Elsevier

Energy

Volume 189, 15 December 2019, 116351
Energy

An assessment of global electric power consumption using the Defense Meteorological Satellite Program-Operational Linescan System nighttime light imagery

https://doi.org/10.1016/j.energy.2019.116351Get rights and content

Highlights

  • Global electric power consumption was estimated with nighttime light data.

  • Good degree of accuracy was achieved against statistical data in China and USA.

  • Temporal variations of electricity consumption were observed in East Asia.

  • Electric power consumption decrease was observed in Russia and the Western Europe.

  • Energy efficiency improvement in Asia contributed to global energy use reduction.

Abstract

Industrialization and urbanization have led to a remarkable increase of electric power consumption (EPC) during the past decades. To assess the changing patterns of EPC at the global scale, this study utilized nighttime lights in conjunction with population and built-up datasets to map EPC at 1 km resolution. Firstly, the inter-calibrated nighttime light data were enhanced using the V4.0 Gridded Population Density data and the Global Human Settlement Layer. Secondly, linear models were calibrated to relate EPC to the enhanced nighttime light data; these models were then employed to estimate per-pixel EPC in 2000 and 2013. Finally, the spatiotemporal patterns of EPC between the periods were analyzed at the country, continental, and global scales. The evaluation of the EPC estimation shows a reasonable accuracy at the provincial scale with R2 of 0.8429. Over 30% of the human settlements in Asia, Europe, and North America showed apparent EPC growth. At the national scale, moderate and high EPC growth was observed in 45% of the built-up areas in East Asia. The spatial clustering patterns revealed that EPC decreased in Russia and the Western Europe. This study provides fresh insight into the spatial pattern and variations of global electric power consumption.

Introduction

The rapid progress of urbanization, population increase, and economic development has led to the increase of electric power consumption (EPC) during past four decades. Worldwide, per capita EPC increased from 1200 kWh in 1971–3126 kWh in 2014 [1]. Despite its important role in supporting economic activity [2], the increased EPC has accelerated global warming and climate change due to its long-run association with greenhouse gas emissions [3]. Decarbonization of power sector could play an important role in carbon emission reduction and climate change mitigation actions in national contexts [4]. Future projections implied that deployment of low-carbon technologies in electricity sector could make great contributions to emission reduction in developing countries such as India [5]. Information on EPC and how it changes spatially and temporally are thus important for policy makers to improve energy efficiency and reduce carbon emissions [6]. However, the primary source of EPC data is census data reported at the scale of administrative unit; such data may even be unavailable for many developing countries.

The nighttime light (NTL) data obtained by the Defense Meteorological Satellite Program’s Operational Linescan System (DMSP-OLS) have served as a proxy for a number of socio-economic variables, including population, Gross Domestic Product, and electric power consumption. Elvidge et al. [7] reported that the lit area detected by DMSP-OLS imagery was highly correlated to EPC and suggested potential applications of DMSP-OLS data for modeling EPC. Lo [8] demonstrated a logarithmic relationship between DMSP-OLS data and statistical EPC data for 35 Chinese provincial capital cities. Amaral et al. [9] observed a linear relation between the EPC from public and commercial illumination and the lighted area from DMSP imagery in the case of Brazilian urban settlements. Chand et al. [10] reported the correlation between increase of power consumption and increase in number of nighttime lights in India. Efforts to compare DMSP-OLS nighttime light imagery against ground-based electrical infrastructure and electricity use data have further confirmed the close correlation between nighttime light imagery and electrification in urban and rural areas [11,12].

