Collaborative Research: Abrupt changes in the structure and function of forest and woodland ecosystems in the Western United States: non-linear responses to changes in climate, CO2 levels and fire regimes


Intellectual Merit: This Our interdisciplinary research effort proposes towill test the hypothesis that forest and woodland ecosystems in the Southwestern US are vulnerable to rapid non-linear changes in ecosystem composition, structure and function in response to predicted projected changes in climate and climatic variability, including warmer temperatures and changes in the precipitation regime. Evaluating the above hypothesis is challenging because tTheir structure and composition species composition of these systems are determined by multiple interactingons among causal agents including climate, fire disturbance history, and pathogen disturbances, making it difficult impossible to predict from direct observation the the long-term, large-scale consequences of correlated changes in these factors simply from observations. We will test the hypothesis that predicted changes in climate and climatic variability will cause rapid non-linear changes in the composition, structure and function of forest and woodland ecosystems in the Southwestern US through the development of a constrained, regional, terrestrial biosphere model capable of assimilating a diverse, comprehensive suite of long- and short-term observations available for Southwestern ecosystems. The model will be used in conjunction with a broad array of observational constraints that span timescales ranging from hours (eddy flux measurements) to decades (ecological data) and centuries (tree ring data) toThe key to testing the central hypothesis is consequently to embody each sub-hypothesis in a mechanistic, process-based terrestrial biosphere model that is capable of representing the interactions among the full set of possible explanatory factors, and then to test the model against comprehensive, multi-scale, multi-dimensional observations spanning timescales of hours (eddy flux data), decades (ecological measurements), and centuries (tree rings). The cross-cutting constraints provided by comprehensive retrospective observations in this model framework remove the degrees of freedom that allow the notorious “tuning” of models, allowing us to construct strong quantitative constraints on the interactions postulated in the hypotheses.

We will test 5 inter-related sub-hypotheses: to explain that the rapid changes in ecosystem composition that occurred during the 1950s, and again in 2000-2003 were caused by, testing the effects of changes in: precipitation (H1), temperature (H2), and rising atmospheric CO2 concentrations (H3), the history of fire management (H4), and beetle outbreaks (H5). The resulting constrained biosphere model will be capable of simulating ecosystem physiology and will link changes in ecosystem structure and function to individual and community response to external forcing.

The model will provideformulation will improve ourd understanding of the causes of rapid shifts changes in ecosystem composition and function on ponderosa and pinyon-juniper ecosystems, and thusgiving us the capability to more accurately forecast the ecosystem consequences effects of future environmental changes. We will then use apply the constrained regional biosphere model to predict the long-term fate of Southwestern ecosystems over the next hundred years under a range of future climate, CO2 and fire-suppression scenarios, and. We will also couple the biosphere model to a state-of-the-art mesoscale meteorological model examine the potential role of regional-scale biosphere-climate feedbacks in reinforcing or mitigating changes in ecosystem composition that occur in response to climate variability. This work will test an additional sub-hypothesis (H6), that reductions in evapotranspiration provide positive feedback for the transition from Ponderosa® Pinyon-Juniper® Juniper-grassland, increasing aridity and inhibiting reversibility.

Broader Impacts:

This research will provide much-needed knowledge about the underlying causes of observed rapid changes in the composition, structure and functioning of Southwestern ecosystems. A key feature of this analysis is to simultaneously consider the impacts of land-use management, fire suppression, climatic effects, succession, and pests and their interactions. Our goal is to develop the capability to predict non-linear changes in vegetation that may arise from increasing climate variation and climate trends. In addition to educating graduate and undergraduate students in a highly interdisciplinary research project, this project will provide educational materials for K-12 students about the dynamic linkages between climate and atmospheric composition, ecosystem composition and disturbance regimes.

Results from Most Relevant Prior NSF Support

ATM- 0221850 BE/CBC: Continental, Landscape and Ecosystem Scale Fluxes of Atmospheric Carbon Dioxide (CO2) and Carbon Monoxide (CO) Gases 01/01/03-6/31/07 S. Wofsy, D. Hollinger & P.R. Moorcroft $1,600,000 This ongoing project is has improving improved our understanding of terrestrial carbon sinks by linking process-level biological information obtained for individual plants or ecosystems over short time scales with observations and models that characterize the large spatial domain and long time scales of regional and global concern. As part of this project we haveWe developed RAMS-ED2 a new, regional-scale coupled atmosphere-ecosystem model (see below, Methodology: RAMS-ED2) that was. The RAMS-ED2 model has been used to integrate data collected across a range of spatial and temporal scales, including tower and aircraft CO2 and water flux measurements and forest biomass inventories, into a constrained biosphere-atmosphere model that accurately captures both slow and fast ecosystem carbon dynamics in New England and Quebec. The regional-scale constrained biosphere model has wasbeen tested against both atmospheric and ecosystem observations, demonstrating that it can quantitativelyly link properties of the terrestrial biosphere-atmosphere system with the underlying fundamental biological and physical processes, and. It is now being used to determine the principal sources of long-term trends and interannual variability in carbon fluxes. Publications resulting from this research include:

