Font Size: a A A

Data Assimilation Of Global Leaf Area Index By Using Community Land Model, And Its Application In Global Climate Change

Posted on:2017-02-27Degree:DoctorType:Dissertation
Country:ChinaCandidate:X L LingFull Text:PDF
GTID:1220330485460995Subject:Climate systems and climate change
Abstract/Summary:PDF Full Text Request
The leaf area index (LAI) influences the exchanges of momentum, carbon, energy, and water balance between the terrestrial biosphere and the atmosphere. As such, it is a key variable in regulating the global carbon, energy, and water cycles. Modern land surface models have included prognostic carbon and nitrogen components to simulate LAI and other related vegetation variables, but these LAI simulation shows biases in both amplitude and phase. Such biases can be corrected by assimilating satellite-derived LAI into land surface models, which is the focus of this study.The land data assimilation system in this study is the Community Land Model version 4 (CLM4.0) with the prognostic carbon-nitrogen (CN) option (hence CLM4-CN), which is linked with the Data Assimilation Research Testbed (DART). An ensemble of 40-member atmospheric forcing fields is generated by running the DART and the Community Atmosphere Model (DART/CAM). This ensemble forcing is used to drive the CLM4-CN with or without assimilation. In our study, LAI data from the Global Land Surface Satellite (GLASS LAI) is assimilated into the CLM4-CN at a frequency of every 8 days, and the spatial resolution of 0.9°x1.25°In order to achieve ensemble initial conditions,3 steps of spin-up process were conducted to make all pools in carbon and nitrogen model reach their steady-state. First of all, single Qian’s forcing was used to drive the model for 4000 years (Shi et al.,2013). Next, the ensemble mean of 40 atmosphere forcing members for the year of 1998 was used to driving the single CLM4-CN for 1000 years. And for the last step, 40 ensemble atmosphere forcing members were used to drive the ensemble CLM4-CN members for 40 plus 4 years from 1998 to 2001, and we can finally achieve 40 ensemble initial conditions.Three experiments were designed to find out the best assimilation method, in which the experiment without assimilation is called the CTL experiment, with LAI assimilated while no carbon-nitrogen control is called the NO-CN experiment, and with both LAI assimilated and carbon-nitrogen control is called the C-N experiment. The results showed that the C-N experiment performed the best among all the designed experiment, and the assimilation results were better in low latitude regions than higher latitude regions. In detail, the CLM4-CN simulated LAI systematically overestimated the global LAI, especially in low latitude. If there was no C-N control during assimilation, the analyzed LAI is almost the same as simulation. While if updating both LAI and Leaf Carbon (and nitrogen), the analyzed LAI can be corrected.The proportion of modeling and observation is another topic of this research. The results showed that, when the LAI value and its amplitude are small, the analyzed LAI is more relied in modelling, which situation is obvious in mid and high latitude region in the northern hemisphere. In contract, if the amplitude is large for LAI value, and the bias of modelling and observation is large, the analyzed results relied more in observation.Furthermore, different assimilation mathematical algorithm, including Ensemble Adjust Kalman Filter (EAKF), Ensemble Kalman Filter (EnKF), Kalman Filter (KF), and Particle Filter (PF) were used to find out the best performance during assimilation. The results showed that the Ensemble Adjust Kalman Filter as well as including carbon-nitrogen control is the best choice.Based on the assimilated LAI and potential LAI data, the difference for land surface variables, and land-atmosphere exchange variables between assimilation and simulation are also analyzed, as to see if assimilation can improve the prediction ability of the land surface model. When LAI improves, the canopy intensity improved and decrease the solar radiation entered into the land surface; as such, net longwave radiation also decrease with the LAI increase, as well as the 2-metre high air temperature. The increased canopy density can decrease the drop off entered into the land surface, as well as the runoff. Furthermore, the increased LAI can intensify the evaporation and transpiration of vegetation, and decrease the soil moisture. In addition, the region with largest LAI variations is not the region with largest variations for the other variables, which may be associated with the land surface vegetation cover type.The effect caused by LAI variation on other variables by the model of running coupled CAM and CLM (as well as CESM) is also analyzed. The result showed that the atmosphere can enlarge the effect caused by LAI on land surface, especially for the 2-metre high air temperature, and precipitation. Furthermore, when accounting into the effect of ocean and ice, the effect caused by land surface LAI change shows its largest variation on middle and higher latitude regions.This study needs longer time designed experiment for verification.
Keywords/Search Tags:Community Land Model, Leaf area index, Data Assimilation Research Testbed, Global climate change, Carbon-nitrogen cycle
PDF Full Text Request
Related items