Current Research Themes include:

• Deconvolution of belowground ecosystem processes.
• Climatic controls on alpine and sub-alpine soil and ecosystem processes.
• Labile carbon, tree growth and mortality, and forest and woodland dynamics.
• Synthesis of the functional ecology of trees.
• Carbon and water dynamics in deserts of the Southwest.
• Scaling of plant hydraulic architecture.
• Bayesian modeling in ecology.

We are developing biophysically-inspired methods for unraveling belowground processes. Two specific problems that we are addressing include: (1) teasing-apart the contributions of different sources (e.g., heterotrophs vs. autotrophs) to soil carbon fluxes; and (2) reconstructing plant water sources and active rooting distributions.

Plant water sources. Ogle et al. (2004) developed the RAPID (Root Area Profile and Isotope Deconvolution) algorithm, a semi-mechanistic Bayesian deconvolution modeling approach for estimating plant water uptake and rooting profiles. Data on stable isotopes of hydrogen and oxygen (δD, δ¹⁸O) in plant and soil water are used to reconstruct plant water sources and rooting patterns. The RAPID algorithm overcomes the nonidentifiability problem associated with traditional “simple linear mixing models,” whereby the number of water sources that can be estimated is limited by the number of isotope members used. The RAPID algorithm overcomes the nonidentifiability problem and can estimate many sources (and continuous profiles) by incorporating a biophysical model for root water uptake into a Bayesian framework such that the biophysics, empirical data, and prior distributions (for root-related parameters) place biologically realistic constraints on the water uptake and root area profiles. Additionally, the modeling approach and results provide insight into the mechanisms associated with plant water uptake and active root dynamics. We are developing dynamic (nonlinear) Bayesian models to estimate temporal changes in rooting behavior and implications for water uptake patterns and responses to precipitation events.

Soil respiration. We are currently extending the Bayesian deconvolution approach to the problem of partitioning soil respiration into its component sources. The relative contribution of different sources to soil CO₂ efflux can vary within the soil matrix, depending on spatial and temporal variability in soil properties, resource and substrate availability, and microclimate. Our approach allows for simultaneous analysis of multiple data sources related to soil respiration dynamics, and the data are analyzed within the context of process-based models. The data include repeated measurements of soil CO₂ efflux, stable isotopes of carbon (δ¹³C) in the emitted CO₂, δ¹³C of potential sources, relative activity or concentration of different sources, and soil properties (e.g., bulk density, temperature, water availability). The process-based models include flux equations that describe respiration rates of different sources within the soil profile and flux-based, mechanistic isotope-mixing models. We are using the approach to partition the components soil respiration in desert ecosystems characterized by episodic precipitation and highly variable carbon flux dynamics. The field data were collected by Jessie Cable (post-doc in the Ogle lab, former Huxman lab member) and Travis Huxman’s lab at the University of Arizona. The data are representative of sites typical of the Sonoran Desert, including grasslands on clay and sand-dominated soil in the Santa Rita Experimental Range, and a riparian grassland site along the San Pedro River Basin undergoing mesquite encroachment. We are coupling the deconvolution algorithm with the field data to explore how shifts in precipitation regimes (e.g., precipitation pulse timing, frequency, and magnitude) affect soil respiration across different sites differing in properties, between and within seasons, and associated with different plant species (e.g., native grasses, invasive grasses, and encroaching shrubs).

Alpine and sub-alpine ecosystems are sensitive to environmental change due to a short growing season, extreme (low) temperatures, and the potential reliance on both summer rainfall and winter snowfall. Thus, it is possible that changes in seasonal moisture inputs (rain and snowfall) that are predicted due to global warming may impact ecosystem and hydrological processes in these systems by, for example, changing plant community structure and soil biochemistry. To develop a better predictive understanding of how these high-elevation ecosystems will respond to climate change, it is imperative that we characterize historical and potential ecosystem responses to environmental variability. This is the motivation behind the studies that we are beginning at the Glacier Lakes Ecosystem Experiments Site (GLEES) in the Snowy Range of southeastern Wyoming.

We are collecting preliminary data to characterize the riparian ecosystems along chains of small lakes within the GLEES. The ecosystem types that we are currently examining include meadow (wet, dry, scree) and upland coniferous forest (30-60% cover). We are collecting data related to soil physical properties, soil water and nutrients, soil microbial communities, and plant functional traits. We will use the data to explore how soil properties vary within the watershed and to develop hypotheses about spatial variability in relevant ecosystem properties. The plant data will provide insight into plant carbon-water-nutrient relations, how they are affected by a plant’s position within the watershed, and how they are related to soil characteristics. We will use these data to inform future studies aimed at teasing-apart the linkages between plant functional traits (e.g., water uptake patterns, growth, allocation, etc.), soil properties (e.g., available N, P; water), and ecosystem processes (soil microbial dynamics; coupling of soil and plant processes). Our ultimate goal is to learn how these linkages and processes (1) vary within the watershed due to shifts in the relative contribution of winter vs. summer water inputs, (2) are affected by atmospheric nitrogen deposition and buffering of N deposition by lakes, (3) may be controlled by climatic thresholds that mark seasonal transitions, and (4) how they may be affected by future changes in seasonal and annul precipitation and temperature.