Moorcroft, P.R. (2006) (Invited Feature). How close are we to a predictive science of the biosphere? Trends in Ecology and Evolution (in press). doi:10.1016/j.tree.2006.04.009

Matross, D.M., A. Andrews, M. Pathmathevan, C. Gerbig, J.C. Lin, S.C. Wofsy, B.C. Daube, E.W. Gottlieb, J.T. Lee, C. Zhao, P.S. Bakwin, J.W. Munger, and D.Y. Hollinger. Estimating regional carbon exchange in New England and Quebec by combining atmospheric, ground-based, and satellite data. Tellus 57B, in press.

Ise, T and Moorcroft, PR (2006). The gGlobal-scale temperature and moisture dependencies of soil organic carbon decomposition: analysis using a mechanistic decomposition model Biogeochemistry (in press).

Albani, M., P.R. Moorcroft, G. C. Hurtt The contributions of land-use change, CO2 fertilization and climate variability to the carbon sink in the Eastern United States. Global Change Biology (accepted).

Pathmathevan M., S.C.Wofsy, D.M. Matross, X. Xiao, A.L. Dunn, J.C. Lin, C. Gerbig, J.W. Munger, V.Y. Chow, E. Gottlieb (2006). A Satellite-Based Biosphere Parameterization for Net Ecosystem CO2 Exchange: Vegetation Photosynthesis and Respiration Model (VPRM), Global Biogeochemical Cycles (in review)

Medvigy, D., P.R. Moorcroft, M. Albani, R. Avissar, R. L. Walko. (2006) RAMS-ED2: a Coupled Atmosphere-Ecosystem Model: Formulation and Results (in prep.)

Urbanski, S., C. Barford, S. Wofsy, C. Kucharik, E. Pyle, J. Budney, K. McKain, D. Fitzjarrald, M. Czikowsky, J. W. Munger (2006). Factors Controlling CO2 Exchange on time scales from hourly to decadal at Harvard Forest (in prep.submitted to J. Geophys. Res.)

Medvigy, D., P. R. Moorcroft, R. Avissar, and Robert L. Walko. (2005). Mass Conservation and Atmospheric Dynamics in the Regional Atmospheric Modeling System (RAMS). Environmental Fluid Mechanics 5: 109-134.

Moorcroft, P.R. (2003). Recent advances in ecosystem-atmosphere interactions: an ecological perspective. Proceedings of the Royal Society Series B: 270:1215-1227.


Recent studies suggest that Forests forests in the Southwestern US may be particularly vulnerable to rapid, non-linear changes in ecosystem composition in response to climate change. It was observed (Allen and Breshears, (1998) showed that the ecotone boundary between ponderosa forests and pinyon-juniper woodlands shifted by 2 km in less than 5 years in a broad region of New Mexico during 1950s, and the ponderosa trees have never recovered. A regional drought that caused widespread mortality of ponderosa appears to have been the proximal trigger, however, other factors were in play: ; in particular, the long-term persistence of the vegetation shift change appears attributable to decadal/centennial shifts in climate, and abetted by prior fire-suppression that allowed pinyon-juniper to establish in the understory (Allen and Breshears 1998).

Regional drought recurred in 2000-2003, resulting in widespread mortality, this time of pinyon pine, possibly due to accompanying extremely high temperatures (Breshears et al., 2005). The ecosystem now appears to be shifting toward a juniper-grassland —a dramatic degradation from the ponderosa forests just 50 years earlier. These recent rapid changes in ecosystem structure and composition are super-imposed on a backdrop of large-scale shifts in the regional distributions of ponderosa and pinyon-juniper throughout the Holocene in response to the long-term changes in climate and atmospheric CO2 levels.

In this study, we propose to test the hypothesis that the forest and woodland ecosystems of the Southwestern US are vulnerable to rapid non-linear change in ecosystem composition, structure and function in response to predicted projected changes in climate and climatic variability, in order to develop prediction capability for future vegetation of the Southwestern US. Our principal tool will be the Ecosystem Demography Model Version 2 (ED2). We will use ED2 to integrate data for past and present climate and vegetation with information on the ecological and eco-physiological characteristics of these ecosystems. The proposed work will enable us to determine the causes of past changes in vegetation distribution, structure and function and thus better predict how future changes in temperature, precipitation and atmospheric CO2, and fire suppression will affect Southwestern ecosystems.


