Urbanization as a global process began with the industrial revolution. Starting in 1850, Europe and North America embarked upon an exponential increase in urban population, followed by the developing world around 1950. In 2000, the world achieved not only a total population of 6 billion, but surpassed the 50% urban mark. As we move into a new century, urbanization rates in developing mega-cities far exceed any type of classical European or North American urban spread. For example, the Brazilian Amazon is one of the most urbanized societies, with over 70% of the population in urban areas, while cities in Italy and some other parts of Europe and the Eastern United States are losing population. Meanwhile, urbanization in the developed world has seen massive decreases in living densities, telecommuting from rural zones, and a rise in the number of single person households. Three factors are evident: first, that existing models of the urban transition are failing as a new type of urban transition occurs; second, that human habitation has become immensely more consumptive of geographic space; and third, that the conversion of land to urban use has become the primary engine of global environmental change. We face the ironic situation that urbanization is happening where we least understand it, in ways previously unknown, and it is continuing as a land transition even in areas where populations are decreasing.

This presentation will review the methods and results of the USGS's Urban Dynamics program. In this work, historical data on urbanization has been compiled with the goal of training a model that can also simulate future growth. This model, called SLEUTH, predicts both urbanization and the impact that urbanization has on other land transitions. In recent work, funded under NSF's Urban Research Initiative, we have developed systems of coupled models that allow urbanization's impact to be traced as it impacts the build and physical environments. Our work has examined the role of wildfire and hydrology, but others have also examined run-off's impact on marine systems, the selection of suitable sites for landfills, and the impact of land use change on local microclimates. Taken together, these impacts have had a negligible global impact in the past, but now may represent the single most important driver of global change, and so cannot be ignored in global level planning for biodiversity, global warming, sea level change, and sustainable land use.
 
 

Concern over declines of amphibian populations has increased considerably in the past decade. Declines have been documented globally and are increasing in severity, and climate change has been hypothesized to be one of the causes. There are correlations between El Niño-related drought and amphibian extinctions in Costa Rica, but other evidence linking amphibian declines to climate change has not been found. However, life history of amphibians, particularly species occurring in temperate climates, is closely tied to weather, and changes in breeding phenology have been recorded for some species. Future climate change may be beneficial in some cases, but is most likely to cause problems for amphibians, exacerbating declines or population fragmentation from other causes. The typically large amounts of variation observed in amphibian life history and population dynamics will reduce the precision of models of the effects of global change. Data from 15 yrs of amphibian monitoring in northern Colorado will be used to illustrate that responses to climate change are dependent on local conditions.
 
 

Trees shifted latitude or elevation in response to Quaternary climate change. Because many modern trees display adaptive differentiation in relation to latitude or elevation, it is likely that ancient trees were also so differentiated, with environmental responses of populations throughout the range evolving in conjunction with migrations. Very rapid climate changes challenge this process by imposing stronger selection and by distancing populations from environments to which they are adapted. Extinctions may thus be more probable at times of rapid change than at times of severe climate, suggesting that future rapid change may threaten the persistence of many species.
 
 

Global climate change could have profound effects on the earth's biota, including large redistributions of tree species and forest types. We used a deterministic regression tree analysis model (DISTRIB) to examine environmental drivers related to current forest species importance values, then examine potential future suitable habitat using five global circulation model (GISS, GFDL, UKMO, Hadley, Canadian Climate Centre) outputs of estimated 2xCO2 climate regimes. We evaluated potential shifts for 80 common tree species in the eastern United States, using forest inventory data and 33 environmental variables. Results showed a wide variation in potential future species suitable habitat, depending on the GCM used. The UKMO model produced the greatest potential shifts, while the Hadley model produced the least. Several species could potentially be extirpated from the United States, and the dominant species and forest types could shift considerably. Given these potential future suitable habitats, actual species redistributions will be controlled by migration rates attainable through fragmented landscapes. We are using a cellular automata model, SHIFT, which migrates species through landscapes based on habitat quality and species abundance near the range boundary, to model possible future distributions over the next 100 years. Preliminary results show very slow migration when forests are scarce and species abundance is low near the current range boundary.
 
