Spatiotemporal Modelling of Agricultural Greenhouse Gas Emissions for Sustainable Farm Management
Journal of Smart and Sustainable Farming · 22.09.2026 · 212
DOI: 10.64026/jssf/2026018 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Smart and Sustainable Farming, online erschienen am 22.09.2026. Verfasst von Ketema Niguse Baye und 1 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben.
Abstract (Originalsprache)
Greenhouse-gas (GHG) emissions from farms are highly variable in space and time: nitrous oxide (N2O) is released in short bursts after fertilisation and rain on wet soil, methane (CH4) is emitted from flooded rice, and both differ between parcels. Annual emission factors and a few chamber measurements cannot resolve this variability, which limits farm-level inventories and mitigation planning. This paper develops a spatiotemporal modelling framework for daily, parcel-level N2O and CH4 emissions that combines sparse chamber measurements, weather, soil-moisture, management and satellite-vegetation covariates with a process-guided gradient-boosting model, validated under random, spatial, temporal and spatiotemporal splits and upscaled to unsampled parcels with regression kriging. A simulated 144-ha farm with 24 chamber parcels and 3,230 chamber observations over two years was used, in which the true fluxes are known. The simulated farm emitted about 406 t CO2e per year, of which 67 % was rice methane; a mean N2O emission factor of 1.0 % was reproduced, and 40 % of annual N2O was emitted on 10 % of days. The hybrid model explained 27 % of the variance of daily N2O flux on unseen parcels in an unseen year (59 % on the square-root scale) and 80 % of CH4, whereas study-blind random splits overstated N2O accuracy by a factor of 2.1. Upscaled annual CO2e per parcel reached R² = 0.97 on unsampled parcels against 0.80 for IPCC Tier 1 factors, although the farm total was underestimated by 5 %. Chamber networks that covered all crops and soil classes gave much smaller farm-total errors than random placement when few parcels were sampled, and scenario analysis showed that water management in rice and nitrification inhibitors dominated mitigation, with N-rate reduction and application timing worth only a few per cent.
Zitieren
APA
Baye, K. N., & Dalgaard, T. (2026). Spatiotemporal Modelling of Agricultural Greenhouse Gas Emissions for Sustainable Farm Management. Journal of Smart and Sustainable Farming, 212. https://doi.org/10.64026/jssf/2026018
Vancouver
Baye KN, Dalgaard T. Spatiotemporal Modelling of Agricultural Greenhouse Gas Emissions for Sustainable Farm Management. Journal of Smart and Sustainable Farming. 2026:212. doi:10.64026/jssf/2026018
BibTeX
@article{Baye2026,
author = {Baye, Ketema Niguse and Dalgaard, Tommy},
title = {Spatiotemporal Modelling of Agricultural Greenhouse Gas Emissions for Sustainable Farm Management},
journal = {Journal of Smart and Sustainable Farming},
year = {2026},
pages = {212},
doi = {10.64026/jssf/2026018}
}
EUVIMED