Calibration and Assessment of Burned Area Simulation Capability of the LPJ-WHyMe Model in Northeast China

• Main Authors: Dandan Yue, Junhui Zhang, Guodong Sun, Shijie Han
• Format: Article
• Language: English
• Published: MDPI AG 2019-11-01
• Series: Forests
• Subjects: fire; parameter optimization; lpj-whyme; simulation uncertainty
• Online Access: https://www.mdpi.com/1999-4907/10/11/992

Fire is one of the major forest disturbances in northeast China. In this study, simulations of the burned area in northeast China from 1997 to 2015 were conducted with the Lund−Potsdam−Jena wetland hydrology and methane (LPJ-WHyMe) model. The fire modeling ability in northeast China was assessed by calibrating parameters in the model. The parameters in the model were calibrated against the satellite-based global fire product (Global Fire Emission Database, version 4.1 (GFEDv4)) for the simulated burned area over the calibration period (1997−2010). Compared to the results with the uncalibrated parameters, the results obtained with the calibrated parameters in the LPJ-WHyMe model better described the spatial and interannual variability of the burned area. The spatial correlation coefficient between the GFEDv4 and the simulations increased from −0.14 for the uncalibrated version to 0.46 for the calibrated version over the calibration period. The burned area simulation ability was also improved over the validation period (2011−2015), and the spatial correlation coefficient between the GFEDv4 and the simulations increased from 0.20 for the uncalibrated version to 0.60 for the calibrated version. The mean absolute error (MAE) between the GFEDv4 and the simulations decreased from 0.018 for the uncalibrated version to 0.011 for the calibrated version (a decrease of 39%) over the calibration period and decreased from 0.020 to 0.016 (a decrease of 20%) over the validation period. Further numerical results showed that the improved simulation abilities of soil moisture and total aboveground litter had an important contribution to improving the burned area simulation ability. Sensitivity analysis suggested that determining the uncertainty ranges for parameters in northeast China was important to further improving the burned area simulation ability in northeast China.