Summary: | 학위논문 (석사)-- 서울대학교 대학원 : 농업생명과학대학 협동과정농림기상학, 2018. 2. 류영렬. Vegetation in the high latitude ecosystem is most responsive to climate variables, leading to high year to year variability of gross primary productivity (GPP). Therefore, understanding the spatiotemporal patterns of GPP and how climate variables drive its interannual variability (IAV) is important to account for their present and future status. In this study, we examine the spatiotemporal patterns of Alaskan GPP and further investigate how their relation to climate drivers. We use GPP derived from four different approaches, a process-based approach (Breathing Earth System Simulator), a semi-empirical approach (Moderate Resolution Spectroradiometer 17A2) and the machine-learning approaches (Support Vector Regression and FLUXCOM). Model evaluation with eddy covariance data from 17 sites showed that the models explained 65% to 85% of the monthly variation with relative bias ranging from -22% to 33%. Model performance was better in the boreal forest compared to tundra and fire disturbed ecosystems. The spatial and temporal variation of GPP in the models displayed a consistent pattern, where the deciduous broadleaf forest showed the highest variability of GPP IAV by 14%, followed by fire and evergreen forest (13%) and then tundra (10%). Tundra accounted for the largest fraction of IAV of GPP with 55%, exceeding evergreen needleleaf forest (38%), deciduous broadleaf forest (7%) and areas that had been disturbed by fire (0.8%). GPP in tundra has the smallest variation among the PFTs. 68% of Alaska is tundra which led to the largest contribution to the IAV of GPP. The IAV of GPP from 2001 to 2011 had a similar pattern to the IAV of both air temperature and radiation, where warmer years had a larger GPP anomaly compared to the colder years. Therefore, warming and cooling as a result of climate change could significantly impact the IAV of land-atmosphere interaction of carbon dioxide. 1. Introduction 1 2. Material and Method 5 2.1 Study Region 5 2.2 Flux Tower Data 6 2.3 Satellite-based GPP Datasets 9 2.4 Dataset of climate variables 14 2.5 Landcover map 15 2.6 Evaluation and analysis of GPP 16 3. Results 20 3.1 Evaluation of Models against flux tower data 20 3.2 IAV of GPP 23 3.3 Relationship between IAV of GPP and Climate Variables 31 4. Discussion 37 4.1 Model Performance across different PFTs 37 4.2 IAV of GPP 39 4.3 Controlling factors in IAV of GPP 41 5. Conclusion 43 References 44 Abstract in Korean 53 Master
|