BACKGROUND
State-and-transition models are an increasingly common conceptual method of organizing knowledge of successional dynamics in ecosystems. Unlike older successional theories they allow for multiple successional pathways, do not assume a single “climax” and do not assume that successional change is reversible. They assume that a given ecosystem might be capable of shifting among multiple stable states, when a specific “trigger” leads to a transition. The model seeks to identify the primary stable states, and key phases within those states, and the transitions among them. States and phases are defined by structural properties such as dominant plants and other biota or biodiversity, and functional properties/processes such as frequent fire return intervals or high soil aggregate stability. The properties of states and phases may be dynamic (e.g. plant cover & community composition) or inherent (e.g. climate, soil properties, landscape physiography), and may confer resistance (the ability to maintain the same set of structural/functional properties, i.e. not change) or resilience (the ability to experience change but return to the original set of structural/functional properties) to ecosystem change. (examples: 1) a dynamic property conferring resistance might be biological crust cover which decreases erodibility, 2) an inherent property conferring resistance might be high rock cover which prevents various soil disturbances from resulting in erosion, 3) a dynamic property conferring resilience might be rhizomatous grass cover: grazing animals remove biomass, and removal of the grazers when the forage quantity becomes poor results in a recovery of biomass due to resprout from tillers, 4) an inherent property conferring resilience might be a climate which favors a wet season after the grazing season, and facilitates recovery. Phases within states are different varieties of a state, and transitions from one phase to another tend to be fairly common, easily triggered, and often reversible. However there are particular phases which are at-risk to transition to degraded states, due to their structural or functional properties. Dynamics within a state, i.e. shifts from phase to phase, tend to be characterized by negative feedbacks which confer resilience. In contrast, transitions among stable states, may be abrupt, may exhibit threshold behaviors. Shortly after a threshold crossing, simple facilitating practices (e.g. cessation of grazing) may allow a reversal of the state change. If the transition progresses too far beyond the threshold, neither resiliency of the reference state or facilitating practices can reverse the transition on time scales relevant to management, and must be addressed using active –and often costly- restoration practices. Every transition has an underlying trigger, for example introduction of grazing. The trigger initiates a change in structural or functional properties of the ecosystem (e.g. loss of vegetative cover, and disruption of soil surface aggregates), which may engage new processes which bring about a transition to a new state (e.g. a positive feedback loop wherein lack of vegetation and soil aggregates leads to erosion which prevents the recolonization of vegetation or creation of aggregates). These transitions may be monitored with indicators.
If you would like more background on the current state-and-transition modeling framework we suggest reading:
Briske et al. 2008. Rangeland Ecology & Management 61:359–367.
Bestlemeyer et al. 2009. Rangeland Ecology & Management 62:1-15.
The best way to understand a state-and transition model is to study one. Please review the following model and answer the questionnaire below it.
CANDIDATE MODEL
Limey Uplands (Wupatki)
Fig. 1. Location of limey uplands within Wupatki, and NPS monitoring plots within them (DeCoster et al. 2009).
Limey uplands are ecosites situated on top of fairly level basalt flows, experiencing 8-10” of rainfall per year. Despite the name, these are not limestone soils. Rather the name reflects the presence of carbonate in the upper horizons. The soil is weathered in place from the underlying basalt, and from later cinder deposits. The surface is currently gravelly with a large amount of the surface covered with cinders. These are vegetated by rhizomatous grasses, or more rarely savannahs, though there is sufficient rooting depth for trees.
States:
[Due to a history of volcanism and long-term human occupation, appropriate reference vegetative communities in Wupatki should be defined with care. Here we emphasize what is likely possible under management scenarios today. Paleoecological studies have described reference communities from a variety of time periods. Some are not possible under current conditions. States are divided into pre-historic and historic-modern (manageable) to make this distinction]:
Fig. 2. A state-and-transition model for Limey Uplands of Wupatki and surrounding areas.
Pre-historic States (not possible under current management scenarios):
P1: Pre-eruption. An original vegetation state under approximately modern climate regimes is difficult to reconstruct, and sketchy at best. Based upon packrat midden and palynological analysis, it is likely that grasslands were common (Cinnamon 1988). Plant macrofossils demonstrate that Juniperus monosperma was present (Cinnamon 1988; but not confirmed by Ironside 2006) and also seems to suggest a greater prevalence of some grasses (Achnatherum hymenoides, Enneapogon desvauxii), and the shrub Artemisia bigelovii. Other common species, also present today, include Hillaria jamesii, Chrysothamnus nauseosum, and Opuntia erinaceae. Notably Bouteloua spp. -a current co-dominant- were only sparsely represented. These vegetation reconstructions reflect the general area, rather than Limey Uplands specifically. Because biological crusts have very low potential on basaltic soils on the Colorado Plateau, they were likely a negligible ecosystem component (Bowker and Belnap 2007). A negative feedback of frequent ground fires (15-20 yr return interval) likely maintained grasslands and limited tree and shrub establishment (Cinnamon 1988, Hassler 2006).