 

U.S. natural landscapes and habitats will experience four types of stress during the next hundred years: (i) conversion and spillover pressures from urban growth; (ii) changes in resource demands such as farming and forestry; (iii) continued pressures from invasive species; and, (iv) pressures from global climate change, particularly changes in temperature and precipitation regimes. This presentation will deal with the first of these stressors, urban growth. Drawing on the results of a statistically calibrated and spatially explicit model of urban growth, it will present the results of three simulations of urban land-use change among California urban counties through 2020. The three scenarios, entitled baseline, compact growth, and sprawl, start from a common set of population projections but differ in terms of infill proportions (the share of urban growth accommodated in existing urban areas) and development densities. The results are presented in table and map form and then compared with resource and habitat inventories to determine the potential impacts of urban growth on farmland and on threatened and endangered species. The resulting changes form the backdrop for a more detailed investigation of the likely effects of climate change. The modeling methodology makes use of widely available digital data as well as categorical data modeling and could easily be duplicated elsewhere in the United States.
 
 

During the last decade of the twentieth century, the scientific community began focusing on the interactions between the biosphere and the atmosphere and hydrosphere as a mechanism for affecting environmental systems. In addition, there was increasing recognition that perhaps one of the key drivers of environmental change at all scales was local land use practices and dynamics. Real progress in modeling and assessment, however, was hindered by the lack of current and accurate land use and land cover data at appropriate scales. As a result, national and international scientific groups organized teams to produce the needed national and global land use and land cover databases. As a result of these initiatives, the past 10 years has been an incredible period of discovery in which unprecedented effort has been focused on mapping and characterizing land use and land cover at local, national, and global levels. For example, a detailed global land cover database produced under the auspices of the International Geosphere Biosphere Program shows the full extent of agriculture across the globe - a land use that now covers almost one-third of the planet. At the national level, land cover with high spatial resolution provides a consistent picture of the patterns of settlement and land uses that affect environmental quality and economic activity. These studies provide baseline data that are now being used to parameterize models addressing climate, biogeochemistry, biological diversity, water quality, and many other phenomena. As we step into the twenty-first century, the focus is expanding to studies of land use and land cover dynamics. Of particular importance is the development of an improved understanding of the spatial and temporal dimensions of local to regional rates and driving forces of land use and land cover change. Systematic studies of land use dynamics spanning local to regional levels are essential if we are to understand how to scale and connect observed changes with regional and global environmental consequences.
 
 

Global change is a natural reality over the long time scales of Earth. The planet has never been static in its (by current best estimates) 12 Gy history, and there is no reason to expect that it should become so now just to accommodate the human presence. The geologic record is one of continual change punctuated by major events, some biogenic. Perpetual change is an ongoing reality, and human continuance depends on successful entrainment to the global dynamic rather than reactionary prescriptions for stasis. Human adaptation, however, must begin with a correct understanding of the systemic relationships between planet and mankind.

The Earth and its systems are not fragile, as often stated. On human scales the planet is relatively permanent, stable, and well behaved. Factors in this include a large, hard, durable mass, a stable orbit around an even larger mass, the existence of many feedback checks & balances in systemic organization, and the fact that change is lawful and, if only grudgingly predictable, at least understandable.

The Earth is unitary. Its core, mantle, and crust operate not as separate entities but as a unified whole. Factors in this are rotational dynamics, crustal weathering, crust-mantle mixing, and core differentiation & evolution. Weathering brings massive exogenous energy to crustal change. It contributes to chemical differentiation and translation of crustal rocks, plate tectonics and continental drift, subduction of the oceanic crust (which is little more that 80 My old as a result), and massive shifts in the mantle as recently revealed by seismic tomography.

The Biosphere is a superficial crustal phenomenon, established only during the last third of planetary existence (4.6 Gy). The energies involved in biogenic phenomena are small compared to those powering the grand processes of the globe. Life itself is a globally interconnected network whose continuance occurs at the interface between speciation and natural selection. Most species that ever evolved are extinct, implying tight coupling of biotic change to global change as a property of the unitary planet. Some biological milestones and their consequences in global evolution are as follows: first life forms (4.6 Gy); first anaerobic bacteria (4.2 Gy); first stromatolites (3.8 Gy); anaerobic procaryotes and oceanic plankton diversify (2.5 Gy); aerobic photosynthesis, rapid increase in atmospheric oxygen (2.4 Gy); multicellular eucaryotes (2.1 Gy); aerobic respiration (2.0 Gy); eucaryote radiation, high atmospheric oxygen (1.9 Gy); aerobic procaryotes diversify (1.7 Gy); present oxygenated atmosphere achieved (0.6 Gy); metazoa abundant (0.5 Gy); first terrestrial plants (0.4 Gy); snothites (lithotrophic H2S-metabolizing microorganism communities) (0.2 Gy); dinosaurs and angiosperms (0.1 Gy); genus Homo (2.0 My); Homo sapiens (0.5 My); Homo holisticus (1.1 Cy).