Mathematical models have been developed to measure and monitor electric power consumption using nighttime light imagery at multiple spatial scales (Table 1). The linear correlation between cumulative digital number (DN) values of corrected NTL images and EPC was used to estimate energy consumption in Asian countries such as Japan, India and China by Letu et al. [13]. Zhao et al. [14] created disaggregated maps of EPC in China in 1995, 2000 and 2005, using the urban population identified by the areal extent of lighting from annual NTL composites. Cao et al. [15] built a linear model to map annual EPC in China with DMSP-OLS intensity and population data as independent variables. He et al. [16] used linear regression models between the cumulative DN value of saturation-corrected NTL image and provincial statistic EPC to estimate pixel-by-pixel EPC in Mainland China annually from 2000 to 2008. The linear regression model was further applied to Enhanced Vegetation Index (EVI) and population-adjusted NTL data to assess the spatiotemporal dynamics of electricity consumption in urban cores and suburban regions in China from 2000 to 2012 [17]. Two recent studies for global EPC mapping examined changes of spatiotemporal changes in energy consumption at large scales. Shi et al. [18] subdivided the world into 48 regions and performed a linear regression model for each region to map global EPC at 1 km resolution. Hu & Huang [19] combined different correction methods for nighttime light data and regression models to produce an improved global 1 km gridded EPC. Nonetheless, both studies mainly used the information from NTL data and barely considered other factors that can affect the relationship between EPC and NTL. Satellite imagery obtained from the Visible Infrared Imaging Radiometer Suite Day Night Band (VIIRS DNB) scanning radiometer aboard the Suomi National Polar-Orbiting Partnership satellite has better qualities than NTL data, and was used to model energy use in the United States [20]. To accommodate for the possible nonlinear relationship between electricity consumption and NTL and other predictors, Jasiński [21,22] made the first effort to model EPC with VIIRS DNB using artificial neural networks.

Despite its extensive applications in EPC modeling, the saturation problem of DMSP-OLS nighttime data in urban centers limits its application for accurately estimating EPC at local scales. Different techniques for saturation correction have been developed to adjust OLS data prior to applying it for EPC prediction (Table 1). Letu et al. [13] used a cubic regression equation to correct saturated stable light and reported a higher correlation between the corrected data and EPC in Japan, China, India and 10 other Asian countries. Townsend and Bruce [23] reported an increase of 25.4% in accuracy estimation of the electricity consumption in Australia by applying an Overglow Removal Model to reduce the blooming effect of NTL data. Using provincial population data, Cao et al. [15] derived a linear model to correct saturated pixels in time-series of nighttime light data and increased the accuracy of EPC estimates for China. Shi et al. used the calibrated global radiance nighttime light data from 2006 as a reference for reducing the saturation effect in NTL data, and detected the spatiotemporal dynamics of EPC at global [18], regional [24] and local city scales [25]. In addition to approaches using empirical models, a combination of the normalized difference vegetation index (NDVI) and DMSP-OLS NTL data has been demonstrated to be very effective at increasing the variation in NTL luminosity [26]. In contrast to NDVI, Jasiński [22] obtained high correlation between energy consumption and the total useable floor area of dwellings, and suggested built-up area may be a better variable for electricity consumption modeling.

The increasing availability of global geospatial data in recent years has made it possible to optimize NTL correction and improve NTL-based EPC estimation at the global scale. The objective of this study was to provide better understanding of the spatiotemporal dynamics of global EPC at different spatial scales by using DMSP-OLS NTL data and the newly available global population and built-up data.

Section snippets

Datasets

The datasets used in this study mainly consisted of statistical EPC data, DMSP-OLS NTL data, Gridded Population Density data (GPW V4), global human settlement layer (GHSL), Landsat Enhanced Thematic Mapper Plus (ETM+) images, and administrative boundaries (Table 2).

The annual statistics for electric power consumption at the country level were collected from the World Bank Open Database and were used to calibrate the EPC estimation models. Chinese energy consumption data (excluding Tibetan

Results

In the following sections, the improvement of adjusted NTL was first demonstrated, followed by accuracy assessment of EPC estimation, and spatial and temporal analysis of EPC. The pixel-level electric power consumption was estimated for 2000 and 2013 in Section 3.1 Evaluation of adjusted nighttime light data, 3.2 Accuracy assessment of estimated electric power consumption. The adjustment of nighttime light data and electricity consumption estimation results were compared and evaluated with

Discussion

In this study, NTL data was intercalibrated and adjusted by using population and built-up area data to map and analyze the EPC across the world. Previous studies have demonstrated the usefulness of built-up and population data to minimize and to even eliminate saturation to produce spatial electricity consumption maps [17,22]. The GPW V4 population layers were released in 2014 and updated in 2015 and 2017. These provide more detailed and precise population grids compared to previous versions of

Conclusions

In this study, a methodology for estimating global electric power consumption at a 1 km spatial resolution was developed. The original NTL data were inter-calibrated and adjusted using both GPW population and global built-up area data to reduce the saturation effect. EPC estimation models were further developed and calibrated using statistical data from the World Bank for six regions globally in 2000 and 2013. Our analyses showed that satisfactory results could be achieved when they were

Acknowledgements

The research was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (No. XDA19090107); the National Key Research and Development Program of China (No.2017YFE0100800); and the National Natural Science Foundation of China (No. 41471369).