P2: Post-eruption, post occupation. State P1 experienced two drastic changes which almost certainly led to rapid transitions. First the Sunset Crater eruption deposited up to several centimeters of volcanic ash and cinder possibly killing vegetation. Cinder deposition is thought to have improved water infiltration dynamics, and likely strongly altered hydrology (Sullivan and Downum 1991). Within a few decades an approximately century-long occupation of dry-farming societies ensued. These are though to have been abandoned due to declining soil nutrient stores (Sullivan and Downum 1991). Midden and palynological data indicates shifts in relative abundance of some taxa (Cinnamon 1988) including a decrease in Enneapogon, Artemisia bigelovii, and Achnatherum hymenoides, an apparent arrival of Heterostipa comata, and no major change in Hilaria jamesii. A negative feedback of frequent ground fires (15-20 yr return interval) likely maintained grasslands and limited tree and shrub establishment (Cinnamon 1988, Hassler 2006).
Modern manageable states:
S1: Current potential Grassland
S1Phase1: Rested Grassland. Prior to the initiation of grazing the area had over 500 years to recover from agricultural disturbance. We know little about this period, except that midden analysis suggests that relative abundance of plant macrofossils does not appear to have changed much over this time. Presumably cover increased. A range assessment of the region from the early 1970’s provides some grossly overoptimistic estimates of “climax” cover at 60-90%. The assessment also suggests that a greater prevalence of tussock species, and a lesser prevalence of the rhizomatous Hilaria spp., and Bouteloua spp. (Doughty 1971). These assertions were based upon general knowledge of the range assessor rather than specific knowledge of the ecosystem.
Midden evidence suggests that H. comata (a tussock grass) may well have been more important in the past, but the present rhizomatous dominants Bouteloua and Hilaria were also well-represented macrofossils from that time. The supposedly frequent ground fire interval (Cinnamon 1988, Hassler 2006) would tend to favor the rhizomatous species such as Hilaria and Bouteloua (Jameson 1962, Ford 1999). This fire return cycle also likely discouraged shrub and tree colonization, conferring resistance to state change in this phase.
S1Phase2: Denuded grassland.
Around 1900, grazing was introduced (Jameson 1962), and its intensity peaked early in that century. Based upon general knowledge of behavior of common plant species under grazing, the relative abundance of palatable grasses such as Heterostipa comata would be expected to decrease, and unpalatable species such as Guiterrizia and Chrysothamnus or grazing tolerant species such as Bouteloua gracilis might increase. Grazing decreases the standing biomass of fine fuels and their connectivity, decreasing the susceptibility of this system to fire. Jameson (1962) estimated that a 1956 wildfire in a grazed grassland had 13% total vegetative cover pre-burn, which might be taken as a reasonable first approximation of a minimum cover to sustain fire. This removal of biomass can proceed to extreme levels (Fig 2, left panel). This in turn allows more Juniperus recruits to invade. Invasion by Salsola may also be possible at this stage. The fire cycle which maintains S1P1 can be strongly compromised, making this phase at-risk of further change. However, due to resprout of rhizomatous grasses, it displays considerable resilience.
Fig. 2. (Jameson 1962). Left panel shows an extreme version of S1P2 in 1906. Right Panel shows S2P2 in 1961. [Photos do not appear to be specifically on Limey Uplands, but provide some evidence regarding the general dynamics of these systems.]
S2Phase1: Savannized-Denuded understory. To date, limey uplands have been less susceptible to juniper encroachment than surrounding areas but it is clear that the prevalence of juniper in increasing in the grasslands and former grasslands of Wupatki (Cinnamon 1988, Hassler 2006, Parker 2009, Ironside 2006) A comparison of basalt soils to limestone soils indicated that tree growth rate or density did not differ, but average age of establishment did occurred later on basaltic soils (Hassler 2006). There is sufficient rooting depth, but perhaps the soil texture challenges the junipers’ ability to colonize delaying the process (Bowker et al. unpublished). Simulation modeling based upon 20th century climate regimes indicate that limey uplands have a high probability of invasion (Ironside 2006), if that climate regime were to continue. If grazing keeps fines fuels low, fire cannot cull colonizing junipers making savannization likely. It is not known if this can proceed to a woodland state like some nearby areas on different soils, but this possibility is not favored by future climate projections thus is not considered here (Ironside 2006). Due to resprout of rhizomatous grasses, the understory displays considerable resilience.
