Development of New Earth System Model and Investigation of the Impacts of Vegetation on Hydrological Cycle Using Developed ESM

Abstract

This thesis is to develop a version of the earth system model (ESM) with the interactive vegetation parameterization scheme and investigate the influences and feedback processes driven by terrestrial vegetation on climate and global hydrological cycle. Previous observation and modeling studies suggested that the temporal variation of atmospheric CO2 concentration depends largely on the carbon uptake by terrestrial biosphere, compared with that by oceanic counterpart. In the state-of-the-art ESMs, the parameterization uncertainty in the carbon cycle between land and atmosphere is still large, which tend to produce a significant spread in the model simulations for the ambient CO2 amount and its secular trend. The huge spread of CO2 in turn causes a wide spread in the simulated climate through vegetation-climate feedbacks, which make their future climate projection less accurate and less reliable. One aspect of the parameterization uncertainties lies in the nitrogen process, a critical process coupled with the carbon cycle by limiting supplement of nutrients both to subsurface soil organic matter (SOM) and plants at the surface, which is either absent or represented poorly in current ESMs. A few ESMs with a primitive version of the carbon-nitrogen (C-N) coupling exhibit a significant model deficiency such as the underestimation in the carbon pools or storages in the vegetation and sub-surface soil systems and substantially weak carbon uptake, whereas the other models without C-N coupling tend to overestimate them. The interaction and feedbacks between vegetation and climate should be, therefore, understood based on a more improved understanding of the C-N cycle and its representation in the ESMs. Before developing new ESM, the model intercomparision using 11 CMIP5 ESMs has been done first to evaluate the overall representation of the terrestrial biogeochemical cycle and the model dependences. Using the MODIS satellite estimates, the study validates the simulation of gross primary production (GPP), net primary production (NPP), and the carbon use efficiency (CUE) depending on plant function types (PFTs). The models show noticeable deficiencies from MODIS in the simulation of horizontal patterns of GPP and NPP as well as large simulation differences, although their multi model ensemble (MME) mean represents realistic global mean value and the spatial distributions. Larger model spreads in GPP and NPP than those in surface temperature and precipitation suggest that the simulation difference in terrestrial carbon cycle is largely attributed to the uncertainties in the dynamic vegetation model parameterizations. The models also exhibit large differences in the simulation for CUE geographically and at the change of dominant PFTs, primarily due to the differences in parameterization. While the MME of CUE shows strong dependence on surface temperature, the observed CUE from MODIS shows more complex and non-linear sensitivity. To developing new ESM, this study uses the Geophysical Fluid Dynamics Laboratory Earth System Model version 2 (GFDL-ESM2M) as a base model for the further development. GFDL-ESM2M has a full capability of interactive vegetation parameterizations, and with decent simulations in the long-term climatology and climate variability compared with other ESMs in the Fifth Phase of the Coupled Model Intercomparison Project (CMIP5). The model has comprehensive carbon cycle both for the terrestrial and aquatic environments, yet with no C-N coupling. In order to implement the C-N cycle in the model, this study replaces the Land Model version 3 of GFDL-ESM2M with the Community Land Model version 4 (CLM4). The latter land model is the most recent version in the development suits by NCAR, with more detailed vegetation types and biogeochemistry as well as the interactive C-N parameterizations. However, the Community Earth System Model (CESM) with CLM4 has a well-known bias of the pronounced underestimation in the terrestrial carbon uptake. This is one of the major motivations in this research. The developed ESM named as UNIST-ESM shows the reasonable simulation skill for climate conditions and terrestrial carbon fluxes. The features of distribution of carbon cycle is depended onto land surface models and the features of spatial distribution of climate conditions is dominant to characteristics for atmospheric model in UNIST-ESM. In terrestrial carbon cycle, UNIST-ESM still has systematic bias for simulating GPP in the globe. The underestimation (overestimation) of GPP over high latitude region (tropics) in UNIST-ESM is major deficiencies of simulation of terrestrial carbon cycle. For improvement of deficiencies of terrestrial carbon cycle in UNIST-ESM, this study develops a new parameterization method for determining Q10 by considering the soil respiration dependence on soil temperature and moisture obtained by multiple regression. This study further investigates the impacts of the new parameterization on the global carbon cycle budget. Our results show that non-uniform spatial distribution of Q10 tends to enhance heterogeneous anomaly of soil respiration comparing with the control simulation with uniform Q10. Moreover, it tends to improve the simulation of observed relationship between soil respiration and soil temperature and moisture, particularly over cold and dry regions. The new parameterization improves the simulation of gross primary production (GPP) by reducing bias in the global mean comparing with the FLUXNET-MTE observation data. Besides, GPP over high latitudes is significant increased by about two times from the control simulation. The realistic Rs and GPP simulation induced to represent carbon balance between release at the subsurface and uptake at the surface over terrestrial biosphere reasonably. Overall, enhanced heterogeneous temperature sensitivity in the soil decomposition process in the model showed the improvement of production and respiration. For evaluation of vegetation feedback on the climate, this study investigated the vegetation-climate feedback and impacts of vegetation change on the hydrological cycle in the East Asian monsoon region using comprehensive and developed UNIST-ESM. The intensity of East Asian summer monsoon (EASM) in the future increase due to enhance moisture flux at the surface by increased vegetation in EA. However, low-level relative humidity in the future decreases due to relatively increasing temperature comparing with enhanced moisture increase over EA. This process induces to suppress the formation of low-level cloud in the future. Therefore, enhanced incoming solar radiation over China regions is occurred by suppressed low-level cloud in the future climate. On the other hand, the exception of additional anthropogenic heating in future climate, the role of increase vegetation on the variation of EASM is opposite signal from the future climate scenarios. In terms of local hydrological process, the increased vegetation induces incareasing evapotranspiration and surface moisture flux over vegetated regions. This increment of moisture source from increased vegetated area tends to increase formation of low-level cloud due to small perturbation of temperature increase in this simulation. Enhanced amount of low-level cloud tends to decrease incoming solar radiation at the surface due to reflecting by clouds. It emphasized surface cooling in the increased vegetation regions. This surface cooling also affects to depress local vertical updraft over China where has large plants. Suppressed vertical updraft makes atmosphere stabilization over land area. The convective precipitation is affected to decrease by depressed vertical updraft associated with cooling temperature. In the focus of the large-scale circulation, the cooling temperature over vegetated area tends to decrease the land-sea contrast. Thermally more uniformed land and ocean surface make weak low-level circulation and meridional moisture transport over EA regions. This study suggested that the roles of vegetation to climate variation and hydrological cycle are homogenous not only location but also environmental climate conditions.