References (51)

  • T. Hu et al.

    A novel locally adaptive method for modeling the spatiotemporal dynamics of global electric power consumption based on DMSP-OLS nighttime stable light data

    Appl Energy

    (2019)
  • D. Fehrer et al.

    Spatial distribution of building energy use in the United States through satellite imagery of the earth at night

    Build Environ

    (2018)
  • T. Jasiński

    Modeling electricity consumption using nighttime light images and artificial neural networks

    Energy

    (2019)
  • K. Shi et al.

    Exploring spatiotemporal patterns of electric power consumption in countries along the Belt and Road

    Energy

    (2018)
  • K. Shi et al.

    Evaluating spatiotemporal patterns of urban electricity consumption within different spatial boundaries: a case study of Chongqing, China

    Energy

    (2019)
  • Q. Zhang et al.

    The vegetation adjusted NTL urban index: a new approach to reduce saturation and increase variation in nighttime luminosity

    Remote Sens Environ

    (2013)
  • S. Leyk et al.

    Assessing the accuracy of multi-temporal built-up land layers across rural-urban trajectories in the United States

    Remote Sens Environ

    (2018)
  • Y. Xie et al.

    Spatiotemporally enhancing time-series DMSP/OLS nighttime light imagery for assessing large-scale urban dynamics

    ISPRS J Photogrammetry Remote Sens

    (2017)
  • A. Huete et al.

    Overview of the radiometric and biophysical performance of the MODIS vegetation indices

    Remote Sens Environ

    (2002)
  • Q. Zheng et al.

    A new source of multi-spectral high spatial resolution night-time light imagery—JL1-3B

    Remote Sens Environ

    (2018)
  • Z. Zhu et al.

    Understanding an urbanizing planet: Strategic directions for remote sensing

    Remote Sens Environ

    (2019)
  • C. Zhang et al.

    Panel estimation for urbanization, energy consumption and CO2 emissions: a regional analysis in China

    Energy Policy

    (2012)
  • K. Jayanthakumaran et al.

    CO2 emissions, energy consumption, trade and income: A comparative analysis of China and India

    Energy Policy

    (2012)
  • H. Kim et al.

    Energy demand and supply, energy policies, and energy security in the Republic of Korea

    Energy Policy

    (2011)
  • U. Al-Mulali et al.

    The effect of energy consumption, urbanization, trade openness, industrial output, and the political stability on the environmental degradation in the MENA (Middle East and North African) region

    Energy

    (2015)
  • Cited by (25)

    • Estimating natural disaster loss using improved daily night-time light data

      2023, International Journal of Applied Earth Observation and Geoinformation
    • Automated Extraction of Energy Systems Information from Remotely Sensed Data: A Review and Analysis

      2022, Applied Energy
      Citation Excerpt :

      Some other studies performed estimation on a country level, estimating energy consumption in Tanzania [211] and India [213]. Lu et al. [214] did it at a lower geographic resolution where estimation was conducted on country, continental, and global scales. Energy Information: Energy Consumption Estimation Using Other Remote Sensing Data.

    • Research on energy conservation and carbon emission reduction effects and mechanism: Quasi-experimental evidence from China

      2022, Energy Policy
      Citation Excerpt :

      Most research results argue that PM2.5 particulate matter emissions are mainly derived from human energy consumption activities, such as industrial production, coal and other fossil energy consumption, and automobile exhaust (Lin and Wang, 2016). Some studies have shown that night city lights are highly correlated and consistent with carbon emissions (Lu et al., 2019). If the LCCP policy effectively reduces PM2.5 and city night-lighting levels, it shows that the policy has achieved energy conservation and carbon emission reduction.

    View all citing articles on Scopus
    View full text