Fred E. Smeins

State-and-Transition Models, Thresholds, and Rangeland Health: A Synthesis of Ecological Concepts and Perspectives

Abstract This article synthesizes the ecological concepts and perspectives underpinning the development and application of state-and-transition models, thresholds, and rangeland health. Introduction of the multiple stable state concept paved the way for the development of these alternative evaluation procedures by hypothesizing that multiple stable plant communities can potentially occupy individual ecological sites. Vegetation evaluation procedures must be able to assess continuous and reversible as well as discontinuous and nonreversible vegetation dynamics because both patterns occur and neither pattern alone provides a complete assessment of vegetation dynamics on all rangelands. Continuous and reversible vegetation dynamics prevail within stable vegetation states, whereas discontinuous and nonreversible dynamics occur when thresholds are surpassed and one stable state replaces another. State-and-transition models can accommodate both categories of vegetation dynamics because they represent vegetation change along several axes, including fire regimes, weather variability, and management prescriptions, in addition to the succession-grazing axis associated with the traditional range model. Ecological thresholds have become a focal point of state-and-transition models because threshold identification is necessary for recognition of the various stable plant communities than can potentially occupy an ecological site. Thresholds are difficult to define and quantify because they represent a complex series of interacting components, rather than discrete boundaries in time and space. Threshold components can be categorized broadly as structural and functional based on compositional and spatial vegetation attributes, and on modification of ecosystem processes, respectively. State-and-transition models and rangeland health procedures have developed in parallel, rather than as components of an integrated framework, because the two procedures primarily rely on structural and functional thresholds, respectively. It may be prudent for rangeland professionals to consider the introduction of these alternative evaluation procedures as the beginning of a long-term developmental process, rather than as an end point marked by the adoption of an alternative set of standardized evaluation procedures.

A Unified Framework for Assessment and Application of Ecological Thresholds

Abstract The goal of this synthesis is to initiate development of a unified framework for threshold assessment that is able to link ecological theory and processes with management knowledge and application. Specific objectives include the investigation of threshold mechanisms, elaboration of threshold components, introduction of threshold categories and trajectories, and presentation of an operational definition of ecological thresholds. A greater understanding of ecological thresholds is essential because they have become a focal point within the state-and-transition framework and their occurrence has critical consequences for land management. Threshold occurrence may be best interpreted as a switch from the dominance of negative feedbacks that maintain ecosystem resilience to the dominance of positive feedbacks that degrade resilience and promote the development of post-threshold states on individual ecological sites. Threshold categories have been identified to serve as ecological benchmarks to describe the extent of threshold progression and increase insight into feedback mechanisms that determine threshold reversibility. Threshold trajectories describe the developmental pathway that post-threshold states may follow once a threshold has been exceeded. These trajectories may produce a continuum of potential post-threshold states, but the majority of them may be organized into four broad states. This framework lends itself to management application by providing an operational definition of thresholds that is based on a probabilistic interpretation. Probabilities associated with 1) the occurrence of triggers that initiate threshold progression, 2) the trajectory of post-threshold states, and 3) threshold reversibility will provide an operational procedure for threshold assessment and application. If thresholds are to play a central role in rangeland ecology and management, then the rangeland profession must accept responsibility for their conceptual development, ecological validity, and managerial effectiveness.

Simulation of a fire-sensitive ecological threshold: a case study of Ashe juniper on the Edwards Plateau of Texas, USA

A model was developed to represent the establishment of a fire-sensitive woody species from seeds and subsequent survival and growth through five size classes. Simulations accurately represent structural changes associated with increased density and cover of the fire-sensitive Ashe juniper (Juniperus ashei, Buckbolz) and provide substantial evidence for multiple steady states and ecological thresholds. Without fire, Ashe juniper increases and herbaceous biomass decreases at exponential rates until a dense-canopy woodland is formed after approximately 75 years. Maintenance of a grass-dominated community for 1.50 years requires cool-season fires at a return interval of less than 25 years. When initial cool-season fires are delayed or return intervals are increased, herbaceous biomass (fuel) decreases below a threshold and changes from grassland to woodland become irreversible. With warm-season fires, longer return intervals maintain grass dominance, and under extreme warm-season conditions even nearly closed-canopy stands can be opened with catastrophic wildfires.