S2Phase1: Savannized-Grassy. In this phase, Juniperus and other woody vegetation are already established, but the rhizomatous grasses of the understory have recovered to conditions reasonably similar to S1P1. The grass-fire-shrub recruitment feedback may reestablish itself, but cannot cull previously established trees (Fig. 1, right panel)
S3Phase1: Highly eroded. If heavy grazing continues unabated, a transition to a severely eroded state is possible. This state is theoretical in that sites in Wupatki were known to have been strongly denuded, yet subsequently recovered (Fig. 2). This state is dominated by bare ground, and would be characterized by erosional features such as rills, gullies, terracettes, and sheeting. Although the cinder-covered surface and generally flat slopes are inherent properties of this ecosite that lend it low erosion potential, such a reduction in vegetation cover could conceivably initiate erosion since water-stable aggregate structure of exposed soil surface is very poor (Generally <2 using Herrick soil stability test; Bowker and Belnap 2007). This state would be maintained because high erosion rates prevent revegetation, and the lack of vegetation prevents stabilization of soil.
S2Phase2: Highly eroded-Juniper overstory. If junipers were previously allowed to establish and grow to adulthood, and heavy grazing later reinstituted, we would expect a degradation of the soil surface similar to S3P1, with an overstory of high-lined juniper trees.
Transitions:
Transitions not linked to climate change
T1: Triggers: a) Cinder and deposition due to volcanism beginning in 1064 likely initially destroys much vegetation, but enhances infiltration and water retention dynamics of soils. b) From ~ 1200 - 1300 agricultural societies practiced dryfarming.
T2. Triggers: a) Abandonment and long term rest (centuries) allowed succession to proceed to a fire-maintained grassland.
T3. Triggers: a) Introduction of grazing b) fire suppression reduce both amount and connectivity of fine fuels.
Spatial scale: site-level, or level of management unit (e.g. range allotment, or NPS vs. BLM management)
Related indicators: grass standing biomass, total or basal cover of fine fuels, connectivity of fine fuels.
T4. Trigger: a) Reduction or removal of grazing, b) cessation of fire suppression practices allow recovery of fuel amount and connectivity.
Spatial scale: site-level, or level of management unit (e.g range allotment, or NPS vs. BLM management).
Related indicators: grass standing biomass, total or basal cover of fine fuels, connectivity of fine fuels.
T5. Trigger: woody plant colonization, linked to previous grazing or fire suppression activity. Must coincide with propagule availability of woody plants.
Spatial scale: site-level
Related indicators: Density or frequency of woody plants.
Likely obeys threshold behavior, but may be reversible in early stages with facilitative practices such as the cessation of grazing, or cessation of fire suppression practices.
T6. Trigges: a) Reduction or removal of grazing, b) cessation of fire suppression practices allows allow recovery of fuel amount and connectivity, but cannot cull established trees.
Spatial scale: site-level, or level of management unit (e.g. range allotment, or NPS vs. BLM management).
Related indicators: grass standing biomass, total or basal cover of fine fuels, connectivity of fine fuels (Leonard 2009, Knapp & Keely 2006).
T7. Trigger: Re-introduction of grazing reduce both amount and connectivity of fine fuels in understory.
Spatial scale: site-level, or level of management unit (e.g range allotment, or NPS vs. BLM management).
Related indicators: grass standing biomass, total or basal cover of fine fuels, connectivity of fine fuels.
T8. Triggers: Continued high intensity grazing, conceivably in combination with drought leads to plant mortality which initiates a feedback mechanism wherein erosion prevents plant colonization, and erodibility is high due to denuded plant cover.
Spatial scale: site-level, or level of management unit (e.g range allotment, or NPS vs. BLM). Drought may reflect a regional spatial scale.
Related indicators: total or basal cover of plants, erosional features such as rills, gullies, terracettes.
Likely obeys threshold behavior.
T9. Triggers: Continued high intensity grazing, conceivably in combination with drought leads to plant morality, initiating a feedback mechanism wherein erosion prevents plant colonization, and erodibility is high due to denuded plant cover.
