COOPERATIVE STUDY PROPOSAL
Approved by the Inyo/Los Angeles Standing Committee March 23, 2000.
Project Title:
DEVELOPMENT OF A DEMOGRAPHIC MODEL FOR NEVADA SALTBUSH
Principal Investigators:
Inyo County Water Department - Sally Manning
Los Angeles Dept. of Water and Power - Dave Martin
Consultant(s) - to be selected
Purpose:
According to the Green Book (V.B.9, p. 121), a study of demographics of Owens Valley plant species should be conducted. The purpose of this study is to develop a stage-based demographic model for the native, invasive shrub, Nevada Saltbush (Atriplex lentiformis ssp. torreyi (S. Watson) = ATTO). ATTO populations in the Owens Valley have been intensively monitored for several years, but the data have been used only retroactively: to document changes that have already occurred to ATTO populations. Development of a demographic model would allow existing data to be used pro-actively: to estimate the probability of populations reaching certain sizes in the future given various assumptions about environmental factors. In addition, the model could be used to simulate ATTO=s response to known or hypothetical management conditions. Model development would also allow a sensitivity analysis to be performed in which the points in the species= life cycle having the most impact on population growth would be identified. Identification of such points could be extremely useful to determine the nature and timing of intervention which could be implemented to control ATTO in places where its invasion could cause a conversion in vegetation type that is not allowed under the long-term water agreement. Another benefit of model development would be the potential improvement of Green Book monitoring protocols. Modeling may show certain data currently gathered to be of little value, and/or that other data would be very valuable to record. Finally, by developing demographic modeling tools for ATTO, the Technical Group would be equipped to apply the tools to other important species and thus achieve the Green Book objective of understanding vegetation responses to hydrologic management.
Background:
Relevance of Nevada Saltbush
ATTO is a native, perennial, phreatophytic shrub. The species appears to thrive in areas of saline soils where humans or their activities have negatively affected the native grass cover (Manning 1997). According to the long-term water agreement, conversion of native alkali meadows (dominated by native grasses) to sites dominated by phreatophytic shrub species (e.g. ATTO) is not allowed to occur if the conversion can be attributed to groundwater or surface water management practices. It is possible that a better understanding of ATTO population dynamics would allow managers to prevent ATTO invasion and avoid subsequent conversion of meadow sites to scrub sites.
Existing Monitoring Data
Monitoring for shrub recruitment at the permanent monitoring sites as well as other monitoring activities have resulted in a substantial data base for ATTO and other important Owens Valley species. Since 1991, data on shrub recruitment have been collected at all permanent monitoring sites in a 1m belt running the length of and adjacent to the 100m permanent transect. Within these belts, new recruits have been identified and notes were taken on each recruit=s location and size. In subsequent years, these recruits were located, if possible, and measured. Following eight years of this data collection, a data base has been amassed for each site from which survival and mortality can be estimated. In addition, a long term monitoring plot of young ATTO shrubs was implemented in 1992. Here, all shrubs within a 250m2 area were tagged, and their fates have been tracked annually. These data provide opportunities to quantify proportions of individuals within each stage and how they change from one stage to another over subsequent growing seasons.
Previous analysis of ATTO recruitment data revealed trends that could serve as the basis for designing scenarios to test with the model. ATTO exhibits prolific germination under certain precipitation conditions and a fair number of seedlings persist at sites where soil water conditions are adequate. The data also show that ATTO recruits may reach reproductive maturity within a few years. ATTO individuals may grow excessively in wet years, and although they will die back in dry years, the larger mature individuals persist (Manning 1994; 1995a; 1995b). A pattern observed but not yet explored in the line point transect data is the apparent increase in ATTO in a step-wise manner, as opposed to a gradual increase. Finally, because virtual monocultures of ATTO exist in the Owens Valley, it is possible that within an unknown time period, a meadow could convert to ATTO scrub. One hypothesis states that once near total cover of ATTO at a site has been attained, the site remains at a successional dead end, with no native perennial species moving in to replace ATTO (Manning 1997; 1999).