Predicting grassland community changes with an artificial neural network model

Abstract Artificial neural networks are parallel processing systems with the ability to learn by example and generalize from inferred patterns. In this application, a neural network model of the feedforward, backpropagation type is designed to predict future community composition from knowledge of present climatic factors and species cover. Training and testing data are drawn from a 30-year record of the environmental and vegetative variables of a grassland community. The resulting trained network is capable of forecasting accurately up to 4 years into the future. The results indicate a potential usefulness of neural network technology for non-mechanistic modeling in ecological research and management.

The influence of soil depth on plant species response to grazing within a semi-arid savanna

Grassland patches within a semi-arid savanna were evaluated over 45-years for (1) local temporal dynamics of basal area for five dominant grass species within long-term heavily grazed and ungrazed treatments, (2) the influence of soil depth (resource availability) on vegetation dynamics, and (3) the applicability of community-level grazing response groups over fine-scale patterns of soil heterogeneity. Temporal patterns in species composition and basal area were dependent upon soil depth. In the heavy grazed treatment, Hilaria belangeri dominated deep soils while Erioneuron pilosum and Bouteloua trifida were restricted to shallow soils. In the ungrazed treatment, removal of grazing resulted in successional changes that were significantly different across soil depths. After 45 years without grazing, Eriochloa sericea was most abundant on deep soils while Bouteloua curtipendula was more abundant on intermediate and shallow soils. Community-level functional groups that are based on grazing were not appropriate when multiple pattern-driving variables were considered across multiple scales indicating that functional groups should only be applied to certain processes at specific scales. Within the ungrazed treatments, variable soil depths have resulted in a shifting mosaic in time and space where early- and late-successional species co-exist continuously but spatially separated within the community. In the heavily grazed treatment, species are somewhat spatially arranged by soil depths, but much of the inherent heterogeneity is eliminated and species composition is dominated by the three grazing-resistant short-grasses. Broad scale successional changes may appear linear and predictable while at finer scales, the same changes may be described as non-linear and dependent upon soil depth resulting in thresholds that are partially explained by weather patterns, seed bank limitations and competitive inhibitions.

Differential Influence of Weather on Regional Quail Abundance in Texas

Although weather variables are known to influence quail abundance in some habitats, most studies have addressed only limited geographic areas and indices to weather conditions. The few replicated studies addressed relatively similar climate zones. We used 21 years (1978-98) of quail abundance data collected by the Texas Parks and Wildlife Department (TPWD) biologists to address the relationship between both simple precipitation and Palmer drought indices and Northern Bobwhite (Colinus virginianus) and Scaled quail (Callipepla squamata) abundance in 6 ecological regions of Texas. Three 12-month Palmer indices were more highly correlated with changes in Northern Bobwhite abundance in the South Texas Plains ecological region than was raw precipitation alone. The 12-month Modified Palmer Drought Severity Index (PMDI) was correlated (r s ≥ 0.78, P ≤ 0.001) with the mean number of Northern Bobwhites visually observed per survey route in the Rolling and South Texas Plains ecological regions, while a 12-month, raw precipitation index was correlated (r s = 0.64, P = 0.002) with Northern Bobwhite abundance in only the South Texas Plains. The PMDI and raw precipitation were correlated (r s ≥ 0.67, P ≥ 0.001 and r s ≤ 0.57, P ≤ 0.007, respectively) with the mean number Scaled Quail observed per survey route in the Edwards Plateau, South Texas Plains, and Trans-Pecos Mountains and Basins ecological regions. There was no relationship (P ≥ 0.437) between changes in quail abundance and the PMDI or raw precipitation in the Gulf Prairies and Marshes physiographic region, where precipitation was relatively high. The monthly PMDI was a better indicator of changes in both northern bobwhite and Scaled Quail abundance among years than was monthly precipitation alone. Both monthly and 12-month precipitation-based weather indices were more correlated with changes in Northern Bobwhite and scaled quail abundance among years in relatively dry as opposed to wet ecological regions. Our approach should help wildlife biologists and managers better account for annual variability in quail productivity in semi-arid environments so that long-term populations trends can be better elucidated.