The aim of this project is to improve our understanding of how forests and woodlands are affected by environmental change. Towards this goal, we are developing a process-based, quantitative framework for linking tree physiology and growth to community and ecosystem dynamics. For example, we are developing the Allometrically-Constrained Growth and Carbon Allocation (AC-GC) model, an individual-based model that explicitly links labile carbon dynamics, tree allometries, growth, allocation, and mortality. The AC-GC model includes: (i) a mechanistic basis for predicting tree death, (ii) an allocation scheme that simultaneously satisfies flexible allometries and physiology-based carbon dynamics, and (iii) a range of physiological conditions that are consistent with real tree behavior (e.g., healthy, static, shrinking, recovering, recovered, dying trees). In collaboration with Jarrett Barber (University of Wyoming, Statistics), we beginning to build a hierarchical Bayesian melding algorithm that links the process-based AC-GC model to diverse data sources, including a forest inventory database of stem diameters and heights for over 300 tree species in the U.S. and a literature database of species-specific physiology and allometries (see Synthesis of functional ecology of trees).

Allocation of labile carbon to leaves, fine roots, and sapwood is nearly impossible to measure, and it is difficult to identify functional relationships governing allocation. However, our AC-GC modeling approach does not require explicit allocation functions; allocation is a product of the model whereby labile carbon is allocated in such a way that the structural carbon dynamics are consistent with allometric relationships (e.g., height:diameter, sapwood area:leaf area). Key assumptions and predictions include: (i) allocation is dynamic, varying with plant age and environmental conditions; (ii) labile carbon storage capacity is determined, in part, by xylem anatomy; (iii) retranslocation of labile carbon occurs in order to satisfy excessive respiration demands; (iv) tree mortality is coupled to carbon starvation (depletion of stored carbon); (v) trade-offs exist such that species with high capacity to store labile carbon are also those that grow slow but are more likely to survive extended periods of stress compared to species with a low capacity to store carbon (grow fast, but susceptible to short-term stress). For example, the model reproduces growth and mortality patterns of temperate forest species that are consistent with their shade-tolerance and successional strategies. So far, which project has been a collaborative effort with Steve Pacala’s lab at Princeton University, and we have also been interacting with Christian Wirth’s lab at the Max-Planck Institute for Biogeochemistry. And, the work has focused on producing a modeling and scaling framework for linking individual-tree and species-specific physiological and structural traits to community and ecosystem dynamics of eastern temperate forests.

The Ogle Lab will extend the AC-GC model, database, and Bayesian melding work to explore the individual- and species-specific traits (or processes) that control woodland dynamics in pulse-driven, water-limited regions in the western U.S. This work will lend insight into species-specific characteristics - especially as related to labile carbon traits - that underlie drought-tolerance strategies and susceptibility to mortality during severe drought episodes. This work is also paramount to predicting the potential effects of climate change on semiarid woodlands because extreme precipitation events, such as anomalous droughts, are expected to become more frequent due to global warming.

We are developing and compiling a huge database of functional traits for 300+ tree species that occur in the United States. We are developing a relational database in MS Access to store data that we are mining from the literature related to physiological and structural traits of mostly temperate, but some boreal and Mediterranean, tree species. The database contains functional traits such as, but not limited to, specific leaf area (SLA), wood density, photosynthetic and respiration parameters, and demographic traits. Christian Wirth’s group at the Max-Planck Institute for Biogeochemistry is conducting a parallel effort whereby they are developing a nearly identical database for tree species found throughout temperate and boreal forests of Europe, South America, Asia, and potentially New Zealand. Both databases will eventually be merged and synthesis efforts will be carried-out to understand the global functional ecology of temperate and boreal trees.

The Ogle Lab is developing statistical methods for synthesizing data contained in such databases. The methods involve modern computation statistics and Bayesian approaches that are expected advance meta-analysis and data synthesis effort in the field of ecology (e.g., comparative ecology, ecosystem ecology, functional ecology, etc.). Another goal of this synthesis effort is to produce species-specific estimates of the various structural and functional traits. Such estimates will be extremely valuable to modeling studies that require good, empirically-based parameter estimates that, for example, describe plant functional traits. We are also initiating more focuses data mining efforts to learn about the functional ecology of arid and semiarid woodland species, and how species-specific traits scale-up to affect woodland dynamics under episodic and changing environmental conditions. The database work is also a key part of the project related to Labile carbon, tree growth and mortality, and forest dynamics.