Homo holisticus, Systems Man, arose during the Industrial Revolution and radiated in rapid development throughout the 20th Century. First of a kind in organic evolution, the selective forces acting on this new form of life are global, not local as they have been for all prior species. Entrainment of a global species to global change poses special survival problems because loss of locality entails loss of refugia also, among other things. The internet, mass transportation, and economic globalization are early factors locking humanity into irreversible and virtually instantaneous interdependence of all its parts on a nonlocally determined global whole. The rise of a biospheric "superorganism" is a not unrealistic prospect, paralleling at other scales the symbiotic unions of free-living microbiota that earlier traded relative independence for the constrained interdependence of the metazoan habit.

Anthropogenic contributions to global change are probably widespread but hard to pin down by the small energies, complex causalities, and long time scales involved. In China from 1950 to 1990 human population doubled from 540 million to 1140 million, and the human-built environment expanded from 10.9 million to 32.8 million hectares. This had a number of ripple effects over five decades (data from T. X. Yue, Chinese Academy of Sciences): reduction in diversity of Holdridge life zones (26 biomes; reduction rate= 0.8497 + t^-0.00853, t = time), increase in flood and dust storm calamities (rate of increase = 17.67 + t^0.7625), reduction in ecological buffer capacity (rate = -1.00034 + t^-0.00057), increased economic losses due to floods (rate = 27.0794t + 25.4206) and seldom seen sand + dust weather, with consequences for both regional and global climate change.

To adapt to global change, human change must accommodate to certain principles of organization that apply to the syncytial network H. holisticus is presently establishing worldwide. Two of these network principles seem particularly germane:

1. Network aggradation. The second law of thermodynamics (entropy principle) contraindicates life, which is antientropic. Network aggradation is a principle of systems organization that allows order to be generated against 2nd-law energy degradation. It asserts that departure from equilibrium (aggradation) and implied growth of order (negentropy) may exceed the 2nd-law generation of disorder (entropy). The canonical sufficient condition for network aggradation is simple electromagnetic coupling (energy-matter exchange) between non-equilibrium entities. This will be illustrated with example networks. The implication is that by establishing unlimited interactions within a system, fixed boundary inputs can support indefinite internal aggradative development. H. holisticus would appear to be on this track at the present time through widescale technological increases in electromagnetic coupling.

2. Network synergism. This is the increase in aggradation/degradation ratios with increasing connectance. However, the rate of this increase decreases as the number of interconnected units rises. This will be shown by Monte Carlo simulations. It gives a cautionary prescription for adaptive evolution of a globalizing species in tracking global change: Humanity should self-organize into geographically distributed, discrete population aggregations that are strongly interconnected but weakly cross-connected. These would mimic the "near decomposable systems" of Herbert Simon's theory of hierarchical organization.

From cells to demes to cities and beyond, localizing spheres of influence seems prudent from the dual perspectives of network aggradation and network synergism. A global reach may contraindicate sustainable entrainment to global change by reducing network synergism even though network aggradation may rise.
 
 

The representation of hydrologic processes in Atmospheric General Circulation Models dates to the late 1960's with the classic bucket model of Manabe. Since that time, the development of land surface models (a.k.a. Land Surface Parameterizations -- LSPs; or Soil-Vegetation-

Atmosphere-Transfer Schemes -- SVATS) has focused on vertical detail such as the detailed canopy representation of Sellers' Simple Biosphere (SiB) Model and Dickinson's Biosphere Atmosphere Transfer Scheme (BATS). With increased recognition of the critical role of the hydrologic cycle in climate change as well as a desire to assess potential water resources impacts associated with climate change, the focus is shifting towards improving the representation of hydrologic processes in land surface models. This shift is partly motivated by Koster and Milly's contribution to the Project for the Intercomparison of Land surface Parameterization Schemes (PILPS), in which it was shown that long-term hydrologic budgets are as sensitive to runoff generation processes as to evapotranspiration processes.

