open access

https://doi.org/10.14943/doctoral.k16101

https://hdl.handle.net/2115/93890

空間明示的個体ベース動的全球植生モデルの改良型火災モジュールによるシベリア林野火災とそれによるエアロゾル排出量の将来予測

英語

2024-09-25

Hokkaido University

v，144p

Wildfire is one of the most significant disturbances, globally affecting biogeochemical cycles, atmospheric chemistry, carbon cycle, and ecosystem structure and function. The effects of wildfires are becoming increasingly severe due to climate change, it is estimated that the global mean CO2 emission intensity has increased by 0.9 ± 0.9% year-1 from 2000 to 2019. Siberia has the largest forest biome and one-third of global forest cover and one of among the regions impacted by intense wildfires annually that affects the atmosphere‒land interactions of the global climate. Modeling of fire regimes using dynamic global vegetation models (DGVMs) is a key approach to analyzing these factors. However, including interactive fire disturbances in vegetation models is critical for accurately simulating vegetation dynamics. Therefore, in this study, the widely used process based SPread and InTensity of FIRE (SPITFIRE) module was integrated into the spatially explicit individual-based dynamic global vegetation model (SEIB-DGVM) to predict fire, vegetation, and burned biomass emission variables in Siberia in the future. The SEIB-DGVM was modified to add two new input variables for fire ignition: population and lightning data, and then integrated the complete SPITFIRE equation. The model was run in three phases under fire-on and fire-off mode and each phase repeated 5 time:1) a 1000-year spin-up phase to bring the soil and vegetation carbon pools into equilibrium using daily baseline Climatic Research Unit Time Series (CRU TS)3.22 climate data, 2) a 156-year historical phase also using daily baseline CRU TS3.22 climate data, and 3) a 95-year future phase using daily MIROC-AR5 base V3 RCPs 8.5, 6.0, 4.5, and 2.6 climate data. Finally, to ensure that the simulation results and the projections data are in line with the observational data, the model outputs validated by using Global Fire Emissions Database 4 (GFED4), GFED4s, European Space Agency's (ESA) Biomass Climate Change Initiative (CCI) and Global Biomass Burning Emissions Inventory (GBEI) benchmark datasets. The model is able to reproduce historical data that well compared to the benchmark datasets. Overall, on spatial comparison, the model is able to produce data with the same distribution pattern with a value of 70.7% compared to the benchmark datasets. Numerically and on a long-term average, the model is able to produce values with very high accuracy of around 99% compared with the benchmark datasets. The model estimated that until 2100, Siberia will continue to experience an increase in the frequency of forest fires. Under the RCP8.5 climate scenario, the CO2, CO, PM2.5, total particulate matter (TPM), and total particulate carbon (TPC) emissions in Siberia will continue to increase annually until 2100 by 0.295 ± 0.08 % year-1 or individually by 189.66 ± 6.55, 15.18 ± 0.52, 2.47 ± 0.09, 1.87 ± 0.06,1.30 ± 0.04 Tg species year-1, espectively. Under the same climate scenario and period comparison, we estimated that the number of trees burnt increased by 100 %, resulting in a 385.19 ± 40.4 g C m-2 year-1 loss of net primary production (NPP). The model modifications have led to a more realistic depiction of fire frequency,intensity, and extent, aligning the model outputs more closely with benchmark datasets. The major variables reached an agreement of 70.7% or greater with the observations. Additionally, the improved model accurately simulated forest structure, increasing the agreement between the simulated and observed dataset patterns and further emphasizing the reliability of the model and its emission projections.

2024-09-25

甲第16101号

博士(農学)

(主査) 教授 平野 高司, 教授 濱本 昌一郎, 助教 中島 大賢, 教授 加藤 知道 （国際食資源学院）

農学院（農学専攻）

博士論文

Version of Record