The system dynamics model described in this report is the result of a collaboration between U.S. Geological Survey (USGS) scientists and National Park Service (NPS) San Francisco Bay Area Network (SFAN) staff, whose goal was to develop a methodology to integrate inventory and monitoring data to better understand ecosystem dynamics and trends using salmon in Olema Creek, Marin County, California, as an example case. The SFAN began monitoring multiple life stages of coho salmon (Oncorhynchus kisutch) in Olema Creek during 2003 (Carlisle and others, 2013), building on previous monitoring of spawning fish and redds. They initiated water-quality and habitat monitoring, and had access to flow and weather data from other sources.

This system dynamics model of the freshwater portion of the coho salmon life cycle in Olema Creek integrated 8 years of existing monitoring data, literature values, and expert opinion to investigate potential factors limiting survival and production, identify data gaps, and improve monitoring and restoration prescriptions. A system dynamics model is particularly effective when (1) data are insufficient in time series length and/or measured parameters for a statistical or mechanistic model, and (2) the model must be easily accessible by users who are not modelers. These characteristics helped us meet the following overarching goals for this model:

Summarize and synthesize NPS monitoring data with data and information from other sources to describe factors and processes affecting freshwater survival of coho salmon in Olema Creek.

Provide a model that can be easily manipulated to experiment with alternative values of model parameters and novel scenarios of environmental drivers.

Although the model describes the ecological dynamics of Olema Creek, these dynamics are structurally similar to numerous other coastal streams along the California coast that also contain anadromous fish populations. The model developed for Olema can be used, at least as a starting point, for other watersheds. This report describes each of the model elements with sufficient detail to guide the primary target audience, the NPS resource specialist, to run the model, interpret the results, change the input data to explore hypotheses, and ultimately modify and improve the model. Running the model and interpreting the results does not require modeling expertise on the part of the user. Additional companion publications will highlight other aspects of the model, such as its development, the rationale behind the methodological approach, scenario testing, and discussions of its use.

System dynamics models consist of three basic elements: stocks, flows, and converters. Stocks are measurable quantities that can change over time, such as animal populations. Flows are any processes or conditions that change the quantity in a stock over time (Ford, 1999), are expressed in the model as a rate of change, and are diagrammed as arrows to or from stocks. Converters are processes or conditions that change the rate of flows. A converter is connected to a flow with an arrow indicating that it alters the rate of change. Anything that influences the rate of change (such as different environmental conditions, other external factors, or feedbacks from other stocks or flows) is modeled as a converter. For example, the number of fish in a population is appropriately modeled as a stock. Mortality is modeled as a flow because it is a rate of change over time used to determine the number of fish in the population. The density-dependent effect on mortality is modeled as a converter because it influences the rate of morality. Together, the flow and converter change the number, or stock, of juvenile coho. The instructions embedded in the stocks, flows, converters, and the sequence in which they are linked are processed by the simulation software with each completed sequence composing a model run. At each modeled time step within the model run, the stock counts will go up, down, or stay the same based on the modeled flows and the influence of converters on those flows.

The model includes a user-friendly interface to change model parameters, which allows park staff and others to conduct sensitivity analyses, incorporate future knowledge, and implement scenarios for various future conditions. The model structure incorporates place holders for relationships that we hypothesize are significant but data are currently lacking. Future climate scenarios project stream temperatures higher than any that have ever been recorded at Olema Creek. Exploring climate change impacts on coho survival is a high priority for park staff, therefore the model provides the user with the option to experiment with hypothesized effects and to incorporate effects based on future observations.