Spatial scale: site-level, or level of management unit (e.g range allotment, or NPS vs. BLM). Drought may reflect a regional spatial scale.
Related indicators: total or basal cover of plants, erosional features such as rills, gullies, terracettes.
Likely obeys threshold behavior.
Transitions linked to climate change
[Changes in grassland composition are difficult to predict. Higher CO2 favors C3 grasses, whereas increased prevalence of drought and higher temperatures favors C4 grasses. These opposing forces may balance one another to preserve the current C4 dominance. Because tree death due to drought has actually been observed near Wupatki in savannahs in marginal habitat for junipers (Bowker et al. unpublished, Gitlin et al. 2006), we focus on this change]
T10. Triggers: Extreme drought in combination with high temperatures and edaphic (clayey soils) and physiographic (southerly aspects) stress leads to juniper mortality events, potentially en masse.
Spatial scale: drought is regional, edaphic or physiographic stressors are site-scale.
T11. Triggers: Extreme drought in combination with high temperatures and edaphic (clayey soils) and physiographic (southerly aspects) stress leads to juniper mortality events, potentially en masse.
Spatial scale: drought is regional, edaphic or physiographic stressors are site-scale.
T12. Triggers: Extreme drought in combination with high temperatures and edaphic (clayey soils) and physiographic (southerly aspects) stress leads to juniper mortality events, potentially en masse.
Spatial scale: drought is regional, edaphic or physiographic stressors are site-scale.
Literature Cited
Bowker, M.A. and Belnap, J. 2007. Spatial Modeling of Biological Soil Crusts to Support Land Management Decisions: Indicators of Range Health and Conservation–restoration Value Based Upon the Potential Distribution of Biological Soil Crusts in Montezuma Castle, Tuzigoot, Walnut Canyon, and Wupatki National Monuments, Arizona.
Bowker, MB, Munoz, AA, Martinez, T., Lau, MK. 2010. Rare drought-induced mortality of juniper: edaphic and climatic stressors promote hydraulic failure. unpublished manuscript.
Cinnamon, S.K. 1988. The vegetation community of Cedar Canyon, Wupatki National Monument as influenced by prehistoric and historic environmental change. Master’s Thesis, NAU.
DeCoster, J. K., and M. C. Swan. 2009. Integrated upland vegetation and soils monitoring for Wupatki National Monument: 2008 summary report. Natural Resource Data Series NPS/SCPN/NRDS—2009/022. National Park Service, Fort Collins, Colorado.
Doughty, J.W. 1971. Soil survey and range site and condition inventory. Wupatki National Monuments, Arizona. A special report. USDA SCS.
Ford, PL. 1999. Response of buffalograss (Buchloe dactyloides) and blue grama (Bouteloua gracilis) to fire. Great Plains Research 9: 261-76.
Gitlin, AR, Sthultz, CM, Bowker, MA, Stumpf, S., Paxton, K.L., Kennedy, K., Munoz, A., Bailey, J.K., Whitham, TG. 2006. Mortality gradients within and among dominant plant populations as barometers of ecosystem change during extreme drought. Conservation Biology 20: 1477-1486.
Hassler, F. 2006. Dynamics of juniper invaded grasslands and old growth woodlands at Wupatki National monument, Northern Arizona, USA. Master’s Thesis, Northern Arizona University.
Ironside, K. 2006. Climate change research in national parks; paleoecology, policy, and modeling the future. Master’s Thesis, Northern Arizona University.
Jameson, D.A. 1962. Effects of Burning on a Galleta-Black Grama Range Invaded by Juniper. Ecology 43: 760-763.
Knapp, E.E., and J.E. Keeley. 2006. heterogeneity in fire severity in early and late season prescribed burns in a mixed conifer forest. International Journal of Wildland Fire 15: 37-45.
Leonard, S. 2009. Predicting sustained fire spread in Tasmanian Native grasslands. Environmental Management 44: 430-440.
Miller, ME, Witwicki, DL, Mann, RK, Tancreto, NJ. 2007. Field evaluation of sampling methods for long-term monitoring of upland ecosystems on the Colorado Plateau. USGS Open File Report 2007-1243.
Sullivan, AP, Downum, CE. 1991. Aridity, activity, and volcanic ash agriculture: a study of short-term prehistoric cultural-ecological dynamics. World Archaeology 22: 271-287.
QUESTIONNAIRE
Questions marked *** are required, we cannot use your response if any of these are omitted. Just indicate your responses next to the option, or in the blank space provided The other responses are optional, and much appreciated, we encourage you to at least read them to see if there is anything you can contribute. Please email me the completed questionnaire.
