Analyzing Carbon Emission Reduction Driven by Digital Transformation: Evidence from China's Northwest Region Using GTWR-BP Neural Network Model

Abstract

This study is based on panel data from 30 prefecture-level cities in Northwest China from 2011 to 2021, providing an in-depth analysis of the impact of digital transformation on carbon emissions, with a detailed examination from a spatial perspective. The research findings indicate that digital transformation plays a significant role in curbing regional carbon emissions, exhibiting notable spatiotemporal heterogeneity. Specifically, the average value of the carbon reduction effect of digitalization decreased from -5.0792 to -3.05602 over time, indicating a gradual weakening of the carbon reduction effect. However, neural network predictions suggest a potential rebound in the digital carbon reduction effect from 2022 to 2024, with an expected value of -0.14617 in 2022, eventually reaching -0.5063 in 2024. Despite the relatively weak foundation for digital development in Northwest China, which has led to a diminished carbon reduction effect, the ongoing advancement of digital transformation is expected to overcome technical lags, reduce energy consumption, and lower carbon emissions. The study highlights that the improvement in the level of digital transformation primarily operates through two mechanisms: industrial structure upgrading and economic level enhancement, promoting the transition of traditional industries towards low-carbon directions, while simultaneously increasing production and consumption efficiency, thereby reducing resource and energy consumption. These findings provide important references for formulating relevant policies, suggesting the enhancement of infrastructure construction and the promotion of digital transformation to achieve the "dual carbon" goals and foster sustainable economic development.