Grantee Research Project Results
Final Report: Use of Multi-Scale Biophysical Models for Ecological Assessment Applications in the Southeastern United States
EPA Grant Number: R825157Title: Use of Multi-Scale Biophysical Models for Ecological Assessment Applications in the Southeastern United States
Investigators: Huston, Michael A.
Institution: University of Tennessee
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1996 through September 30, 1999
Project Amount: $1,697,104
RFA: Ecological Assessment (1996) RFA Text | Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Aquatic Ecosystems
Objective:
The primary goal of this project was to assess the relative magnitude and causes of natural variability in selected ecosystem properties that often are used as indicators of human impacts on the environment, specifically changes in productivity and/or biodiversity. The complementary conceptual frameworks that were evaluated for their utility in predicting indicator responses and variability were: (1) the dynamic equilibrium model of disturbances interacting with variation in the rates of population and community processes; and (2) responses of productivity to multiple interacting resources operating according to Liebig's "Law of the Minimum." Environmental properties that were addressed include aquatic and terrestrial biological diversity (species and landscape scales), tree growth and forest productivity, soil nitrogen and carbon content, animal population dynamics, and streamflow variability. To achieve these objectives, we focused on analysis of existing spatial and time-series datasets to support integration of data collected at multiple scales, as well as integration of climatic, hydrologic, and ecological processes.Summary/Accomplishments (Outputs/Outcomes):
Research and analyses completed during the project, plus review and synthesis of published results, have supported the primary hypotheses on which the project was based and confirmed the existence of consistent patterns of ecological response over a range of spatial scales and types of systems. The major conclusion of the project is that much of the apparent high variability and unpredictability of ecological dynamics is the result of failure to appropriately stratify environments for analysis and failure to appropriately match the scales and variables of analysis with the known or hypothesized causal mechanisms.
Patterns and Dynamics of Species Diversity. Species diversity is widely used as an indicator of environmental quality, yet the factors controlling diversity have been the subject of much debate in the ecological literature. Even though mortality-causing disturbances have been known to influence species diversity, recent reviews have concluded that the effect of grazing (a type of disturbance) on diversity in aquatic systems is completely unpredictable.
The dynamic equilibrium model (Huston 1979, 1994), which served as the conceptual framework for this project, makes specific predictions about the patterns and variance of ecological properties. Because the predictions of this hypothesis are derived from the interaction of two independent sets of processes (mortality-causing disturbances and population dynamics, as they affect survival and competitive interactions), the predicted patterns can be detected only when the environmental conditions under which the processes are occurring are appropriately characterized and stratified.
One of the major predictions of the dynamic equilibrium model that is directly relevant to the objectives of this proposal has been validated recently. The central prediction of the model is that the effects of changing disturbance frequency on species diversity should REVERSE between productive and unproductive environments. Proulx and Mazumder (1998) conducted a meta-analysis of published studies on the effects of grazing and mowing on plant diversity in both aquatic and terrestrial systems. They determined that all of the studies conducted in unproductive, oligotrophic environments found that diversity DECREASED with increasing intensity of grazing, while in productive, eutrophic environments, most studies found that diversity INCREASED with increasing intensity of grazing. The dynamic equilibrium model also predicts an analogous response reversal of diversity along productivity (e.g., nutrient) gradients in high-disturbance versus low-disturbance environments.
These response reversals have profound implications for the interpretation of diversity changes as an environmental indicator. Specifically, the potentially opposing diversity responses along major environmental gradients can produce high variance in the observed response along the gradient (primary x-axis), with a weak or no correlation between diversity and the x-axis variable. Only when the second interacting factor (either productivity or disturbance) is used to stratify the observed responses to the first factor (either disturbance or productivity) can the pattern be observed and predictable responses quantified.
Several aspects of these complex diversity dynamics were addressed by graduate student projects.
Analysis of long-term datasets of fish and benthic insect communities in Bear Creek on the Oak Ridge National Environmental Research Park indicates strong physical regulation of both population and community structure, consistent with the predictions of the dynamic equilibrium model for disturbance-driven systems. Results show significant patterns that were related to: (1) differences between the diversity dynamics of certain functional groups of organisms and that of the entire community; (2) strong nestedness of community structure, with the highest number of species at the downstream sites where disturbances from streambed drying were lowest; and (3) reversal of population and community responses to flow variation between upstream and downstream sites. Strong flow-related interannual variation in species distributions and diversity suggests caution in the use of diversity indices as indicators of anthropogenic impacts in first- through third-order streams, at least in this region.