We expect a respondent to answer based on past experience, accumulated knowledge, educated guesses, and general principles. A respondent may use any source of information that is already known to them, but are not expected to conduct an extensive literature review to uncover new information. Treat this as if you were reviewing a manuscript for a journal.
***1) Please identify any states or phases which should be omitted from the state-and-transition model. You may select more than one.
a. P1
b. P2
c. S1
d. S1P1
e. S1P2
f. S2
g. S2P1
h. S2P2
i. S3
j.S3P1
k. S3P2
l. None, all should be retained
2). Please identify any manageable states or phases which are currently not in the model, but should be added to the state and transition model.
[Please briefly list structural properties like dominant species or overall vegetative cover (whatever you feel is important to mention), and functional properties & processes such as fire return intervals, or low soil stability. When you list properties please think about and indicate if they are dynamic or inherent, and if they contribute to the resistance or resilience of the state or phase. Please indicate any feedback mechanisms which tend to maintain these states. If you are identifying a phase, is it at-risk? Let us know about appropriate literature if you know of it.]
***3) Please identify any transitions which should be omitted from the state-and-transition model. You may select more than one.
a. T1
b. T2
c. T3
d. T4
e. T5
f. T6
g. T7
h. T8
i. T9
j. T10
k. T11
l. T12
m. none, all should be retained
2) Please identify any transitions which are currently not in the model but should be added to the state and transition model.
[For each addition provide, the starting state and ending state for which the transition applies. Identify plausible trigger mechanisms. Also please provide a brief explanation of the process that brings about the transition, e.g. fire, insect outbreak, drought, grazing. Please note the dominant scale of the trigger mechanism, and the importance of temporal convergence and order with other mechanisms (e.g. simultaneous drought and grazing may function as a trigger when either alone do not). Also suggest monitorable indicators.]
***5) Please estimate your overall confidence that a new revised model which takes into account your proposed modifications is the correct model of the most important ecosystem states, processes and dynamics of the ecosite in question.
[Please answer on a subjective scale of 0 – 100% certainty. Enter any value in this range. To help you answer: 0% means “It’s anyone’s guess, this model is no better than any other model”, 50% means “Because this model is reasonable I would tend to believe it until evidence to the contrary is presented”, 100% means “The model is so well-supported by evidence and accumulated knowledge, that I am certain it is correct.”]
6) If your level of confidence in any particular state or transition differs from the value above please estimate your confidence for that model component in the appropriate box. In case you are estimating a confidence in a state or transition suggested by you in questions 2 and 4, please use the “other” options to identify it.
[If you do not provide answers to 6 we will assume they are the same as the answers to 5 in all cases.]
a. P1
b. P2
c. S1
d. S1P1
e. S1P2
f. S2
g. S2P1
h. S2P2
i. S3
j.S3P1
k. S3P2
l. T1
m. T2
n. T3
o. T4
p. T5
q. T6
r. T7
s. T8
t. T9
u. T10
v. T11
w. T12
x. Other state or phase (please identify)
y. Other state or phase (please identify)
z. Other transition (please identify)
aa. Other transition (please identify)
***7) Please take a moment to think of any scientist or other person, who is to your knowledge the best qualified to develop a state-and-transition model for this ecosite. This person could be yourself, or any other person. “Best” qualified may or may not mean highly qualified; sometimes no one is highly qualified. Now, in the hypothetical scenario that this person had prepared a state-and-transition model for this ecosite using all of the data, knowledge and experience available to them, estimate how much confidence you would have that it is the correct model of the most important ecosystem states, processes and dynamics of the ecosite in question.
[Please answer on a subjective scale of 0 – 100% certainty. Enter any value in this range. To help you answer: 0% means “It’s anyone’s guess, any person could produce an equally good or bad model”, 50% means “Because this model is reasonable I would tend to believe it until evidence to the contrary is presented”, 100% means “The model is so well-supported by evidence and accumulated knowledge, that I am certain it is correct”.]
8.) Who is the best qualified person (from question 7) to develop this state-and transition model? This response will help us ensure we have contacted all of the right people.
9.) How much time did you spend on this? This is important for improving future iterations of this questionnaire.
10.) How would you improve this survey? Is it too long, too jargony, etc.?
Thank you for taking time out of your schedule! We will use your comments to revise our model, and will contact you again about phase 2 of the survey.
No comments:
Post a Comment