B24C-01 INVITED
Controls over vegetation water use in a small watershed at varying scales of time and space
As physiological ecologists worry about whether stomata respond to humidity or vapor pressure deficit, or differences between isohydric vs. anisohydric species, or the relative importance of chemical signaling from roots in controlling plant transpiration, scientists in related disciplines continue to use the Penman-Monteith equation to model evapotranspiration – typically with a static, species specific parameter for canopy conductance – and wonder what the fuss is all about. At what scales of time and space do the details matter? This presentation uses measurements of transpiration, soil moisture and meterological data from plots along a transect in a small watershed dominated by a single species, Douglas-fir to explore this question.
B24C-02
Coupled hydraulic and photosynthetic feedbacks on forest transpiration throughout the growing season
Ecosystem models account for vegetative controls on water fluxes using environmental drivers and hydraulic and/or biochemical limits on canopy stomatal conductance (Gs), variations in space and time of leaf area index (L), and species or biome specific parameters. However, some parameters, such as maximum stomatal conductance or its reference proxy at vapor pressure deficit of 1 kPa (Gsref), may not be strictly time-independent suggesting as yet undefined mechanisms in the models. We developed a model of coupled canopy water and carbon exchange, which allowed us to examine photosynthetic and hydraulic feedbacks on Gsref spanning the whole growing season for several dominant tree species in wetland and upland positions that collectively account for most a 1600 square km region centered on the WLEF AmeriFlux tower in Wisconsin, USA. The model assimilated half-hourly sap flux and micrometeorological data to quantify and explain temporal variations in Gsref for trembling aspen, sugar maple, and red pine in upland sites, and speckled alder and white cedar in wetland sites. Results show (1) phenological effects on photosynthetic activity with feedback on Gsref in all species, and (2) lags of up to two months between maximum per unit leaf area photosynthetic rates for conifer versus deciduous species. These results show that for given environmental conditions canopy transpiration depends on both L and timing of biochemical activation, both of which have implications for regional ecosystem water cycling.
B24C-03
Dynamics of Amazon Forest photosynthetic increase during 2005 drought
Coupled climate-carbon cycle modeling studies indicate that Amazon forests are vulnerable to drought, and some predict substantial carbon loss from tropical ecosystems, including the drought-induced collapse of the Amazon forest and conversion to savanna. The model-simulated future forest collapse is attributable, in part, to a forest physiological feedback mechanism which should be observable as reductions in transpiration and photosynthesis during drought years under current climates. A widespread drought occurred in the Amazon in 2005, the first such climatic anomaly since the launch of the Terra satellite's MODIS sensor in 1999, providing a unique opportunity to compare actual forest drought response to expectation on broad spatial scales. We recently reported that, contrary to expectation based on model simulations, satellite observations showed a large-scale green-up in intact evergreen forests of the Amazon in response to the 2005 drought. Here, we explore the temporal dynamics of this response, using a simple water-balance model to predict drought response. These findings suggest that Amazon forests, though threatened by human-caused deforestation and fire, may be more resilient to climate changes than many ecosystem models assume.
B24C-04 INVITED
Ecosystem-level water-use efficiency inferred from eddy covariance data: definitions, patterns and spatial up-scaling
In this presentation we discuss ways to infer and to interpret water-use efficiency at ecosystem level (WUEe) from eddy covariance flux data and possibilities for scaling these patterns to regional and continental scale. In particular we convey the following: WUEe may be computed as a ratio of integrated fluxes or as the slope of carbon versus water fluxes offering different chances for interpretation. If computed from net ecosystem exchange and evapotranspiration on has to take of counfounding effects of respiration and soil evaporation. WUEe time-series at diurnal and seasonal scale is a valuable ecosystem physiological diagnostic for example about ecosystem-level responses to drought. Most often WUEe decreases during dry periods. The mean growing season ecosystem water-use efficiency of gross carbon uptake (WUEGPP) is highest in temperate broad-leaved deciduous forests, followed by temperate mixed forests, temperate evergreen conifers, Mediterranean broad-leaved deciduous forests, Mediterranean broad-leaved evergreen forests and Mediterranean evergreen conifers and boreal, grassland and tundra ecosystems. Water-use efficiency exhibits a temporally quite conservative relation with atmospheric water vapor pressure deficit (VPD) that is modified between sites by leaf area index (LAI) and soil quality, such that WUEe increases with LAI and soil water holding capacity which is related to texture. This property and tight coupling between carbon and water cycles is used to estimate catchment-scale water-use efficiency and primary productivity by integration of space-borne earth observation and river discharge data.