Ice storms are major landscape-scale disturbances that occur periodically throughout the eastern United States. Dr. Charles Lafon completed a thesis that investigated the interaction of this major (but little studied) type of forest disturbance with topographic variability in forest productivity and successional dynamics. He made field measurements at several locations within the Appalachian region to quantify the effects of ice storms on forest structure and found consistent effects with respect to elevation, topography, and species properties. He then developed a modified version of the FORET/LINKAGES individual-based forest simulation model to predict long-term dynamics of forest diversity in relation to ice storm disturbance and site productivity, as both varied with elevation and topography. This model represents an implementation of the dynamic equilibrium hypothesis as it may operate in forest systems.
Another study completed by Dr. Lafon demonstrated the effects of past human land use on forest succession and the recovery of forest diversity following agricultural disturbance. Forest succession was dramatically slower on the most heavily eroded sites, with stands remaining dominated by early successional species after more than 50 years' post-abandonment. In contrast, tree diversity was much higher, and succession had advanced to the hardwood stage (primarily oaks) on the least eroded sites (Lafon, et al., 2000).
A second major prediction of the dynamic equilibrium model related to diversity and human impacts addresses the issue of invasive species. The hypothesis predicts that environments with high diversity should be most easily invaded by exotic species. This contradicts the prediction of classical competition theory that high-diversity communities have finer resource partitioning and more complete use of resources, and thus should be more resistant to invasion. Some recent experimental results seem to support this prediction.
While the results of these short-term experimental manipulations seem to support the prediction that more diverse plant communities are more resistant to invasion, field studies of patterns of invasive species in relation to native species diversity have found the opposite pattern. A survey of parks and preserves around the world found a positive correlation between the number of native species and the number of exotic species (Lonsdale, 1999). A higher resolution study of vegetation in the Rocky Mountains and Great Plains of North America also found that native and exotic species richness were positively correlated across sample areas at several levels of sampling resolution (Stohlgren, et al., 1998). All of the field data published to date contradict the primary theory of invasibility and diversity falsifies the classical theory, but they are consistent with the predictions of the dynamic equilibrium model.
A graduate student at the University of Tennessee, Patrice Cole, has begun to examine the spatial distribution of a major invasive grass species, Microstegium vimineum, in the context of the dynamic equilibrium model.
Additional refinements of our understanding of diversity patterns were developed in an evaluation of issues related to the ongoing controversy about the control of diversity by "local" versus "regional" processes (Huston, 1999). Several prominent papers have argued that diversity patterns primarily are controlled by regional processes related to large-scale extinction and speciation patterns. The primary evidence for regional control has been the demonstration that local diversity is linearly correlated with regional diversity, rather than leveling off at some asymptote that would reflect the effect of local processes that cap diversity at some maximum level. However, review of the supporting evidence revealed that the spatial scale at which the "local" diversity was assessed was far too large to show the effect of local processes, and was in reality dominated by the effects of environmental heterogeneity that completely obscured the results of local processes operating at a smaller scale. Evaluation of studies at an appropriate scale revealed that, under conditions consistent with the predictions of the dynamic equilibrium model, local diversity reached an asymptote in relation to regional diversity, demonstrating the effect of local processes in setting maximum levels of diversity.
This work also provides a simple explanation for the criticism that diversity responses along environmental gradients are not consistent across multiple studies. The effect of regional diversity (number of species that could potentially reach a specific location) should increase diversity most strongly under conditions in which local diversity-reducing processes operate weakly (the conditions where the dynamic equilibrium model predicts diversity to be highest). In contrast, under conditions where the model predicts that diversity should be low due to local processes related to competition and survival, regional diversity should have little effect on local diversity. Thus, maximum local diversity will be much higher in areas with high regional diversity, resulting in a conspicuous "peak" in diversity, while there will be a much lower peak, and perhaps no conspicuous peak at all, where the regional diversity is low. This mechanism explains much of the inconsistency in diversity responses along gradients of productivity found in published reviews, which has led to the erroneous conclusion that diversity cannot be predicted on the basis of environmental conditions.
