NSF IRC (Ojima)
NSF 2004 – 2006
Atmospheric CO2 is an important control over climate, and because of recent political initiatives a currency of considerable consequence. From a global perspective, understanding the role of terrestrial ecosystems in the carbon cycle is crucial (Schimel 1995, Schimel et al. 1996, 1997, Houghton et al. 1998). From a domestic perspective, understanding the US carbon budget is a foundation requirement for sound planning and ecosystem management. Great progress has been made in recent years in measurement networks (e.g., LTER, AmeriFLUX, SOMnet), in the re-analysis of inventory data and in modeling (VEMAP 1995, Schimel et al. 1997, Paustian et al. 1996) relevant to the US C budget. In addition, ‘inverse’ analyses of atmospheric data (Enting et al. 1995, Ciais et al. 1995) are beginning to provide information on continental scales (Rayner et al. in press, Fan et al. 1998). Carbon research requires a high degree of integration between disciplines; and, the tools of ecology and atmospheric science have reached a point where an ambitious synthesis is feasible. We can now develop an integrated data-model system that allows us to analyze consequences of different assumptions about biology and land management on spatial patterns of atmospheric CO2 and its stable isotopic composition. This is a powerful complement to traditional model testing against site-specific data, but the development of measurement networks and gradient studies (Hunt et al. 1996, Baldocchi et al. 1996) also greatly improves the power of in situ data-model comparisons (Schimel et al. 1997). Thus the time is ripe for an ambitious integrated analysis of terrestrial ecosystems and the carbon cycle.
The carbon cycle links biological, geochemical and atmospheric processes. For several decades, uncertainty has persisted about a major term in the carbon cycle, the so called ‘missing’ sink. The missing sink is associated with biological processes and stems from an imbalance in the estimated global carbon budget and is widely assumed to result from these processes in Northern Hemisphere ecosystems. However, while atmospheric inversion methods have persistently identified a CO2 sink in Northern Hemisphere ecosystems (and one recent analysis places the bulk in North America; Fan et al. 1998), direct evidence from field studies, inventories and process modeling cannot fully account for this flux. It is unacceptable that the degree of uncertainty and conflict regarding the terrestrial sink has persisted for over a decade. To date, strategies for identifying the sink have been largely directed by geoscientists, have relied on local process studies, or have been reliant only on models without adequate validation.
Our basic hypothesis is that the magnitude and spatial distribution of net carbon fluxes (storage or release) between ecosystems and the atmosphere is controlled mainly by prior disturbance and land use, controlling the sensitivity of ecosystem carbon storage to ‘ecosystem physiological’ controls by CO2 , climate, nitrogen deposition and other factors.
Studies of the terrestrial carbon cycle pose a serious methodological problem of scale. Processes in terrestrial ecosystems exhibit high variability in time and space, yet from the perspective of the global carbon cycle, we are interested in ecosystem’s aggregate impact on the atmosphere. Spatial variability is high enough that measurements alone cannot provide adequate estimates of either fluxes or pools over large regions, implying that models must be used for interpolation of observations. This is particularly true given our basic postulate: that prior disturbance and land use dominate present-day fluxes. This is because disturbance and land use vary in a fine-grained fashion. Yet, without regional observations, such models cannot be convincingly evaluated. Several types of measurements provide regional data not derived from point measurements. Remote sensing techniques are our principal source of wall-to-wall data and have advanced dramatically over the past few years, yet provide only weak constraints on carbon budgets (providing, basically, information on photosynthetic potential and phenology: Running et al. 1994, 1995, Hunt et al. 1996, Asner et al. 1998). Atmospheric CO2 measurements can be analyzed to produce estimates of regional to continental net CO2 exchange, but have wide confidence intervals and almost no spatial resolution (Enting et al., 1995, Fan et al., 1998).
We propose an innovative approach to dealing with this problem of scale mismatch by designing a model-data fusion system to allow simultaneous validation at multiple scales. We will develop a terrestrial ecosystem model drawing on the scientific advances in ecosystem modeling over the past decade (VEMAP 1995, Pan et al. 1998, Schimel et al 1997) coupled to an atmospheric model.