IBIS (Integrated Biosphere Simulator) model is an ecosystem process model, which requires extensive input variables involving vegetation, soil and climate, to simulate net primary productivity (NPP) and other variables. In this study, we validated the modified IBIS based on NPP measurements between 2004 and 2005 for different vegetation types in Maoer Mountain. The results showed that the modified model can be applied to simulate NPP in the eastern area of Northeast China.We also examined the sensitivities of NPP values to the variations of the input variables in IBIS model. Sensitivity analyses were used to determine which inputs potentially cause the greatest uncertainties in calculated NPP change. The sensitivity analysis of different vegetation types, including temperate evergreen needleleaf forest, temperate mixed broad-needleleaf forest, temperate hardwood forest, temperate softwood forest, mongolian oak forest, temperate mixed forest and cool temperate Dahurian larch forest. Three input variables for IBIS, including leaf area index (LAI), temperature and precipitation, were selected for single factor analysis. The analysis was based on change rate (CR) and traditional sensitivity index (SI) with two fixed variables and the third one was given with a change of ±5% (or 1.5 ℃), ±10% (or 2 ℃) and ±20% (or 3 ℃).Taking into account the uncertainties in the values of inputs, increased LAI was predicted to be with CR between -21.2% and -0.04% on NPP output and with a SI between 0.09 and 2.37 in absolutely among case studies, whereas the decreased LAI was predicted to be -12.66% and -0.71%, 0.04 and 2.25 for CR and SI, respectively. Under both LAI increased and decreased, NPP may decrease because of light limitation, water and nitrogen cycle negative effect which may lead to a higher specific leaf area (SLA) and thus may slow down the photosynthetic rate. LAI sensitivity analysis results showed that leaf area index, especially increased or decreased 5%, had the greatest effect on NPP.
Through the climate change sensitivity analysis, it shows: NPP increases with the increase of temperature (except temperature rises up over +3℃) and precipitation, which indicates that rise in temperature can lead to higher NPP, but exist a temperature threshold; Vegetation in the studied area is a little drought and 5% for precipitation decrease had the greatest effect on NPP; Compared to LAI, LAI has the greatest effect on NPP, and then precipation, which indicates that uncertainty could be greatly reduced by calibration of LAI, it is also important to obtain accurate input data for precipitation. We also find the temperate mixed broad-needleleaf forest may serverly response to those three input parameters.
This study is just performed at one of the parameters with pre-setting changes, while the other two parameters are fixed at true values. Definitely, any parameter will responsed to the change of other parameter. Hence NPP will change with them. Therefore, in the future, we need further research on changing more parameters than one simultaneously to study the sensitivity of NPP to input parameters.