Multiple-scale habitat modeling approach for rare plant conservation

Multiple-scale habitat assessment for rare plants is an important component of conservation and development planning. It is challenging, however, due to lack of information synthesis on the ecology of rare plants, lack of effective approaches for habitat assessment at multiple spatial scales, and lack of spatial data for relevant environmental attributes and scales. A multiple-scale habitat modeling approach was developed to meet this need. Regional-, landscape-, and site-scale habitat models were developed for eight rare plant species found in southern Texas, USA. The models were partially validated and used for planning of rare plant conservation and highway construction. Regional-scale habitat models were used to predict, based on coarse-scale geographic information system (GIS) data, spatial distribution of areas containing potential habitat of rare plant species and the probability of encountering potential rare plant habitats. Site-scale models, based on synthesis of the literature and field investigations, were developed for field survey and mapping of rare plant habitats to enable accurate assessment of potential and present habitat suitability of specific locations using fine-resolution field data on soil, landform and vegetation structure. The greatest need for assessing the presence and potential habitat of rare plants is at the landscape scales. Thus, landscape-scale models were developed for spatially explicit assessment of potential and present habitat suitability, based on site-scale models but using GIS and remote sensing-based data. These models can be used as effective tools for conservation planning, monitoring and management of rare plant habitat, as well as for reduction of land use conflicts and development cost. The processes of model development and application synthesizes the diffuse literature, identifies knowledge and data gaps to guide future research, and provides a framework for assimilating new information acquired in the future to improve habitat assessment.

Vegetation of a 25-year exclosure on the Edwards Plateau, Texas.

Bohn, York St., London. 266 p. Daubenmire, R. 1968. Ecology of fire in grasslands. Advance. in Ecol. Res. 5~209-266. Day, Arthur G. 1940. Coronado’s quest. Univ. of Calif. Press, Berkeley. 419 p. Dunmire, R. G., and 0. W. Bidwell. 1959. Soil Survey of Geary County, Kansas. U.S. Dep. A&r. Soil Conserv. Serv. Series 1955, No. 6. 35 p. Gleason, Henry A. 1913. The relation of forest distribution and prairie fires in the Middle West. Torreya 13(8): 173-181. Herbel, C. H., and Kling L. Anderson. 1959. Response of true prairie vegetation on major Flint Hills range sites to grazing treatment. Ecol. Monogr. 29(2):171-186. Kollmorgen, W. M., and D. S. Simonett. 1965. Grazing operations in the Flint Hills-bluestem pastures of Chase County, Kansas. Ann. Ass. Amer. Geogr. 55:260-290. Komarek, E. V., Sr. 1966. Meteorological basis for fire ecology. Proc. 5 th Annu. Tall Timbers Fire Ecol. Conf. p. 85-l 23 Komarek. E. V., Sr. 1968. Lightning and lightning fires as ecological forces. Proc. 8th Annu. Tall Timbers Fire Ecol. Conf. p. 169-197. Kucera, C. L. 1960. Forest encroachment in native prairie. Iowa State J. of Sci. 34(4):635-639. Ktichler, A. W. 1964. Potential natural vegetation of the conterminous United States. Amer. Geogr. Sot. Spcc. Pub. 36. 116 D. Lewis, Meriwether. 1814. The expedition of Lewis and Clark, Vol. 1. Ann Arbor Univ. Microfilms, Inc. 1966. 470 p. Malin, .I. C. 1942. An introduction to the history of the bluestem pasture region of Kansas. Kansas Hist. Quart. 11:3-28. Newberry, .I. S. 1873. Origin of the prairies. Geol. Surv. of Ohio, Vol. I, Part 1:26-31. Owensby, Clenton E., Ed. R. Smith, and Kling L. Anderson. 1973. Deferred-rotation grazing with steers in the Kansas Flint Hills. J. Range Manage. 26: 393-395. Rowe, J. S. 1969. Lightning fires in Saskatchewan grassland. Can. Field Natur. 83(4):317-324. Sauer, C. 0. 1950. Grassland climax, fire and man. J. Range Manage. 3:16-21. Shimek, F. 1911. The prairies. Bull. Natur. Hist. of Iowa 6(1):169-240. Webb, Walter P. 1931. The Great Plains. Ginn and Co., Boston. 398 p.

