longitudinal (rather than cross sectional) data allows for examining changes in electricity production levels over time, which yields a better understanding of how and when production levels change. Second, the inclusion of five years of pre-pandemic data allows for isolating potential pandemic effects (e.g., curfew policies) from non-pandemic effects (e.g., changes in the temperature, population growth, and other monthly and yearly temporal effects). Third, the use of a multilevel framework allows for investigating changes in electricity generation levels both within a day (i.e., hourly effect) and across days (i.e., other temporal effects that vary across days) using both fixed and random effects. Finally, the variety and length of the lockdown periods allow for a detailed examination of electricity supply chain changes in response to changes in curfew policies-mimicking a quasi-experimental design. The study yielded the following conclusions: • Electricity power production during post-pandemic periods decreased by up to 16.4% compared with pre-pandemic levels (potentially due to the closure of commercial and governmental sectors). • The decline in electrical load varied within the day (depending on the hour-of-day) and across days (depending on lockdown policies). • Electricity production declined by approximately 9.0% to 14.0% during the no curfew post-pandemic period compared with pre-pandemic levels. • Curfew enforcement was associated with a further decline in electricity power production up to 16.4%, but only during peak load hours (i.e., 12 PM to 6 PM), as indicated by statistical tests that indicate the significance of the curfew enforcement further declines at the 5% level. The findings of this study can help policymakers understand the impact of a global health pandemic (such as COVID-19) on electricity supply chain, understand economic (e.g., impact on electricity subsidy cost) and environmental (e.g., impact on CO 2 emissions resulting from electricity production process) implications of such changes, and how and when the demand for electricity would be influenced by lockdown measures. The predictions can also help policymakers take precautionary measures on an hourly basis during natural disasters. Furthermore, the results can help policymakers understand how to influence electricity demand during seasons when electricity load reaches capacity of power plants (e.g., during summer season) to avoid potential power outbreaks. For example, policymakers may adopt working-from-home policies in governmental agencies to reduce electricity demand by alleviating the high electricity consumption needed to cool down governmental buildings during peak days when power outbreaks are expected to occur. Despite the thorough longitudinal multilevel analysis of the data in this work, future research could benefit from sector-level analysis, which was omitted in this work due to data collection limitations, to further examine changes in electricity demands across the residential, commercial, industrial, and agricultural sectors. In addition, within-sector analysis could explore whether policymakers can use electricity load data as a real-time indicator of (a) the impact of the pandemic on economic activity (e.g., decline in electrical demand in the commercial sector due to recession) and (b) the compliance of people with lockdown orders (e.g., whether people follow stay-at-home orders). Future research could also benefit from examining the impact of COVID-19 containment measures on other outcomes of interest (e.g., fuel supply chain, environment, economy, etc.) with different contexts (e.g., other regions with different climates and energy subsidy structures). COVID-19 pandemic lessons to facilitate future engagement in the global climate crisis A brave new world: lessons from the COVID-19 pandemic for transitioning to sustainable supply and production, Resources, Conservation, and Recycling doi An Outlook on the Global Development of Renewable and Sustainable Energy at the Time of COVID-19 What opportunities could the COVID-19 outbreak offer for sustainability transitions research on electricity and mobility? Trend analyses of electricity load changes in Brazil due to COVID-19 shutdowns Exploring the impacts of the COVID19 pandemic on global energy markets, energy resilience, and climate change Security of supply, strategic storage and COVID19: Which lessons learnt for renewable and recycled carbon fuels, and their future role in decarbonizing transport? Effects of COVID-19 pandemic on the Italian power system and possible countermeasures COVID-19 and energy: Influence mechanisms and research methodologies, Sustainable Production and Consumption Three pillars of sustainability in the wake of COVID-19: A systematic review and future research agenda for sustainable development