TY - JOUR
T1 - Improving the DNDC biogeochemistry model to simulate soil temperature and emissions of nitrous oxide and carbon dioxide in cold regions
AU - Cui, Guotao
AU - Wang, Junye
N1 - Publisher Copyright:
© 2019 Elsevier B.V.
PY - 2019/10/15
Y1 - 2019/10/15
N2 - The process-based DeNitrification-DeComposition (DNDC) model is widely used to quantify greenhouse gas (GHG) emissions. Soil temperature is an important environmental factor affecting nitrous oxide (N2O) and carbon dioxide (CO2) emissions, however, it is not well described in the original DNDC model due to seasonal snow cover in cold regions. This study aims to modify the original DNDC model with better representations of rain-snow partitioning, snow cover, and soil freeze-thaw cycle to predict soil temperature and GHG emissions in cold regions. Compared to the snow data in Canada, the modified DNDC model better captures snow accumulation and snowmelt with model efficiency EF of 0.64, increased from 0.14 of the original DNDC model. Soil temperature from the modified DNDC model is in good agreement with the measured data (RMSE 1.91 °C and R2 0.97), particularly when snow cover is present in winter seasons because the modified DNDC model accounts for the snow insulation effect and snowfalls above 0 °C. To further improve the simulations, the modified DNDC model runs in a command-line user interface and uses an inverse approach of optimization with a spin-up period. This modeling setup increases the R2 of CO2 emissions from 0.23 to 0.35 and the R2 of N2O emissions from 0.12 to 0.36. Investigations on different modeling setups suggest that optimization and spin-up could improve modeling results and better capture snow processes and soil temperature dynamics in the snowy cold regions, which could contribute to reasonable subsequent assessments of GHG emissions.
AB - The process-based DeNitrification-DeComposition (DNDC) model is widely used to quantify greenhouse gas (GHG) emissions. Soil temperature is an important environmental factor affecting nitrous oxide (N2O) and carbon dioxide (CO2) emissions, however, it is not well described in the original DNDC model due to seasonal snow cover in cold regions. This study aims to modify the original DNDC model with better representations of rain-snow partitioning, snow cover, and soil freeze-thaw cycle to predict soil temperature and GHG emissions in cold regions. Compared to the snow data in Canada, the modified DNDC model better captures snow accumulation and snowmelt with model efficiency EF of 0.64, increased from 0.14 of the original DNDC model. Soil temperature from the modified DNDC model is in good agreement with the measured data (RMSE 1.91 °C and R2 0.97), particularly when snow cover is present in winter seasons because the modified DNDC model accounts for the snow insulation effect and snowfalls above 0 °C. To further improve the simulations, the modified DNDC model runs in a command-line user interface and uses an inverse approach of optimization with a spin-up period. This modeling setup increases the R2 of CO2 emissions from 0.23 to 0.35 and the R2 of N2O emissions from 0.12 to 0.36. Investigations on different modeling setups suggest that optimization and spin-up could improve modeling results and better capture snow processes and soil temperature dynamics in the snowy cold regions, which could contribute to reasonable subsequent assessments of GHG emissions.
KW - Cold region
KW - DNDC model
KW - Optimization
KW - Rain-snow partitioning
KW - Snowmelt
KW - Soil temperature
UR - http://www.scopus.com/inward/record.url?scp=85067000980&partnerID=8YFLogxK
U2 - 10.1016/j.scitotenv.2019.06.054
DO - 10.1016/j.scitotenv.2019.06.054
M3 - Journal Article
C2 - 31202014
AN - SCOPUS:85067000980
SN - 0048-9697
VL - 687
SP - 61
EP - 70
JO - Science of the Total Environment
JF - Science of the Total Environment
ER -