Rice CH4 inventory and validation data + code for Southeast Asia (0.1 deg, 2020-2021)

Data products and Python pipeline for a bottom-up rice methane (CH4) inventory at 0.1 deg resolution for nine Southeast Asian countries (SEA-9: Indonesia, Timor-Leste, Malaysia, the Philippines, Vietnam, Thailand, Cambodia, Laos and Myanmar), produced under the IPCC 2019 Refinement Tier-2 framework with a 10 m Sentinel-derived rice prior (Open-SEA-Rice-10) and a Peat-ML histosol partition for the 2020-2021 window. This Zenodo record provides: (i) The derived 0.1 deg netCDF inventory and counterfactual fields (bottom-up Tier-2 emissions, area-prior-only counterfactual, enhancement against EDGAR v8.0). (ii) A 2000-member Monte Carlo ensemble distributed as the parameter matrix and the SEA-9 totals (NPZ format), together with per-country statistics and diagnostics. (iii) Sequential and Shapley decompositions of the EDGAR v8.0 vs Tier-2 gap as CSV tables, plus the corresponding decomposition against the EDGAR Community GHG 2025 release. (iv) Sensitivity analyses for the country water regime (+/- 25 percentage point envelope), the peat treatment (six configurations), and the cell-country intersection mask (polygon-over-centre vs fractional remapping). (v) The TROPOMI/Sentinel-5P annual-mean XCH4 column field for 2020-2021 retrieved through Google Earth Engine, the cleaned version of the same field (1500-2100 ppb mask), and the Pearson/Spearman correlations of the bottom-up and EDGAR v8.0 emission fields against TROPOMI XCH4 at both cell and country aggregate scales. (vi) The full reproducibility pipeline (22 Python scripts, Python 3.11, no compilation required), the requirements file, a citation file (CITATION.cff) and the Zenodo metadata file (.zenodo.json). Headline numbers: SEA-9 central total 6.61 Tg CH4 yr-1 (Monte Carlo 5-95 percent interval 3.86-12.31 Tg yr-1), 33.3 percent below EDGAR v8.0 averaged over 2020-2021 and 13.3 percent below the EDGAR Community GHG 2025 release. The bottom-up matches Indonesia's official UNFCCC BUR3 submission to within 10 percent and correlates more strongly with the TROPOMI XCH4 column field than EDGAR v8.0 does (Pearson r = +0.279 vs +0.108 at the cell scale, n=12,270; +0.407 vs +0.334 at the country aggregate, n=9). External datasets used as inputs are cited but not redistributed here: Open-SEA-Rice-10 (Ginting et al. 2025, doi:10.5281/zenodo.14627003), EDGAR v8.0 / 2024 / 2025 releases (JRC EDGAR portal), Peat-ML (Melton et al. 2022, doi:10.5194/gmd-15-4709-2022), Natural Earth admin-0, TROPOMI/Sentinel-5P L3 CH4 (Google Earth Engine collection COPERNICUS/S5P/OFFL/L3_CH4), and the UNFCCC Biennial Update Reports of Indonesia, Vietnam, Thailand and the Philippines. The peer-reviewed manuscript and its figures are distributed through the journal Remote Sensing Applications: Society and Environment (Elsevier, manuscript RSASE-D-26-01078, under revision) and are not included in this Zenodo record. Companion to: Goycochea Casas, G. and Ismail, Z.H. "Emission factors dominate the rice methane gap with EDGAR in Southeast Asia: a 10 m harvested-area assessment", Remote Sensing Applications: Society and Environment (under revision).