Thus, the results of our own research and analyses, as well as recent publications in the open literature, indicate that the dynamic equilibrium model provides a powerful framework for predicting the spatiotemporal patterns of community and ecosystem properties in both aquatic and terrestrial systems. While we are still far from being able to make precise a priori predictions of the diversity that will be found in unstudied environments, we now have a much better understanding of the factors that influence the complex patterns of diversity, and some hope that we can develop better predictive capabilities.
Spatial and Temporal Dynamics of Growth and Primary Productivity. A second major objective of the project was to develop a predictive understanding of the spatiotemporal variability of primary productivity at the ecosystem level, as well as the growth of individual organisms and populations that both produce and respond to primary production. Production and growth rates often are used as direct indicators of environmental conditions, and also are critical driving variables for patterns of species diversity, according to the dynamic equilibrium model. Use of the dynamic equilibrium model for predicting ecological dynamics on real landscape requires information on production and growth rates as input.
Our work in this area focused on understanding and predicting the spatiotemporal distribution of environmental conditions that influence growth and productivity over landscapes, and understanding and predicting the effects of multiple interacting factors on the growth of organisms.
We used a combination of field measurements of environmental driving variables (soil carbon and nitrogen, precipitation/soil moisture/streamflow) and organismal responses (tree growth of individual trees and population dynamics of stream invertebrates) to describe the patterns and to evaluate potentially applicable models.
At our smallest and most intensively sampled research site, the Oak Ridge Reservation, we are completing the final analysis of the long-term forest inventory dataset from Walker Branch Watershed, where variation in forest productivity and structure has been strongly affected by the El Nino?Southern Oscillation (ENSO) precipitation cycles over the past 30 years. The effects of decadal or longer climatic cycles are evident in the patterns of tree growth and mortality. The high rates of forest growth during the 1960s and 1970s were associated with a series of wet years, which clearly can be seen in the regional Palmer Drought Severity Index record. Two severe droughts in the 1980s led to a major episode of tree mortality, and the average basal area of the inventory plots declined by nearly 20 percent. Since the end of the droughts, forest growth has been much slower than prior to the drought, with the basal area increment 80 percent lower than its pre-drought value and increasing gradually. Only additional data will reveal whether the forest will recover to its previous levels of growth.
The tree-ring record for tulip poplar on Walker Branch, collected by graduate student Gregory Barlar, clarifies the longer term interaction of climate and tree growth. The low growth rates of tulip poplar, one of the most drought-sensitive species in the forest, are clearly shown by the dramatic decline in ring width following the droughts of the 1980s, and are similar to the growth rates the trees experienced from the 1920s through the 1950s, when the entire country was experiencing major droughts. The wet decades of the 1960s and 1970s, clearly seen in the Palmer Drought Severity Index record, stand out as a period of unusually rapid growth for tulip poplar.
Our analysis of spatiotemporal variation in tree growth, using both long-term forest inventory data and dendrochronological analysis, demonstrates why long-term spatially-spatially distributed data are essential for understanding and interpreting variability in ecological processes, which is the primary theme of this project. Not only does the "typical" growth rate of trees change from one climatic period to another (e.g., 1920-1960 vs. 1960-1980), but the "optimal" conditions for a species shift from one position on the landscape to another. The tree-ring analysis clearly shows that during wet periods the tulip poplars on Walker Branch Watershed grew most rapidly on ridgetop sites, while during dry periods the tulip poplars in the valley bottoms were growing fastest. The spatiotemporal dynamics of ecological processes reveal complexities that must be understood before environmental changes and anthropogenic impacts can be detected and interpreted. Similar phenomena were found in our analysis of aquatic data mentioned earlier.
The other focus of our productivity-related work was the effect of variation in multiple interacting resources on growth and productivity. We used analysis of published data, in combination with modeling and statistical analysis, to evaluate the effects of interacting factors. Our primary conclusion is that in many ecological situations, interacting factors operate in a manner that prevents detection of valid mechanistic relationships, and that produces a pattern of response variation that violates the assumptions of standard statistical analyses. This conclusion is consistent with a growing awareness that standard statistical methods are inappropriate for many ecological datasets, and a movement to develop alternative statistical approaches that are relevant.
