EcoEssay Series Number 2 - Response

Ellison (1997) argues for a transformation of the "culture of ecology" to one in which we see ourselves as a "community working towards a common understanding of the .. biosphere". One benefit of this collective transformation would be increased access to data: Ellison's figurative tapestry. Why? Because a community shares it's resources, and, the key resource of ecologists is our data.

I agree with Ellison. Lack of data access does hamper the growth of our discipline. In consequence, I share my data with ecologists at NCEAS and elsewhere, and I often publish my raw data if a journal's policies permit. However, I do not share my data with everyone, nor do I always publish my raw data. It's these exceptions to Ellison's recommended practice that are my theme here.

I and my Dorset colleagues have generated 20 years of data on a dozen lakes in Ontario. I share my data with other zooplankton ecologists only when: (1) my data can help answer their specific questions, (2) they consult with me about the strengths, limitations and past uses of my data, (3) their proposed research won't exactly duplicate my current work or work proposed for the very near future, and (4) their proposed research won't closely approximate the research and data sources of one of my current graduate students.

The common thread of these restrictions is the requirement for comunication between those who generate and those who want access to the data. This is not a frivolous requirement, rather it protects the "tapestry", reducing the risk of blunders, and unplanned duplications in research. However, it does restrict data access to those with the wit to ask. There are situations where I believe ecologists should publish their data and accompanying metafiles, without specific requests from others. These situations include:

1. when such publication was a condition of financial support for the research,
2. when retirement looms, increasing the risk of effective data loss,
3. when there are any other risks of permanent data loss,
4. if delays in publishing place the environment at any risk,
5. when the major questions of the research have been answered, and
6. if delays in publishing the data are unreasonably long.

Most ecologists would agree with the first four criteria, but the last two should raise some eyebrows. Major questions change over time in long-term ecological studies. For example, I began my 20 years of work on Harp Lake with questions about trophic status, then switched to acidification, and most recently to exotic invaders (Freshwater Biol., 1997, v37:409). I think I should publish the data when the key questions in each research theme are answered, not when my work on Harp Lake is completed. It won't be completed for years yet.

Defining an "unreasonably long delay" of data publication (the last criterion) is not easy. Many ecologists consider a year or two to be reasonable, but longer delays to be less so. I think there is no single answer, because the required scale-dependence of the research must be considered. What if the key questions can't be answered in a few years? It took me more than a decade of data collection to document the pace of recovery of biota in Sudbury lakes after their acidity was neutralized. I didn't publish the raw data until I had answered the central questions, even though this took me almost 20 years? Was I wrong to delay? In this case, I think not?

Let's work towards Ellison's tapestry, but also give each weaver the time they need to finish their little ecological placemat. Some take a longer stint at the loom than others.