Ecology, Biogeography, and Physiology of Ponderosa and Pinyon-Juniper Ecosystems:

Ponderosa pine (Pinus ponderosa) occurs from southern Canada to Mexico, extending from the Pacific coast eastward as far as Nebraska (Figure 1). Moisture commonly limits ponderosa growth throughout its range (Burns and Honkala 1990) and its seedlings are readily killed by fire, but larger individuals are fire-resistant. More than 100 insect species attack ponderosa pine, with the western pine beetle (Dendroctonus bevicomis) the most common cause of mortality of mature trees (Burns and Honkala 1990). Pinyon pine (Pinus edulis) is co-dominant with juniper (Juniperus spp.) in pinyon-juniper woodlands that span the semi-desert zone, covering approximately 42.7 million ha from Texas to California (Burns and Honkala 1990). Areas that shift from ponderosa to pinyon-juniper often show increased soil erosion (Allen and Breshears 1998, Davenport et al. 1998). In both pinyon and ponderosa, recruitment pulses tend to occur during the first sustained wet period following a drought (e.g., Betancourt et al. 1996).

The Southwestern U.S. is characterized by highly seasonal precipitation and climate, with an arid fall, a variable winter/early spring, an arid late spring and fore-summer, and a relatively rainy summer (Swetnam and Betancourt 1998). As a result, drought is a key limiting factor for Southwestern forests (Meko et al. 1995). Climate scenarios for the region predict higher temperatures and increased precipitation variability (see the Climate Projections section below), suggesting increased frequency and intensity of drought events such as the observed in the 1950s and 2000-2003. Predicted increases in air temperature are likely to raise leaf temperatures and thus vapor pressure deficits, increasing water stress even in the absence of drought. The evapotranspiration (ET) rates of trees in the Southwest are poorly quantified, but ET seems to be dominated by the evaporation term, and to be sensitive to individual rain events (Kurc and Small 2004). During the summer monsoon season observed ET rates ranges from 0.5 – 4.0 mm day-1, with a seasonal average of 2 mm day-1 (Kurc and Small 2004, see also Lane and Barnes 1987).

Future changes in either temperature or precipitation could further changealter species distributions, not just at ecotone boundaries, but also across major portions of species’ regional distributionsthe region (Breshears et al. 2005). Relevant physiological attributes of ponderosa and pinyon-juniper are reasonably well known. For instance, pPinyon-juniper woodland species have higher radial growth and water use efficiency (WUE) during droughts than ponderosa (Adams and Kolb, 2004). However, the, but net photosynthesis of pinyon drops off very rapidly with small (~2oC) increases in temperature (Barnes & Cunningham 1987), while ponderosa respond to higher temperatures by increasinges their root:shoot ratios (Maherali & DeLucia. 2000) in warmer conditions. In response to rising CO2,With regard to physiological responses to CO2, pPonderosa exhibits an acclimating growth response (Johnson et al 1997) and nowithout significant changes in stomatal conductance (Maherali & DeLucia. 2000), in response to rising CO2 levels, while both pinyon and juniper appear to decrease their stomatal conductance in response to rising CO2 (Edgar and Koch 2000).

Historical reconstructions of plant communities from materials found in packrat middens in the Southwest indicate that the rising aridity and increasing CO2 concentrations during the last de-glaciation resulted in desert shrubs replacing woodlands across much of the Southwest (Van de Water 1994; Betancourt 1990), consistent with this behavior.

The goal of our proposed work is to determine the environmental causes of rapid, non-linear ecological changes in composition, structure and function of woodland and forests ecosystems of the Southwestern US. If the predicted changes in climate and climatic variability occur, the observed replacement of ponderosa by pinyon-juniper, and then subsequent conversion to juniper-grasslands, may continue to progress, implying radical changes in Southwestern ecosystems.

Rapid Shifts in Ecosystem Structure and Function Induced by Climate Variability:

The Data from the two droughts described in the introduction imply that semi-arid ecosystems in the Southwest are prone to rapid and extensive shifts in vegetation in response to temperature and moisture anomalies. Large-scale mortality and permanent replacement of existing vegetation appears to have occurred when climatic climate stress exceeded a threshold: during the severe regional drought of the 1950s (Figure 2), the ecotone boundary between ponderosa pine forests and pinyon-juniper woodland shifted by up to 2 km in just 5 years (Figure 3). The transition was documented by aerial photographs taken between 1935 and 1975, supplemented by measurements of living and dead ponderosa and of stem diameter increments at plots along the elevation gradient (Allen and Breshears, 1998). Additional measurements of stem diameter increment at plots along the elevation gradient (Allen and Breshears, 1998) highlighted how this shift was fundamentally caused by moisture stress, but was exacerbated by a stress-induced outbreak of bark beetles that amplified the drought-caused mortality.

The extensive die-off of the dominant ponderosa pine released pinyon and juniper that had become established in the sub-canopy as a result of prior fire suppression (see Figure 4), and t; the capability of pinyon and juniper to utilize near-surface water then exacerbated water stress on ponderosa that also and helped drive the vegetation transition. The vegetation change has persisted over the following 40 ~50 years since the drought.