Global Spatial and Temporal Dynamics of Photosynthesis and Transpiration
A central question in ecology is to understand what controls biological variation in space and time -- why are organisms where they are and how do species function and adapt to their respective environments? Our ability to observe spatial and temporal patterns of biological activity at the local scale far exceeds our ability to observe such patterns at the global scale. Satellite remote sensing offers one of the only methods available to determine global spatial and temporal patterns of plant physiological activity. However, direct observation of vegetation-atmosphere fluxes of carbon and water is not currently possible. Satellites can be used to determine fluxes of solar radiation as well as radiation reflected and emitted from the land surface. In addition, satellites can provide measures of light absorptance by vegetation (fAPAR) as well as the fractional vegetation cover (fc). However, ecophysiological theory and models are still required to translate all these observations into meaningful estimates of plant and ecosystem function. Such theory and models are being developed and applied on a global scale by integrating over 25+ years of satellite remote sensing observations acquired by the AVHRR and MODIS sensors with ground-based tower flux micrometeorological measurements and process-based ecophysiological studies. This approach makes it possible to quantify the role of terrestrial ecosystems in the global cycles of carbon and water and to examine the response of ecosystem function to changes in climatic conditions, land use, and biodiversity. See http://landflux.org for more information. See Tu and Fisher 2004, Su et al. 2004, Tu and Fisher 2006, Leuning et al. 2008, Fisher et al. in press for evapotranspiration, Tu 2000, Tu 2002, Misson et al. 2006 for photosynthesis and photosynthetic capacity, Tu et al. 2007 for water use-efficiency, ci/ca, and carbon isotope discrimination.
Spatial Mapping of Keeling Plot Data Using Artificial Neural Networks
The "Keeling plot" method has proven to be a robust and highly informative measure of ecosystem- atmosphere interactions, particularly with respect to photosynthesis, respiration and water use efficiency of terrestrial ecosystems. Applied over many years and locations, the archive of Keeling plot data is steadily increasing, especially in light of recent coordinated collection efforts and advances in laser-based technologies. However, meta-analyses of this valuable and potentially informative record remains challenging because of the discontinuous nature of the largely campaign-based and site-specific collections over the years. One of the main objectives of the Biogeosphere-Atmosphere Stable Isotope Network (BASIN) is to facilitate the synthesis and exchange of stable isotope information related to ecosystem processes in carbon and water cycles at various scales. Towards this goal, we have initiated a BASIN-wide effort for routine synthesis of past and future Keeling plot data in the context of an objective and statistically based approach using an artificial neural network (ANN) to help elucidate coherent patterns in the inherently disparate data. Predictive relationships between Keeling plot intercepts and climate and vegetation developed with this method can help to not only reveal patterns in the data that may lead to future process-based research, but can also provide the means to efficiently translate site-specific, campaign-based data into spatial and temporally continuous maps of Keeling plot intercepts. Using this data-intensive approach, the ANN can be continually updated to increase its accuracy and resolution as new data from more sites becomes available. We will describe the various sites and datasets currently available (BASIN, SIBAE, DOE-TCP, etc.), results related to the training and site-specific validation of the ANN, functional responses of Keeling plot intercepts to environmental conditions and vegetation status as revealed through the ANN, and finally, spatial maps produced with the ANN when applied with global meteorological data and satellite observations of vegetation status. See Tu et al. 2008.
Partitioning Ecosystem Respiration Between Plant and Microbial Sources Using Stable Carbon Isotopes
Keeling plot intercepts integrate all respiratory signals in an ecosystem. Contrary to the longstanding notion of isotopic equilibrium among respiratory components, we have found significant variation in the carbon isotope ratio of CO2
respired from leaves, stems, roots and soil microorganisms across contrasting ecosystems in California (redwood forest,
chaparral, grassland, oak savanna, ponderosa pine plantation). The largest differences are always between leaf and
microbial respiration. Based on these differences, ecosystem respiration can be partitioned between plant and microbial sources. Further study is underway to determine the underlying cause of these isotopic differences Dawson et al. 2002, Tu and Dawson 2003, Tu and Dawson 2005, see also Isotope Fractionation, Biosynthesis and Respiration.
Effects of Rainfall Variability on Carbon Cycling in a California Grassland
Climate models predict that the amount and timing of rainfall in California will change during the next century. We performed both field and laboratory studies to understand the consequences of altered precipitation patterns on the flux and 13C of ecosystem respiration in a water-limited grassland in California. Seasonal changes in the isotopic composition of whole ecosystem respiration were strongly related to the amount and timing of precipitation, and the nature of this relationship depended on the seasonal dynamics of plant growth, litter decomposition, and the depth distribution of water in the soil profile. Whole ecosystem respiration was dominated by plant respiration in the winter wet months and by microbial decomposition during the dry months of summer and early fall. Microbial activity appeared to shift to progressively deeper depths in the soil profile throughout the summer and fall following the availability of soil moisture. Results indicate that surface litter and soil organic matter should be represented as two distinct pools in models attempting to capture the dynamics of soil respiration in response to climate change. See Brener et al. 2005.
Wireless Sensor Networks - A Macroscope in the Redwoods
Wireless sensor network “macroscopes” offer the potential to advance science by enabling dense temporal and spatial monitoring of large physical volumes. In collaboration with David Culler's group in Computer Science and the Intel Research Lab at UC Berlekey, we are working on various projects related to applications of wireless sensor networks in environmental biology. Initial tests have been made using about 30 'motes' measuring spatial gradients of light,temperature and humidity within in a 70m tall redwood crown. See Tolle et al. 2005.