This study is motivated by the potential for altered precipitation regimes (an element of climate change) to greatly impact water-limited ecosystems of the Southwest. Existing field studies exploring the effects of variation in precipitation on C (carbon) and H₂O (water) dynamics in deserts of the Southwest have produced large datasets. However, no syntheses have been conducted with respect to these datasets to explore the broader implications of altered precipitation for arid and semiarid ecosystems. This is a collaborative project with Travis Huxman (University of Arizona), David Tissue (Texas Tech), Michael Loik (University of California, Santa Cruz), and Stan Smith (University of Nevada, Las Vegas). Our objective is to synthesize existing data related to C and H₂O fluxes, spanning leaves to ecosystems, across four major deserts in the Southwest. A second objective is to use the synthesis results to identify novel questions and hypotheses and to help direct future, cross-site experiments.

The datasets are derived from five extensively-studied sites: Santa Rita Experimental Range and San Pedro River Basin, AZ (Sonoran Desert, PI=Huxman), Big Bend National Park, TX (Chihuahuan Desert, PI=Tissue), Nevada Test Site, NV (Mojave Desert, PI=Smith), and Valentine Eastern Sierra Reserve, CA (Great Basin Desert, PI=Loik). The large datasets contributed by the PIs from each site include C and H₂O exchange of leaves, canopies or whole plants, soils, and ecosystems. The synthesis results will be used to address a series of questions, including: How are C and H₂O cycles affected by changes in pulse, seasonal, and annual precipitation? What ecological components appear most critical to ecosystem C and H₂O exchange, and which are most sensitive to precipitation changes? With respect to C and H₂O, how do the four deserts differ in their responses to altered precipitation? What environmental variables (e.g., nitrogen, temperature) interact with precipitation to strongly affect C and H₂O dynamics in these systems?

The synthesis work will be conducted in the Ogle Lab. We will synthesis data from these sites within a Bayesian hierarchical framework that facilitates simultaneous coupling of diverse data sources and mechanistic models; the models contain ecologically-meaningful parameters that, when parameterized with field data, provide important insights into the factors controlling C and H₂O dynamics at different scales. The synthesis results will also be used to inform future field experiments. Feedback to field studies is central to designing experiments that explore effects of multiple global change factors on deserts of the Southwest, and terrestrial ecosystems in general. This study will contribute to building a better understanding of how arid lands may be affected by environmental changes such as altered precipitation.

Kiona Ogle is collaborating with Joshua Weitz (Georgia Tech) on a project that couples empirical and modeling studies to examine variation in xylem properties of trees. This work was motivated by a series of theoretical models that have been proposed to explain the structure and allometry of plant vascular systems, including formulations based on the pipe model, Murray’s law, and resource distribution through hierarchical branching networks. These models yield contrasting predictions about how xylem conduit diameter should taper with plant height or branching order, and their predictions and assumptions remain to be thoroughly tested. The lack of empirical data inspired us to develop a simple method for preparing and measuring numerous conduit diameters from trunk and branch cross-sections obtained from ring-porous trees (Henry Horn at Princeton developed the method). We employed the method to quantify the distribution of vessel diameters in tree rings of cross-sections obtained at different trunk heights for white ash (Fraxinus americana) (Weitz et al. 2006). By comparing the distribution of vessel diameters within a given tree ring at different heights in the trunk we were able to explore vessel tapering as a function height. We showed that vessel radii are determined by distance from the top of the tree, as well as by stem size, independent of tree height or age. The qualitative form for the scaling of vessel radii agrees remarkably well with simple power laws, and it suggests an ontogenetically stable hydraulic design that scales in the same manner as a tree grows in height and diameter. Our work is providing information about how the hydraulic architecture of a plant changes as it ages (or grows), potentially reflecting shifting trade-offs between mechanical stress and resistance to water flow. Ongoing work is exploring how a tree’s hydraulic architecture is constrained by different cavitation mechanisms. We are working on a mathematical model that links hydraulic design with the risk of freezing-induced cavitation. We are using this model and literature data to explore the potential constraints of freezing-induced cavitation on tree height, and to identify optimal xylem tapering that may facilitate growth of taller trees while avoiding catastrophic embolisms.

Nearly all of the above research themes involve data analysis and synthesis efforts. The Ogle Lab often employs Bayesian methods for linking large and diverse data sources to process-based models that quantifying underlying biological, biophysical, and ecological mechanisms, interactions, and feedbacks. We are also working on a variety of other (smaller) problems that integrate mathematical models and Bayesian statistical methods. Some applications include: methods for inferring vulnerability of the plant water transport system (xylem) to cavitation (embolism, loss of transport capacity); statistical methods for estimating scaling or allometric relationships; and, near-real time data analysis and model-data-experiment feedback.