# EML

bug fixes and replies to various input from community on EML package

Thoughts on understanding and using EML in research, including Dryad. See https://github.com/ropensci/EML/issues/82#issuecomment-33558301

# multiple-uncertainty

#### Jim and Mike meeting

- Why the steps in the optimal solution of the deterministic case?
- Irregularities in value function in first iteration: sometimes not exact ties; actually preferring to fish at almost 0 rather than 0.
- address range constraints on
`h`

and`x`

.

### Next steps

- Puzzle out irregular first iteration. Role of state equation,
- Interpolation-based inference of optimum, rather than grid
- When can we optimize over the \(X_{t+1}\) that maximizes the value and just back out the corresponding harvest?

In the stochastic case, multiple harvests could all realize the \(X_{t+1}\) that maximizes the value, though each with different probability. I do not believe that it holds in general that the h for which it is most likely to get the optimal \(X_{t+1}\) is necessarily the h which maximizes the value. Would you agree? (recall we want to maximize the expectation, so we need to integrate over all outcomes of the harvest, not consider only the most likely outcome). I see no alternative to solving for h directly then?

# NCEAS workshop prep

The agenda for the workshop, participant list, goals and intended outcomes can be found at https://projects.ecoinformatics.org/ecoinfo/projects/commdyn/wiki/First_Community_Workshop and is attached. In preparation for the workshop we would like for you to:

- If at all possible, identify and bring a dataset that includes community data for at least ten points in time at somewhat even intervals.

LTER searches; Bestelmeyer maybe

- Bring ideas for metrics that are meaningful measures of community change over time.

Early warning signals? Climate change vectors?

- Develop ecological research questions that may be answered with the metrics of community change that will be developed/used during the workshop. Ideally, these synthesis questions span several ecological systems.

Is this list in reverse sequence? e.g. should develop questions, then metrics, then identify datasets?

Desired Outcomes

- Repository design requirements for sharing workflows
- list of community metrics
- list of data sets
- list of data transformation operations to get from raw data to analyses
- list of visualizations
- research collaborations: Ideas for ecological questions that can be answered with these data and metrics

# Reading

*Nature*on troubled times for the monitoring buoys used by the TAO early warning system for El Nino events 10.1038/505585a- Nice pieces on reproducibility in Nature 10.1038/505612a and Science 10.1126/science.1250475 recently.