The basic problem is that most statistical methods are based on the assumption that the mean response represents the true relationship between the dependent and independent variables. While this is true for situations where variance is caused by random processes that cause positive and negative deviations from the actual response, it is not true when multiple LIMITING factors are operating according to Liebig's "Law of the Minimum." In this situation, the deviation from the true response of a dependent variable to a single driving variable can only be NEGATIVE. In such cases, variance is not randomly or normally distributed, but rather distributed as values between the actual response (observed when no factors other than the focal independent variable are limiting) and zero (observed when one or more factors, including the independent variable, are completely limiting).
In virtually all ecological phenomena, from individual growth to the diversity of a community, it is likely that multiple limiting factors are potentially operating. Unless all factors are known and quantified, the estimated response inevitably will be smaller than the actual response (e.g., a regression with a slope lower than the slope of the actual relationship). Models based on such relationships inevitably will fail to produce accurate predictions, even if they are applied to situations where no other factors are limiting. Approaches for addressing this problem include "quantile regression" (Cade, et al., 1999).
Another outgrowth of this work is a new hypothesis about the conditions under which natural variability in ecosystem processes can be increased (i.e., destabilized) or decreased (i.e., stabilized) by changes in environmental conditions that are not the direct cause of the variability (Huston, 2001). Our work on multiple resource interactions includes both simulation modeling and analysis of published data, and has led to the identification of a potential mechanism for alteration of natural patterns of spatiotemporal variability?the "variance amplification hypothesis"?that is based on a generalized version of Liebig's "Law of the Minimum" applied to any process that is potentially limited by multiple factors. Considerable empirical evidence supports the prediction of this hypothesis that variance (as well as normalized variance, i.e., the coefficient of variation) should increase across a gradient of increasing levels of any one out of the set of multiple limiting factors. This hypothesis has major implications for the analysis and interpretation of responses to changing environmental conditions, and implies that the standard statistical methods used to quantify ecological patterns are inappropriate for many ecological situations because of the assumptions they require.
The complexity of environmental impacts on ecological processes is an issue in experimental ecology, as well as in environmental assessment and impact analysis. Work conducted under this project also has addressed the interpretation of recently published ecological experiments on the effect of biodiversity on ecosystem processes, particularly productivity. A number of serious problems were found in the design and interpretation of these experiments (e.g., Huston, et al., 2000), supporting the conclusion that the effect of species diversity on primary productivity is relatively small compared to the effect of primary productivity on species diversity. The issue of determining causality and identifying limiting factors or methods that may produce spurious correlations is just as critical for environmental monitoring and assessment as it is for experimental ecology.
Primary Results. The results of this project, in conjunction with ongoing review of recent publications related to the project, lead to the following conclusions:
- Complex patterns of species diversity on heterogeneous landscapes can be predicted using the interaction of productivity and disturbance as described in the dynamic equilibrium model. The most important aspect of this interaction is the REVERSAL of the effect of disturbance (or productivity) on species diversity between high and low levels of productivity (or disturbance). Understanding this interaction is essential for assessing the causes of observed changes in diversity, as well as for developing resource management plans and predicting the effects of specific management activities on diversity.
- Different types of organisms (e.g., functional feeding groups, life forms, functional types, trophic levels, etc.) have different patterns of diversity along gradients of productivity or disturbance. These complex patterns are consistent with the predictions of the dynamic equilibrium model. Understanding this relationship is essential for assessing the causes of observed changes in the relative abundance of species and the structure of multi-species assemblages, and for developing plans for management and conservation. The inevitable consequence of this pattern is that no single management action, or conservation strategy, can maximize the diversity of all types of organisms.
- Accurate prediction of the distribution of natural (as well as anthropogenic) disturbances and environmental conditions that affect productivity (e.g., climate, soil nutrients) is essential for prediction of the spatiotemporal responses of organisms, as well as of the patterns of species diversity. Simulation models of the physical environment, and of organismal responses to the environment, can be powerful tools for assessment and management, but require continued improvement to meet their potential.
References:
Cade BA, Terrell JW, Schoeder RL. Estimating effects of limiting factors with regression quantiles. Ecology 1999;80(1):311-323.
Huston MA. A general hypothesis of species diversity. American Naturalist 1979;113(1):81-101.
Huston MA. Biological Diversity: The Coexistence of Species on Changing Landscapes. Cambridge University Press, 1994, 708 pp.
Lonsdale WM. Global patterns of plant invasions and the concept of invasibility. Ecology 1999;80(5):1522-1536.
