National Scale Estimation of Methane Emissions from Rice Paddies in South Korea Using UAV and Sentinel-2 Time Series Data

Methane (CH₄) has a global warming potential approximately 20 times greater than that of carbon dioxide and is a major greenhouse gas emitted from rice paddies, making systematic emission management essential for achieving carbon neutrality. In South Korea, national greenhouse gas statistics are currently derived using IPCC Tier 1 and Tier 2 emission factor–based approaches, which have limitations in accounting for region-specific rice cultivation environments and management practices. Therefore, spatially detailed estimation of methane emissions is required. To address this need, this study first estimated methane emissions using time series unmanned aerial vehicle (UAV) observations and subsequently scaled up the approach using Sentinel-2 satellite time-series imagery to quantify methane emissions from rice paddies across South Korea.

Rice paddies were identified nationwide using a phenology-based classification approach derived from Enhanced Vegetation Index 2 (EVI2) time-series composites. Sentinel-2 data were processed on the Google Earth Engine platform as five-year averaged datasets from 2020 to 2024, with 15-20 day intervals during the rice growing season, considering cloud conditions to ensure image quality. Spatial and temporal consistency was achieved through cloud masking and median compositing. To reflect regional heterogeneity in rice growth processes, region-specific rice cultivar information and growth-stage timing at the municipal level were incorporated into the analysis.

For national-scale methane estimation, the temporal behavior of Sentinel-2 derived EVI2 was evaluated through comparison with UAV-derived EVI2 observations acquired at key rice growth stages. Quantitative agreement between satellite- and UAV-based EVI2 values was limited during the early growth stages, whereas consistent temporal trends were observed after the heading stage. Based on these findings, Sentinel-2 EVI2 variables observed after the heading stage were selected as explanatory variables for the methane emission model, and UAV-based methane flux estimates were used as reference data.

The resulting empirical regression model was applied to all identified rice paddies nationwide to estimate methane emissions at both the parcel and administrative unit scales. The spatial distribution of estimated emissions exhibited pronounced regional variability, reflecting differences in rice cultivation area and growth conditions. Absolute emission estimates were slightly lower than those reported in some previous studies, a result attributed to mixed-pixel effects inherent in moderate-resolution satellite imagery. Despite this difference, the spatial patterns of methane emissions and the relative ranking among regions were generally consistent with results from comparable rice methane studies and national statistics.

This study demonstrates the applicability of a Tier 2.5 level methane emission estimation approach that integrates rice growth information derived from satellite time-series data. The proposed framework provides a scalable and cost-effective pathway for improving national greenhouse gas inventories and supporting the development of region-specific mitigation strategies for methane emissions from rice cultivation.

Acknowledgement:
This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (RS-2021-NR060142), and by the BK21 FOUR program (Grant No. 4120200313708), funded by the National Research Foundation of Korea (NRF).