Dynamic nowcast of the New Zealand greenhouse gas inventory

Malcolm Jones,Hannah Chorley,Flynn Owen,Tamsyn Hilder,Holly Trowland,Paul Bracewell

DOI: https://doi.org/10.1016/j.envsoft.2023.105745

2024-02-17

Abstract:As efforts to mitigate the effects of climate change grow, reliable and thorough reporting of greenhouse gas emissions are essential for measuring progress towards international and domestic emissions reductions targets. New Zealand's national emissions inventories are currently reported between 15 to 27 months out-of-date. We present a machine learning approach to nowcast (dynamically estimate) national greenhouse gas emissions in New Zealand in advance of the national emissions inventory's release, with just a two month latency due to current data availability. Key findings include an estimated 0.2% decrease in national gross emissions since 2020 (as at July 2022). Our study highlights the predictive power of a dynamic view of emissions intensive activities. This methodology is a proof of concept that a machine learning approach can make sub-annual estimates of national greenhouse gas emissions by sector with a relatively low error that could be of value for policy makers.

Machine Learning,Atmospheric and Oceanic Physics

What problem does this paper attempt to address?

The main objective of this paper is to address the issue of time lag in New Zealand's greenhouse gas emissions reporting. Specifically, the paper attempts to solve the problem through the following approaches:

1. **Background and Challenges**:
- New Zealand's National Greenhouse Gas Inventory (NZGGI) has a data lag of 15 to 27 months.
- Accurate and timely greenhouse gas emissions data are crucial for assessing the progress of international and domestic emission reduction targets.

2. **Methods**:
- The paper proposes a machine learning-based approach to "nowcast" New Zealand's national greenhouse gas emissions to reduce the delay in data release.
- This method can reduce the delay to only 2 months, determined by the current data availability.
- The study employs machine learning algorithms such as random forest and extra trees, and cross-validation is used to select the best model.
- Specific indicator variables are used to predict emissions from the two main sectors: energy and agriculture.

3. **Key Findings**:
- The study demonstrates that machine learning methods can achieve sub-annual estimates of national greenhouse gas emissions by sector with relatively low error.
- The models in the paper perform well in terms of prediction accuracy, model stability, and overfitting.
- For example, as of July 2022, the estimated total national emissions have decreased by about 0.2% since 2020.

4. **Significance and Applications**:
- This dynamic estimation method can improve policymakers' understanding and response speed to changes in emissions.
- It provides a proof of concept that machine learning can be used to improve the timeliness of greenhouse gas emissions reporting, which is of significant value for monitoring emission reduction progress.

In summary, this paper addresses the significant time lag in existing greenhouse gas emissions reporting by introducing a new machine learning method, aiming to improve the timeliness and accuracy of emissions data to better support the achievement of emission reduction targets.