27th Annual Meeting and Symposium of the
Desert Tortoise Council, March 22-24, 2002
Abstracts

Long-term Vegetation Dynamics in the Southwestern U.S.: Empirical Needs for Predicting Future Vegetation Change

Julio L. Betancourt
Desert Laboratory, USGS, 1675 W. Anklam Rd., Tucson, AZ 85745

Deterministic and statistically-based biogeographic models, coupled with Atmospheric General Circulation Models, are being used to predict large-scale vegetation responses to future climate change. One deterministic model relates site water balance to the distribution of physiognomic types or biomes by way of the difference between potential and actual transpiration over an annual cycle; the direct effects of CO2 are simulated through changes in stomatal conductance. This particular deterministic model predicts greater summer moisture and eventual conversion of much of our southwestern desertscrub into grassland. By contrast, statistical biogeographic models envision plant distributions as stochastic, spatial realizations of response surfaces, decision trees, and bioclimatic envelopes- statistical functions that describe the way in which each species' expected distribution and abundance depend on the combined effects of several environmental variables. Both the deterministic and statistical biogeographic models yield impressive maps contrasting modern (Map A) with potential (Map B) vegetation under doubled-CO2 climatic scenarios, assuming equilibrium conditions. When forced by climatic conditions under double modern CO2, one statistical model predicts expansion of Joshua tree all the way to west Texas.

What these biogeographic models don't yield is a dynamic view of the slow and complex ecological processes involved in getting from Map A to Map B, a progression that will surely take more than a millennium to complete. These processes include migration, succession, biotic interactions, and hydrological and biogeochemical processes that vary with the life histories of the organisms, the initial conditions and intrinsic rates of change. For most organisms and ecosystems, there are currently few empirical data to support dynamic biogeographic models that can predict transitional stages between Map A and Map B for any point in time and space. These are the kinds of predictions that might prove most useful for anticipating consequences of climate change for ecosystem and resource management on decadal to century scales. I will present a few examples of how historical, large-scale analyses of migration and demography can serve empirical needs for anticipating ongoing and future vegetation change in the western U.S.

Natural migrations provide model systems for understanding biotic responses to global change and invasions of non-native species. Here, I'll draw from the fossil record of plant migration (e.g., creosote bush, Utah juniper, pinyon, ponderosa pine) of western North America to highlight the influence of environmental heterogeneity in dictating patterns of establishment and spread, and of climatic variability in pacing migration. I have purposefully selected ongoing natural migrations that are now being modulated by contemporary land use. Current theory of biotic invasions emphasizes population processes of dispersal, establishment, and expansion, where environmental heterogeneity is typically treated as a binary classification of favorable and unfavorable sites, and climatic variation as stochastic variation about a mean. Favorable sites, however, may range from places that can be occupied only with continual immigration to those where populations can persist independently despite environmental variability. In heterogeneous landscapes, the most favorable sites may occur far from the advancing front, and migration can thus follow unexpected pathways. Climate variability may set the tempo, with unfavorable climate imposing a prolonged colonization phase, and periods of rapid population spread reflecting episodes of favorable climate rather than attainment of population "critical mass." To some extent, the record of non-native species invasions mimics the fossil record. Most introductions of non-native species appear to fail multiple times before they eventually succeed. Both multiple failures and eventual successes are generally attributed to some combination of demographic and environmental stochasticity. A key distinction between past and contemporary invasions is the increasing probabilities of long-distance dispersal associated with our modern road and transportation system.

The second part of my presentation will focus on regional synchrony of disturbance and demographic phenomena. Many biotic and abiotic factors exhibit synchrony over large geographic areas. Understanding regional synchrony is essential to evaluate theory about what regulates local and regional populations, and for predicting thresholds and other nonlinear processes associated with vegetation change. For example, broadscale plant mortality during catastrophic drought and other disturbances (fire, insect and pathogen outbreaks) resets demographic clocks and opens niches for new recruitment across the region. Succession of the dead can involve individuals of the same species, other local species that increase at least temporarily, or extralocal species concurrently undergoing migration. Studies of synchrony require extensive regional networks of long-term data, multiyear for birds and mammals to multicentury for conifers. Tree-ring data can be used to examine regional synchrony, not only in tree growth, but also in vegetation disturbance and demography. Synchrony is embodied best in the ability to cross-date a Douglas fir from southern Arizona with a ponderosa pine from northern New Mexico. Tree growth, fire occurrence, insect outbreaks, and demography are entrained across the southwestern U.S. by cool season (frontal) precipitation, which varies little in space but greatly in time, usually in lockstep with large-scale indices of climate such as SOI, PNA, PDO, and CTI. The cause of regional synchrony is evident not in comparing time series from one site to another, but in correlating spatially-aggregated histories of disturbance or demography to independent, dendroclimatic reconstructions. Understanding the scales and causes of regional synchrony, as well as knowing the disturbance and demographic histories of regional vegetation, are essential for forecasting future vegetation change. I will use disturbance and demographic histories of dominant trees in the southwestern U.S. to illustrate the effects of interannual and decadal-scale climate variability on vegetation.

In the final analysis, I will present evidence that the growing season has lengthened in recent decades. The coincidence of a longer growing season with a warm Pacific and wetter Southwest has yielded unusual warm, wet springs ideal for tree growth at upper treelines, a surge in woody plant recruitment at all elevations, and the spread of winter-flowering, nonnative grasses such as red brome. Because precipitation will always vary both year to year and decadally, predictions of future plant migration and population dynamics should now encompass the consequences of a longer growing season during both wet and dry episodes.