Stage-Based Demographic Modeling
Stage-based demographic models are accepted tools for estimating population parameters such as size and growth rate under different management and/or environmental conditions (Elzinga et al. 1998). This approach to modeling is appropriate for plants because their demographic parameters are usually determined more by size than by age (Schemske et al. 1994). In stage-based models, the members of the population of interest are tallied according to stages typically defined in terms of size and/or age. Field data are used to calculate the percent of individuals proceeding from one stage to another over a given time period. The probabilities that individuals would transition from one to another stage are then calculated. In this manner, a transition matrix, which is the core of the model, is developed (see Figure 1).
Figure 1. A life cycle diagram for a hypothetical species and the structure of its resulting transition matrix. Arrows represent possible transitions and these are identified in the transition matrix by an Ax.@ Those that are biologically impossible are assigned Ao@ in the matrix. (From Elzinga et al. 1998)
From the matrix, important population parameters can be calculated. The most important parameter is the finite rate of population increase, which estimates whether a population is increasing, and if so, how rapidly. The transition matrix can also be used to derive a sensitivity (or elasticity) matrix. The values in the sensitivity matrix indicate the importance of the various transition probabilities to the overall population growth rate. For example, a high elasticity value for a given transition indicates that there will be a greater change in the finite rate of population increase if the transition probability changes.
The model is capable of simulating the population=s response to hypothetical events by assuming certain environmental conditions (e.g. precipitation events) and/or by varying the transition probabilities based on assumed reactions to different perturbations (Ferson 1994). In this way, predictions may be made about an ATTO population structure at some temporal endpoint. AExperiments@ which if performed in the field could be costly and destructive, could then be done by simulation.
Commercial software is available for developing demographic models. The most appropriate package for plant demography studies, however, is RAMAS/stage (Elzinga et al. 1998). RAMAS/stage can be supplied with known or assumed environmental variables to project the behavior of a population. It can be used to estimate the likelihood of extinction or population explosion (Ferson 1994).
Procedure
Build prototype
Sufficient monitoring data exist to derive basic demographic information for ATTO, and from these, an initial model, or prototype, would be constructed. The data would be searched for natural groupings (stages) such as seedlings, immature shrubs, reproductive adults, etc. using histograms or other pattern recognition tools. In addition, other published discussions of stage definitions (e.g. Vandemeer 1978) will be consulted. Once stages have been identified and defined, observations of every plant for every year at the monitoring sites would be assigned to stages. The number of plants in each stage in each year would be tabulated, and all the possible transition paths identified. Then, the number of plants which followed each path over each monitoring time interval (year) would be calculated. From these tabulations, a matrix of probabilities (transition matrix) would be computed for each site for each year.
Refine Prototype
One of the limitations frequently encountered in stage-based modeling is that transition probabilities may vary from year to year, and often only a few years= data are available. However, because the existing data set covers many sites and several years, transition matrices can be generated for each site and time interval. This wealth of data offers many options for model refinement. First, known environmental data would be used with these matrices to project ATTO=s population status following a time period for which data exist, and the projections would be compared with the actual data. Predicting the known past is a standard way to test, calibrate and refine the models. The presence of eight years= data makes this exercise particularly useful. Comparison of annual transition matrices may show that certain sites or years represent extreme values; this may enable the overall ATTO response to be bracketed. Results may also reveal that matrices for similar sites could be combined (averaged) to generate a more comprehensive view of ATTO population change at sites with similar vegetation or management histories. Analysis of variation in the transition matrices themselves, particularly in relation to known environmental conditions such as precipitation, temperature, and water table depth, provides an opportunity for significant insight into ATTO recruitment dynamics and a range of options for refining the models.
Scenario Modeling
When the prototype models (transition matrices) have been thoroughly analyzed and refined models have been created, they can be applied to different scenarios with regard to management activities and possible climatic patterns. Effects of different ATTO control treatments, for example, could be modeled, as could effects of different patterns of precipitation and water table manipulation. The models could show probabilities of conversion of vegetation type under different assumptions.
