Скачать презентацию Patterns and controls of water-use efficiency in an Скачать презентацию Patterns and controls of water-use efficiency in an

16c83b0316a91d79e3ff4e53702c8950.ppt

  • Количество слайдов: 1

Patterns and controls of water-use efficiency in an old-growth coniferous forest Yueyang Jiang 1, Patterns and controls of water-use efficiency in an old-growth coniferous forest Yueyang Jiang 1, John B. Kim 2, Christopher J. Still 1, Bharat Rastogi 1, Steve Voelker 3, Frederick Meinzer 2 1 Department of Forest Ecosystems & Society, Oregon State University, Corvallis, OR; 2 Pacific Northwest Research Station, USDA Forest Service; 3 Department of Plants, Soils and Climate, Utah State University, Logan, UT Oregon State University Introduction Results Water use efficiency (WUE) has been widely recognized as an important physiological link between carbon and water cycling, and used to track forest ecosystem responses to climate change and rising atmospheric CO 2 concentrations. Numerous studies have been conducted to estimate forest-scale WUE through either stable carbon isotope analyses composition or eddy covariance measurements. However, the sign and magnitude of WUE response to climate variability are still highly uncertain, and can vary with the time scale of analysis. This study employed the Ecosystem Demography model version 2 (ED 2) to explore patterns and physiological and biophysical controls of WUE in an old-growth coniferous forest in Pacific Northwest. Long-term eddy covariance flux measurements and the stable carbon isotope composition of CO 2 collected at the Wind River Ameri. Flux site 1. The calibrated ED 2. 1 -WR model well captured carbon and water fluxes, and both model and eddy covariance data indicated that the Wind River forest functioned as carbon sink during 1998 -2015. 2. The unique physiological properties of Pacific Northwest late-successional conifer trees play an important role in determining old growth forest demography, thereby largely affecting ecosystem carbon budget. Figure 5. The ED 2. 1 -WR modeled Ci/Ca ratio under clear and cloudy sky conditions for northern pine (left) and late conifer (right) trees. were used to validate model performance. ED 2. 1 -WR 52 m Wind River tower 19 m Figure 3. ED 2. 1 -WR simulated and tower derived 18 years average daily GPP, Reco and NEP at the Wind River Canopy Crane Research Facility site. Eddy covariance W. hemlock Douglas-fir Figure 1. The Wind River Canopy Crane Research Facility, and species dominancy. Methods We calibrated the ED 2 model to the Wind River forest site by adjusting several key physiological parameters for two dominant PFTs (further referred as ED 2. 1 -WR). Based on eddy covariance, isotope data and ED 2. 1 -WR simulations, we Rin 1) Evaluate ED 2. 1 -WR performance in simulating carbon, water and energy fluxes. 2) Characterize WUE, sapflow and stomatal responses (ci/ca) in wet and dry years. 3) Contrast WUE, sapflow and stomatal response between species & between cohorts. 4) Rank relative importance of varied predictors in determining WUE using machine learning methods. Rout eatm Tatm Figure 4. Comparison of ED 2 modeled and field measured daily transpiration (mmol/m 2/s) for year 2002 at Wind River Canopy Crane Research Facility site. CO 2 ecanopy Tcanopy CO 2 esurface CO 2 Tsurface esoil_n Tsoil_n esoil_i Tsoil_i esoil_1 Tsoil_1 Figure 2. ED 2 framework derived from Medvigy et al. (2009). 3. Different PFTs have distinct WUE, sapflow and stomatal responses to short-term meteorology and radiation including clear and cloudy days, and long-term climate. 4. Multiple linear regression and machine learning methods indicated that the relative importance of predictor for WUE shown different patterns in ED 2. 1 -WR and eddy covariance data. Figure 6. Relative importance of different predictors in determining WUE. Ongoing Work 1. To explore how various WUE metrics (e. g. , intrinsic WUE, inherent WUE) from both measurements and model predictions vary with meteorology through different time scales. 2. To investigate the relative importance of different predictors in determining various WUE metrics, using machine learning methods. 3. To scale up forest WUE across the Pacific Northwest and conducted model simulations (e. g. , ED 2. 1 -WR, 3 -PG and FVS-Climate) to project future WUE under projected climate scenarios.