B24C-05
Estimating Seasonal Changes in Volumetric Soil Water Content at Landscape Scales in a Savanna Ecosystem Using Two-Dimensional Resistivity Profiling
Water distributed in deep soil reservoirs is an important factor determining ecosystem structure of water-limited environments, such as the seasonal tropical savannas of South America. We employed a two-dimensional (2-D) geoelectrical profiling technique to estimate seasonal dynamics of soil water content to 10m depth along transects of 275-m in savanna vegetation. Resistivity values along these 2-D resistivity profiles were converted into volumetric water content (VWC) by soil depth. These resistivity profiles revealed the following soil and aquifer structure: 0-4m of permanently unsaturated and seasonally droughty soil, less severely dry unsaturated soil at about 4-7m, nearly permanently saturated soil between 7-10m, mostly impermeable saprolite interspaced with fresh bedrock of parent material at about 10-30m, and a region of highly conductive water-saturated material at 30m and below. Temporal dynamics in VWC indicate that the active zone of water uptake is predominantly 0-7m, and follows the seasonal cycles of precipitation and evapotranspiration. Uptake from below 7m may have been critical for a short period near the beginning of the rainy season, although the seasonal variations in VWC in the 7-10m layer are relatively small and lag the surface water recharge for about 6 months. Calculations using a simple 1-box water balance model indicate that average runoff was 15-20 mm/mo in the wet season and about 6 mm/mo in the dry season. Modeled ET was about 85 mm/mo in the wet season and 20-25 mm/mo in the dry season. Variation in basal area and tree density along one transect was positively correlated with VWC of the 0- 3m and 0-7m soil depths, respectively, during the wettest months. These multi-temporal measurements demonstrate that the along-transect spatial differences in soil-moisture are quasi-permanent and influence vegetation structure at the scale of tens to hundred of meters.
B24C-06
Does Hawaiian native forest conserve water? Lower uptake rates at tree and stand scale by Metrosideros polymorpha relative to plantation species
Native plants are often claimed to be conservative water users that enhance groundwater recharge compared to faster-growing non-native species that tend to dominate watersheds. This argument would have implications for motivating conservation and restoration of native forest in Hawai'i. However, few studies have examined differences in native and non-native plant transpiration (water use) at species or at stand level. Our aim was determine whether species matter to stand-level water use. We measured plant transpiration in a continuous mosaic of native forest and non-native tree plantation in Honaunau, Hawaii, focusing on endemic dominant tree Metrosideros polymorpha, alien timber trees Eucalyptus saligna and Fraxinus uhdei, and dominant understory Cibotium tree ferns. We measured xylem sap flow for six individuals of each species continuously for over eight weeks, and we estimated stand water use by scaling up these measurements using stand sapwood area and tree fern leaf area values obtained through vegetation surveys. Native forest dominant Metrosideros had the lowest rates of whole-tree daily water use at 8 kg day-1 (200kg m-2sapwood day-1), less than half the daily rates for Eucalyptus or Fraxinus; Metrosideros also had the lowest maximum transpiration rates of the three tree species. At the stand level, Fraxinus-dominated stands had higher water use than Eucalyptus- and Metrosideros- dominated stands due to the species' high sap flow rates, five-fold greater sapwood allocation, and the stands' two-fold greater dominant tree density. In Metrosideros-dominated stands, high Cibotium tree fern leaf area contributed to nearly 60% of water use, indicating the fern's critical role in forest water balance. Stand water use was influenced by factors at various scales, including species composition, stem density, tree sizes, and tree species' sapwood allocation, and was affected significantly by understory contributions. These findings highlight the importance of constituent species in forest water use, and in the case of this Hawaiian forest, indicate conservative water use by native forest.
B24C-07
Analysis of eco-hydrological control and feedback using data-derived entropic process networks
We hypothesize that plant ecosystems form self-organizing systems on the landscape which function to control their immediate surroundings towards the end of improving the efficiency of community carbon assimilation. Self- organization is difficult to define, but the concept requires the presence of feedbacks. Flows of control and feedbacks may be studied using network theory. This research uses entropy-based statistics of information flow to render the eco-hydrological system as a process network empirically derived from multivariate timeseries datasets. The resulting process network is analyzed to identify ecosystem controls and feedbacks and to separate different modes of system behavior. This approach is applied to the central corn belt eco-region using FLUXNET eddy-covariance timeseries data. Results indicate that plant respiration is a dominant controller of the interaction in the network of variables, including CO2 flux and sensible heat flux under well-watered conditions, and latent heat flux (but not CO2 flux) under drought conditions. Respiration is not controlled directly by other processes in the network, indicating that respiration is an independent (information-driven) mechanism of control by plants. Under drought conditions the ecosystem loses its ability to control CO2 assimilation through respiration, in agreement with the Ball-Berry model. CO2 flux inhabits a control feedback loop via latent and sensible heat flux, precipitation and cloud conditions, suggesting that carbon assimilation activity forms the basis of a self-organizing system spanning the Atmospheric Boundary Layer. Our finding that plants regulate their environment and CO2 uptake by modifying respiration, and that carbon assimilation feeds back on itself via atmospheric processes, supports the hypothesis that this ecosystem is self-organizing.