As model spatial resolutions improve and more information is sought regarding the coupling of hydrologic and biogeochemical cycles, two critical categories of issues related to the representation of hydrologic processes in climate models can be identified. The first category is that of spatial representations, and includes issues related to the largely missing concept of a catchment -- the fundamental hydrologic unit, streamflow routing, the assumption of grid-scale spatial homogeneity or "effective parameters", the coupling of land surface and atmospheric models, and the special case of representing urban areas and urban hydrology. The second category relates to hydrologic process models and includes issues such as precipitation downscaling, runoff generation by the saturation excess mechanism, canopy conductance modeling including advances in root hydraulics, and grid-averaged fluxes of momentum (shear stress) and scalar quantities, as well as linkages between soil moisture, temperature, biogenic VOC and soil NOx emissions. These categories will be discussed along with some new approaches to the land-atmosphere modeling problem.
 
 

Human activities have radically transformed the landscape. Richards (1990) has estimated that over the last three centuries, 12 million km² of forests have been lost (19%), grasslands and pastures diminished by 5.6 million km² (8%, but many grasslands were converted to pastures), and croplands expanded by 12 million km² (466%).

With the advent of remote sensing, global land cover can now be monitored consistently from space. However, satellite data are only available for the last two decades, and are not useful for characterizing historical land use and land cover change. Furthermore, although remote sensing can characterize the land cover, it is often unable to identify the land use practices within a land cover, or to distinguish between natural landscapes and human dominated landscapes (e.g., a pasture from natural grassland).

By synthesizing remotely sensed global land-cover data with current and historical land use census data, we have recently developed a geographically explicit global data set of croplands from 1700 to 1992. We are currently extending this work to include the description of historical changes in grazing land and urban areas.

Furthermore, we are also building techniques to anticipate where future changes in land use and land cover are likely to be. By combining several climatic and soil indices, we have developed a data set describing the suitability of global land for cultivation. This map indicates that there is an additional 67% potentially cultivable land in the current climate, primarily in Tropical Africa and northern South America.

We have further incorporated the dynamics of land transformation between natural vegetation and croplands within the IBIS global dynamic ecosystem model and examined the influence of changing atmospheric CO₂ concentrations, climate, and croplands over 1860-1992 on the global carbon cycle. Terrestrial ecosystems have been a small source of CO2₂ to the atmosphere from 1860 to 1960 and been a sink since then. Cropland changes contributed to a net release of 109 Gt-C over the whole period.
 
 

Global climate change, brought about by human impacts on the earth's ecosystems, has the potential to massively alter the structure and functioning of natural systems on a scale and extent that is unprecedented in our planet's history. Attempts to foresee the ecological consequences of global change have primarily focused on the redistribution of species on a continental scale. These analyses have been completed at two levels: process-based modeling of animal species' ability to tolerate heat loads given anticipated future climate scenarios, and long-term monitoring of the geographic extent of species distributions in response to coincident climate warming. To offer some insight into emerging trends in species redistribution, I present a synthesis of efforts to anticipate and quantify the effects of global climate change on animal species populations. Species redistribution, however, is only one effect of climate change. A growing body of research indicates that there also may be shifts in the timing of life-history events, which in turn have important effects on population demography and ultimately community dynamics. This paper will, therefore, offer synthesis of the nature of such life-history shifts. It will also suggest how one might begin to integrate this empirical insight into process-based models of climate change that attempt to scale form the level of individual life-history to population and community dynamics.
 
 

This paper will illustrate the importance of evaluating climate change from a multiple scales. The approach is to use different ecological models of the dynamics of the boreal forest to illustrate changes in the world's boreal forest over the past 50 years. At least two sets of processes combine to determine ecosystem responses to disturbance or environmental change. Slower plant community processes, which depend on species physiology, morphology, and life-history characteristics, substantially complicate the consequences of first-level, rapid responses by basic photosynthetic and microbial processes. In forests, community dynamics, including competition for light, space, nutrients, water, and other resources may well be as important or even more important in determining ecosystem responses on time scales from several years to centuries than are the primary photosynthetic processes. In the case of boreal forests in the regions that have seen substantial warming over the past 50 years, models of forest responses at different levels point to multiple plausible and independent explanations of a purported increase in growing season and productivity of these regions. Each of these responses to climate change could be acting simultaneously.
 