Spatial scale influence on longterm temporal patterns of a semi-arid grassland

Longterm (45 years) temporal data were used to assess the influence of spatial scale on temporal patterns of a semi-arid west Texas grassland. Temporal basal area dynamics of common curlymesquite (Hilaria belangeri (Steud.) Nash) collected from permanent plots within two areas that were released from disturbance (longterm overgrazing and drought), were evaluated at two spatial scales (quadrat, site). Wiens (1989) proposed hypotheses to characterize the influence of scale on variability, predictability, and equilibrium. These hypotheses were tested for this grassland and temporal patterns observed were different for each spatial scale. The large scale (site) was characterized by low variation between units, high variation within units, high potential predictability, and possible movement toward a fluctuating but relatively stable or equilibrial state. At the small scale (quadrat), variation between units was high, predictability low, and there was no indication of movement toward a stable state; chaotic behavior may be expressed at this scale although the length of the temporal record may not be sufficient to evaluate this phenomenon.

The Combined Influence of Grazing, Fire, and Herbaceous Productivity on Tree–Grass Interactions

Although Juniperus communities are native to most regions of North America, they have proliferated in many areas of the Great Basin and Great Plains that historically supported grasslands, shrublands, and savannas. Explanations for the observed increases in Juniperus dominance, as well as other woody plant communities, are the subject of ongoing debate. The balance between herbaceous and woody vegetation is regulated by complex interactions between climate (e.g., amount and seasonality of rainfall), soils (e.g., soil texture and depth), and disturbance regimes (e.g., fire, gazing, browsing) (Walker 1987; Scholes and Archer 1997; Higgins et al. 2000). Changes in one or more of these factors can potentially elicit a change in the ratio of woody to herbaceous plants. Accordingly, climate change, intensification of grazing, elimination of fire and browsing (Hastings and Turner 1965; Grover and Musick 1990; Archer 1994; Fuhlendorf et al. 1996), atmospheric CO 2 enrichment (Idso 1992; Johnson et al. 1993), and nitrogen deposition (Kochy and Wilson 2001) have all been invoked as potential reasons for woody plant proliferation over the past century (see reviews by Archer 1994; Van Auken 2000). However, because these factors are correlative and interact across multiple spatiotemporal scales, it is neither feasible nor realistic to assess their relative importance using traditional, short-term factorial experiments. Field studies based on space-for-time substitutions and comparisons of landscapes with differing management histories have been used to assess long-term changes, but results from such studies are difficult to replicate, interpolate, or extrapolate and do not explicitly test causality. As a result, there is still considerable debate as to the relative importance of grazing, climate, and fire influences on community dynamics in drylands (O’Connor 1995; Fernandez-Gimenez and Allen-Diaz 1999; Illius and O’Connor 1999; Fuhlendorf et al. 2001). Dynamic simulation modeling is an underutilized tool that can be used to evaluate how climate or climate–disturbance interactions potentially affect tree– grass ratios and to test competing hypotheses attempting to account for woody plant increases over the past century. Grazing, fire, and climate strongly interact to influence woody plant abundance via effects on herbaceous production and composition. High rainfall can promote