Climate change and depletion of fossil fuel are two of the major global challenges calling for urgent actions. Localised generation of renewable energy such as wind power has been adopted by farms as an effort for decarbonisation. It is important to develop the capability to accurately predict wind power generation featured by intermittence and fluctuation so that optimal renewable development plans can be formulated. In this work, the Autoregressive Distributed Lag modelling approach was employed to study the influences of economic and environmental factors (pressure, wind speed, temperature, and electricity price) on wind power generation on a Scottish farm. The proposed Autoregressive Distributed Lag model well explain the wind power generation with an accuracy of 91.8%. The results showed that when wind speed increases by 1%, the wind power output increases by 0.256% in the long run. We forecasted a total wind generation capacity of 1894.9 MWh from September 2020 to September 2021 based empirical environmental and economic data. In this case, the annual carbon emission of on-farm wind power usage was estimated to be 5.3664 tonnes. The on-farm wind power generation would reduce the electricity-related carbon emission by 278.87 tons over the 13 months.
Keywords sustainable development, renewable energy, wind power, ARDL model.