Proulx M, Mazumder A. Reversal of grazing impact on plant species richness in nutrient-poor versus nutrient-rich ecosystems. Ecology 1998;79(8):2581-2592.
Stohlgren TJ, Binkley DA, Chong GW, Kalkhan MA, Schell LD, Bull KA, Otsuki Y, Newman G, Bashkin M, Son Y. Exotic plant species invade hot spots of native plant diversity. Ecological Monographs 1999;69(1):25-46.
Journal Articles on this Report : 12 Displayed | Download in RIS Format
Other project views: | All 64 publications | 16 publications in selected types | All 13 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Harden CP, Scruggs PD. Infiltration on mountain slopes: a comparison of three environments. Geomorphology 2003;55(1-4):5-24. |
R825157 (Final) |
Exit Exit |
|
Hastwell GT, Huston MA. On disturbance and diversity: a reply to Mackey and Currie. Oikos 2001;92(2):367-371. |
R825157 (Final) |
not available |
|
Huston MA. Local processes and regional patterns: appropriate scales for understanding variation in the diversity of plants and animals. Oikos 1999;86(3):393-401. |
R825157 (1999) R825157 (Final) |
Exit |
|
Huston MA. Microcosm experiments have limited relevance for community and ecosystem ecology: synthesis of comments. Ecology 1999;80(3):1088-1089. |
R825157 (1999) R825157 (Final) |
not available |
|
Huston MA, Aarssen LW, Austin MP, Cade BS, Fridley JD, Garnier E, Grime JP, Hodgson J, Lauenroth WK, Thompson K, Vandermeer JH, Wardle DA. No consistent effect of plant diversity on productivity. Science 2000;289(5483):1255a. |
R825157 (1998) R825157 (Final) |
Exit Exit |
|
Huston MA, Marland G. Carbon management and biodiversity. Journal of Environmental Management 2003;67(1):77-86. |
R825157 (Final) |
Exit Exit Exit |
|
Huston MA. Management strategies for plant invasions: manipulating productivity, disturbance, and competition. Diversity and Distributions 2004;10(3):167-178. |
R825157 (Final) R828897 (2002) |
Exit |
|
Huston MA. The three phases of land-use change: implications for biodiversity. Ecological Applications 2005;15(6):1864-1878. |
R825157 (Final) |
Exit |
|
Lafon CW, Graybeal DY, Orvis KH. Patterns of ice accumulation and forest disturbance during two ice storms in southwestern Virginia. Physical Geography 1999;20(2):97-115. |
R825157 (1998) R825157 (1999) R825157 (Final) |
not available |
|
Lafon CW, Huston MA, Horn SP. Effects of agricultural soil loss on forest succession rates and tree diversity in east Tennessee. Oikos 2000;90(3):431-441. |
R825157 (1998) R825157 (1999) R825157 (Final) |
Exit |
|
Lafon CW, Speer JH. Using dendrochronology to identify major ice storm events in oak forests of southwestern Virginia. Climate Research 2002;20(1):41-54. |
R825157 (1998) R825157 (1999) R825157 (Final) |
Exit |
|
Scanlon TM, Raffensperger JP, Hornberger GM, Clapp RB. Shallow subsurface storm flow in a forested headwater catchment: observations and modeling using a modified TOPMODEL. Water Resources Research 2000;36(9):2575-2586. |
R825157 (Final) |
Exit |
Supplemental Keywords:
landscape, dynamic equilibrium model, disturbance, diversity, exotic species, wildlife models, nonequilibrium, environmental gradients, biodiversity, land use impacts, simulation models., RFA, Air, Geographic Area, Ecosystem Protection/Environmental Exposure & Risk, exploratory research environmental biology, Ecosystem/Assessment/Indicators, Chemical Mixtures - Environmental Exposure & Risk, Ecosystem Protection, climate change, Monitoring/Modeling, Ecological Effects - Environmental Exposure & Risk, Ecological Effects - Human Health, Southeast, Ecological Indicators, multi-scale biophysical models, geomorphology, environmental monitoring, aquatic biota , assessment models, climate change impact, ecosystem assessment, multi-level indicators, biodiversity, microclimatic conditions, hydrological, ecological assessment, aquatic ecosystems, soil carbon, landscape characterization, spatial and temporal patterns, climate variability, ecological researchRelevant Websites:
http://www.ornl.gov/ornlreview/rev29_3/text/life.htmProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.