Resources

• Variable timestep needed.
• Add belowground competition / competition by things smaller than a target plant.
• Long-term trends in soil nutrient dynamics. Currently nutrients would need to be specified as an externality - but someday we would want to simulate nutrient cycling.
• Trends in climate.
• "Memory" effects across multiple timesteps - like moving cations out of the system due to acid rain or soil compaction from harvesting.
• Ecological classification (site index) - so you can use different growth functions/conditions for variation in a site within a single run (if the climate is like that here, use this function; over there use another function).
• Site effects on reproduction - like climate-cued masting.
• One way to address model generalization - do cross-site comparisons - so similar types of forests (i.e. temperate hardwood) can start to share parameters - and you will get a sense of those parameters that absolutely must come from data.
• Switches from single resource to multiple resource limitation as a function of density - perhaps this would introduce the stochastic element needed.

Recruitment

• Spatial vs. non-spatial - there's some exploration of non-spatial recruitment - let the model tell us if spatial dispersal is important.
• If you have a non-spatial dispersal - would you get the wrong answer? For long-term co-existence it matters a lot, but we may be able to demonstrate that on shorter time-scales and for other questions that it doesn't matter (management).
• Animal behavior
• Masting
• Microclimate for germination and establishment. It makes a huge difference in any given season, but do we try to capture the long-term average instead? Use a stochastic driver to replicate these effects?
• Stochasticity in recruitment - the variable recruitment of rare species can be important for co-existence.
• It would be good to look at regeneration from a landscape perspective - managers could save a lot of money over trying to force the right species mix from the start if you could predict better the dynamics of natural seeding.
• It's worth thinking more about relative species abundances instead of long-term coexistence.
• Generalized substrate - general agreement that that would be useful.
• Epiphytic establishment.
• Small mammals.
  • Nested models for functional and numeric responses
  • Neighborhood effects
  • Control efforts
  • Parsing out the effects of invertebrates and alternate food sources - although the effects of seed availability probably swamp this and it doesn't need to be taken into account at first.
• Vegetative reproduction
• Effects of light
• Pathogens

Growth

Multiple resource limitation

• Stochasticity
• Parsimony (noise vs. signal) - when is light enough?
• R² of light: how high is high enough? Predictive power vs. explanation
• Water might be important

Life history stages - when do we hand off an individual from one set of functions to another? Especially seedling to sapling. Can we have just one function for the entire life cycle? So far not enough data. It might be worth investing more at the empirical level to get seedlings right.

Mortality

• Self-thinning
  • Crowding
  • Density dependence
  • Either we just empirically thin when it's needed, or we actually try to reproduce self-thinning in a mechanistic way
• Understanding thinning in a more general way - generalize self-thinning to a crowding effect
• Pathogens
• Senescence
• Adult tree crowding
• Relative size matters
• Resource dependent
• How do you decide which approach to use? Relying on crowding in growth to get mortality right vs. just "whacking" excessive trees that escape your mortality function with self-thinning

Herbivory

• Size matters
• Deer as externality but with functional responses
• Exotics vs natives - herbivore overabundance - natives with predators don't have the same effect
• Offtake rates
• Episodic outbreaks (pests)
• How to model random extreme events? Sometimes you let them get lost in the noise but some people are specifically interested in them - especially changes in frequency in these events.

Issues

• Initial conditions - chaos - are some sites more sensitive to initial conditions because the growth trade-offs are weaker?
• Iterative linking of other models to SORTIE - feeding each other input and output
• Extension to new sites - need to go through the painful process of parameterization for new sites? Are there some parameters that we could generalize, that could be separated out from those parameters which absolutely must be calculated? (PP networks)
• Spatial vs. non-spatial analyses
• Capturing microsite variability through stochasticity
• Now that we have so many different sites parameterized - is it time to do a synthesis paper to find similarities?
• Genetics / evolution - might genetic variability across space confound management approaches?
• Multiple stable states - have they ever been exhibited? We don't have enough drivers right now - but you might see it once you get enough stochasticity.
  • Stochasticity must not be in there for its own sake - it must be derived from the extremes around the mean
  • We underestimate the noise in the data
  • How do you deal with what you cannot explain? Do you put in a process that explains 20% of the variance and invoke stochasticity to explain the other 80%?

Collaboration

• It would be interesting to include water effects in the model to look at climate change
• Could we link the physiological model with the growth function of SORTIE? This project might be outside the SORTIE framework.
• Have to figure out how to handle collaboration so that groups don't compete to answer the same questions - need to coordinate
  • There will be natural competition - need to facilitate and turn it into complementary work
  • We could do more together than if everyone went their own separate ways
• SORTIE network
  • Has a list of proposed projects
  • The power structure must be managed - new researchers coming in can have ideas but the established people can take it and make it happen much faster
  • Want to avoid "land claims" - maybe only list projects you have funding for
  • List of researchers and their research interests so everyone can initiate collaboration
  • Communication forum important - web site, discussion groups
  • Software and tool repository, full user/science documentation
  • It's been made clear by NSF that something they are looking to fund is an educational component and the addressing of inequities of participants
  • Knowledge of new publications important
  • Workshops to show people how to really use the model correctly
• How do you share your work? If you make a behavior, you want to have a chance to use it before it's shared - so how do you share it? When do you make something available?
• SORTIE book? It's on the horizon. A meeting in Montreal in 2005 may take symposium form with book chapters. Let's do a symposium proposal for ESA.