Dataset on bitcoin carbon footprint and energy consumption

Due to data limitations on bitcoin-related emissions, assessing the environmental impacts of bitcoin appear difficult. This data in brief article presents constructed daily frequency dataset on bitcoin annualised carbon footprint spanning July 7, 2010 to December 4, 2021 with 4,158 observations. The 12 data variables capture floor, ceiling, and optimal annualised carbon footprint from coal, oil, gas, and the average from the 3 sources. The constructed bitcoin carbon footprint data are measured in kgCO2 using emission factors for electricity generation from IEA World Energy Outlook. The data will benefit multidisciplinary research on cryptocurrency from environmental, energy, and economics disciplines.


Specifications
Economics, Econometrics and Finance Specific subject area  Cryptocurrency and Fintech  Type of data  Tables, and Figures  How the data were acquired Bitcoin energy consumption data were acquired from the Cambridge Centre for Alternative Finance. The raw data is processed and modelled to produce bitcoin carbon footprint using STATA (version 16) and R (version 4.1.2) software. Data format Raw and analysed data formats submitted alongside the data article Description of data collection The daily frequency data capture 4,158 observations from July 7, 2010 to December 4, 2021. First, the raw data has 3 variables namely minimum, maximum, and optimal bitcoin annualised energy consumption. The energy consumption variables capture the total annual electricity consumption of the Bitcoin proof-of-work consensus network expressed in kilowatt-hours (kWh). The annualised measure of electricity assumes continuous power usage (i.e., minimal, maximal, and optimal) over one year period--with a subsequent application of a 7-day moving average to control for short-term hash-rate variabilities

Value of the Data
• The dataset consists of daily frequency measurements on bitcoin annualised carbon footprint with huge data points spanning July 7, 2010 to December 4, 2021. • The dataset can facilitate empirical research on environmental and energy sustainability of bitcoin, thus, improving the global debate. • The data can benefit multidisciplinary research on cryptocurrency from environmental, energy, and economics disciplines. • The estimation of bitcoin carbon footprint using global parameters makes it generally applicable and reusable in any crypto-based studies on bitcoin sustainability assessment. Table 1 presents the data description of the 12 data variables constructed using 3 initial raw data from CBECI [1] . The dataset comprises daily frequency variables with their units of measurement. Notes: The raw data were converted from the original measurements in TWh to kWh before constructing the emission dataset using IEA emission factors.   Table 2 presents the descriptive statistical analysis of data variables showing the mean, median, maximum, minimum, standard deviation, skewness, kurtosis, and Jarque-Bera test. Notes: * denotes the rejection of the null hypothesis of normal distribution. JB is the Jarque-Bera test for assessing normal distribution. Fig. 2 shows the annualised bitcoin energy consumption measured in kWh. Using Fig. 2 , the maximum, minimum, and optimal energy consumption of the bitcoin network is compared using the Games-Howell test.  Fig. 3 shows the annualised minimum bitcoin carbon emissions measured in kgCO 2 . Fig. 3 compares the distribution of the constructed minimum carbon footprint from coal, oil, gas, and average of the 3 energy sources.  Fig. 4 shows the annualised maximum bitcoin carbon emissions measured in kgCO 2 . Fig. 4 compares the distribution of the constructed maximum carbon footprint from coal, oil, gas, and average of the 3 energy sources.   5 shows the annualised optimal bitcoin carbon emissions measured in kgCO 2 . Fig. 5 compares the distribution of the constructed optimal carbon footprint from coal, oil, gas, and average of the 3 energy sources.   6 presents the effect of counterfactual change in energy consumption on carbon emissions in the bitcoin network. The change in bitcoin carbon footprint was estimated using the dynamic ARDL simulations--an empirical procedure expounded in Sarkodie and Owusu [2] to examine the relationship between energy consumption and carbon footprint based on the bitcoin network.

Experimental Design, Materials and Methods
Following the estimation procedure presented in Stoll, Klaaßen and Gallersdörfer [3] , the bitcoin carbon footprint [kgCO 2 ] CF is calculated as: Where EC denotes energy consumption [kWh] and EF represents emission factor [kgCO 2 /kWh] that captures carbon intensity of the energy mix namely coal, oil, and gas. Thus, Eq. (1) underpins the daily frequency data on bitcoin carbon footprint constructed using STATA (version 16) and R (version 4.1.2) software. The raw data from CBECI [1] were converted from the original measurements in TWh to kWh before developing the emission dataset using emission factors from IEA World Energy Outlook 2017 Annex A Tables for Scenario Projections. The global emission factors for coal, oil, gas and average are 1.019, 0.854, 0.514, and 0.554 kgCO 2 /kWh, respectively. Which is nearly closer to emission factors presented in de Vries, et. al, [4] . To construct the bitcoin carbon footprint, the following assumptions corresponding to the outlined emission factors are made: First, energy used for data centres and mining equipment regardless of hardware type is derived solely from coal. Second, energy used for data centres and mining equipment regardless of hardware type is exclusively from oil. Third, energy used for data centres and electricity for mining equipment regardless of hardware type is derived specially from gas. Fourth, energy used for data centres and electricity for mining equipment regardless of hardware type is derived from all the energy mix.
Based on the four assumptions, each of the 3 scenarios of energy consumption namely minimum, maximum, and optimal power consumption are subsequently multiplied by the four emission factors to develop 12 daily frequency data variables of bitcoin carbon footprint. However, caution should be taken in using the bitcoin emission dataset, as the global emission factors are static, yet, emission factor differs across countries. For example, the emission factor for coalbased electricity for Bitcoin miners is ∼50% higher than the global average [4] .

Declaration of Competing Interest
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.

Data Availability
Dataset on bitcoin carbon footprint and energy consumption (Original data) (Figshare).