Review on government action plans to reduce energy consumption in buildings amid COVID-19 pandemic outbreak COVID-19 and the politics of sustainable energy transitions Emergency measures to protect energy consumers during the Covid-19 pandemic: A global review and critical analysis COVID-19 crisis response in MENA countries COVID-19 energy sector responses in Africa: A review of preliminary government interventions How did the German and other European electricity systems react to the COVID-19 pandemic? Current expectations and actual values for the clean spark spread: The case of Spain in the Covid-19 crisis Analysis of the electricity demand trends amidst the COVID-19 coronavirus pandemic Impacts of COVID-19 on energy demand and consumption: Challenges, lessons and emerging opportunities Implications of COVID-19 for the electricity industry: A comprehensive review The impact of the COVID-19 on households hourly electricity consumption in Canada Impacts of the COVID-19 lockdown on energy consumption in a Canadian social housing building A cross-domain approach to analyzing the short-run impact of COVID-19 on the US electricity sector Exploring the effects of California's COVID-19 shelter-in-place order on household energy practices and intention to adopt smart home technologies Coronavirus comes home? Energy use, home energy management, and the social-psychological factors of COVID-19 Quantitative assessment of US bulk power systems and market operations during the COVID-19 pandemic COVID-19 and household energy implications: what are the main impacts on energy use? Drivers behind residential electricity demand fluctuations due to COVID-19 restrictions Energy transition in a lockdown: An analysis of the impact of COVID-19 on changes in electricity demand in Lagos Nigeria The impact of different COVID-19 containment measures on electricity consumption in Europe Electricity demand during pandemic times: The case of the COVID-19 in Spain A retrospective analysis of the impact of the COVID-19 restrictions on energy consumption at a disaggregated level Effects of COVID-19 on Indian energy consumption COVID-19: Impact analysis and recommendations for power sector operation Environmental impacts of shifts in energy, emissions What does the China's economic recovery after COVID-19 pandemic mean for the economic growth and energy consumption of other countries? When pandemics impact economies and climate change: Exploring the impacts of COVID-19 on oil and electricity demand in China Impact of COVID-19 measures on electricity consumption Impact of the COVID-19 pandemic on the US electricity demand and supply: An early view from data Impact analysis of COVID-19 responses on energy grid dynamics in Europe Pandemic Governance Requires Understanding Socioeconomic Variation in Government and Citizen Responses to COVID-19 A hybrid multi-objective optimizer-based model for daily electricity demand prediction considering COVID-19 Sustaining prosperity through strategic energy management Energy conservation in electricity generation: a case study of the electricity and water industry in Kuwait Energy and water in Kuwait Part I. A sustainability view point Desalination and water resource management in Kuwait Three years in jail, 10,000 KD fine for breaching curfew A review on implementation strategies for demand side management (DSM) in Kuwait through incentive-based demand response programs COVID-19 updates Unlocking the Impacts of COVID-19 Lockdowns: Changes in Thermal Electricity Generation Water Footprint and Virtual Water Trade in Europe An overview of the Global Historical Climatology Network temperature database Application of hierarchical linear models to assessing change Applied longitudinal data analysis: Modeling change and event occurrence NC: SAS Institute Inc., SAS/STAT 14.3 Users Guide, version 9 Econometric analysis Energy Demand in the State of Kuwait During the COVID-19 Pandemic: Technical, Economic, and Environmental Perspectives Kuwait's electric power grid Spatiotemporal analysis of water-electricity consumption in the context of the COVID-19 pandemic across six socioeconomic sectors in Doha City Validation, Visualization, Formal analysis, Investigation, Resources, Data curation, Writing -original draft Conceptualization, Visualization, Formal analysis, Investigation, Writing -original draft, Writing -review & editing, Supervision. Nawaf F. Aljuwayhel: Conceptualization, Investigation, Resources, Writing -review & editing The authors would like to thank Kuwait University for the support. Special thanks to Suhaila Marafi and Faisal Alshammari from the Engineering and Environmental Department in the Ministry of Electricity and Water (MEW) and the National Control Center (NCC) of Kuwait for providing the electricity data. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.