 

The relationships between environmental management policies and supporting scientific knowledge range from strong to obscure or nonexistent. This entire range may apply in the case of managing atmospheric CO₂ concentrations. The available, relevant scientific information on the role of land use, land use change and forestry in the storage and release of terrestrial carbon was summarized in a special report by the Intergovernmental Panel on Climate Change (IPCC). The report was generated to permit the Conference of Parties (COP) policymakers to define the carbon credits countries would receive or lose for carbon sequestered or released by a-, re-, and deforestation during an initial 5-year commitment period, and by these and other land use activities in subsequent periods. The objective has not been reached, in part because the science available to policymakers is inadequate for the negotiation portion of the task. Consider how precisely we can measure carbon flux from individual small plots over large, contiguous regions. Think of the methods available to verify the amount of carbon sequestered by each country annually. Is the science strong enough to distinguish between carbon stored by forests because of recent afforestation projects or past land use history, and that stored by coincidental fertilization from increasing CO₂ and nitrogen deposition associated with use of fossil fuels? How might we implement land management policies that fit specific negotiated reductions in emissions and increases in terrestrial carbon sequestration? How could we accurately calculate impacts of the management policies on atmospheric CO₂ to gain acceptance by governments of negotiated reductions? These and similar questions have dogged the COP and will require careful consideration by the science community for some time to come.
 
 

Ecologists are being asked to make sweeping predictions about the future of Earth's ecosystems. Yet predictions are only as good as the assumptions behind them, and some of those assumptions bear critical examination. Here I briefly review some current assumptions about climatic controls of vegetation distribution and carbon dynamics, identify some potential problems, and suggest some possible solutions. Even after decades of research, our understanding of the climatic controls of vegetation distribution remains somewhat weak. The impracticality of conducting detailed mechanistic studies on thousands of species within each of several plant life forms (or functional types) means that our understanding of the climatic controls of vegetation distribution relies mostly on correlation, informed by only a handful of detailed physiological and life-history studies. But mechanisms that are seemingly plausible based upon correlation can prove to be incorrect. I examine this issue in greater detail, focussing particularly on a case study of the climatic controls of the distributions of coniferous and hardwood forests. Other models are concerned with predicting changes in terrestrial carbon dynamics. Most current models of terrestrial carbon dynamics at continental and global scales focus on plant physiology (growth rates) while ignoring demography (birth and death rates). In contrast, I present a simple model suggesting that (1) demographic processes influence carbon pools and fluxes, (2) very small changes in demographic rates may be amplified into very large changes in carbon dynamics, and (3) large effects are possible without any change in growth rates. These findings, coupled with evidence that ongoing global environmental changes already may be altering demographic rates in otherwise undisturbed ecosystems, lead me to conclude that models ignoring demography may significantly misrepresent environmentally-induced changes in carbon dynamics. We need ways of scaling up from individual-based models that do consider demography.
 
 

Earth's climate is changing. There is compelling evidence that human activity is playing a substantial role in global climate change but just how large a role is still uncertain. Predicting how climate may change in the 21st century is a challenging research problem set within a broader context of economic and political debate. This presentation will address a series of questions that often arise in discussions on future climate change: How is climate changing now? What is the evidence for global warming? What tools do we have to predict climatic change? How can we predict climate if we can't predict weather? What we think we know about climate change will be discussed ranging from the more confident (overall global warming) to less confident (regional and episodic impacts). Important unknowns in climatic change prediction include: The response of clouds, feedbacks from changes in the carbon cycle and vegetation, and climate-altering shifts in ocean circulation. The status of global climate simulation will be discussed. Can we hope to have a credible predictive capability before we get the answer from the global experiment we are already performing?
 
 

Discussion session:
 
 

The notion that the poor and hungry are most directly affected by environmental change is a common assertion in the development, economics and geography literatures. This paper examines this hypothesis, as well as how cotton production has influenced the rural poverty-environment dynamic in the Sahelian nation of Mali. The findings of the paper suggest that it is the vulnerability of the wealthy that has been more affected by environmental change than that of the poor. Critical to the analysis is an examination of changing livelihood strategies over time in areas with different rates of environmental change. While the rich are less vulnerable to hunger at any one point in time, the two major components of vulnerability, exposure to production shocks and ability to recover, have been eroded over time for the rich while the situation of the poor has improved. Not only did the vulnerability of the rich in the degraded area increase relative to that of the poor, but it increased more rapidly than that of the rich and poor in an ecologically stable area that served as a control. The paper's findings are based on household interviews describing past and present food economies, discussions with national policy makers, and land-cover change analysis based on remotely sensed NDVI trends.

Key words: poverty-environment interactions, political ecology, vulnerability, cotton, West Africa.