Implementation
It is desired to develop in-house expertise in the theory, application, and limitations of demographic models, because this approach could prove useful in ongoing monitoring and analysis of various Owens Valley plant species= population trends. It is proposed that an experienced demographer be retained for consultation. Appropriate software (RAMAS/stage) would need to be purchased.
With advice as needed from the consultant, and with data tabulation and spreadsheet assistance from research assistant staff, the PIs would develop the prototype, then field test it. The consultant would review the preliminary analysis and suggest strategies for deriving the most possible useful information from the ATTO data sets. Results of the prototype development and testing would be reported to the Technical Group. The next step would be to refine the model(s) and perform scenario modeling, as described above. Again, the expert would be consulted on an as-needed basis. When a useful, defensible, and tested approach to demographic modeling of ATTO is achieved, a written report describing methods and results will be presented to the Technical Group for review. Results would also be prepared in manuscript form and sent to a refereed journal for peer review and possible publication.
Schedule:
| Activity |
2000 summer |
2000 fall |
2001 winter |
2001 spring |
2001 summer |
2001 fall |
2002 winter |
seek and hire consultant, collect Y2K field data, and build spreadsheets |
X |
X |
|
|
|
|
|
| build prototypes | X |
|
|
|
|
||
| status report to Technical Group | X |
|
|
|
|
||
| refine model and run simulations | X | X |
X |
|
|||
| prepare Technical Group report | X | ||||||
finalize report including Technical Group comments |
X |
Products:
This project will result in the development of a demographic modeling tools for ATTO in the Owens Valley. Models will be used to evaluate different ATTO management strategies to identify the most cost-effective techniques and thus help realize the goals of the long-term water agreement. Refined model results will be presented to the Technical Group in a written report. Depending on the utility of the final models, existing monitoring data for other species, both rare and common, could be used to develop similar demographic models. Population models for individual species could be an essential step toward developing a valley-wide model of vegetation dynamics, because they improve our understanding of the complex responses of the various species to environmental influences.
References:
Elzinga, Caryl L., Daniel W. Salzer, and John W. Willoughby. 1998. Measuring and monitoring plant populations. BLM Technical Reference 1730-1. 477 p.
Ferson, Scott. 1994. RAMAS/stage: Generalized stage-based modeling for population dynamics. Applied Biomathematics. Setauket, NY. 108 p.
Manning, Sally. 1994. Shrub recruitment 1991-1993: Results from permanent monitoring sites; update on Mazourka Canyon Road Atriplex torreyi; recommendations for future recruitment monitoring. Inyo/LA Technical Group report.
Manning, Sally. 1995a. Shrub recruitment 1994: A report to the Inyo/LA Technical Group on results of monitoring for recruitment at permanent monitoring sites. Inyo/LA Technical Group report.
Manning, Sally. 1995b. Mazourka Canyon Road recruitment site: Report on findings 1992-1994. Inyo/LA Technical Group report.
Manning, Sara J. 1997. Plant communities of LADWP land in the Owens Valley: An exploratory analysis of baseline conditions. Inyo County Water Department report. 160 p.
Manning, Sara J. 1999. The effects of water table decline on groundwater-dependent Great Basin plant communities in the Owens Valley, California. In: McArthur, E. Durant; Ostler, W. Kent; Wambolt, Carl L. comps. Proceedings: Shrubland ecotones. 1998 August 12-14, Ephraim, UT. Proc. RMRS-P-11. Ogden, UT. U. S. Department of Agriculture, Forest Service, Rocky Mountain Research Station.
Schemske D.W., B.C. Husband, M. H. Ruckelshaus, C. Goodwillie, I. M. Parker, and J.G. Bishop. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75(3): 584-606.
Vandermeer, J. H. 1975. Choosing category size in a stage projection matrix. Oecologia 32: 79-84.