ideawork/Question_scoping.md

In diazrenata/BBStrends: What the Package Does (One Line, Title Case)

The intuition for this project comes strongly from the work using the BioTIME database.

There are three really relevant papers from that database.

One objective is to flesh out what we learn from that work (which looks at population trends, species richness, and total abundance in individuals (or biomass, but not both)). This includes additional currencies, energy and biomass, and population vs assemblage.

A second objective relates to trade-offs between the currencies. Relevant papers there are White et al 2004 (Tradeoffs through time) and 2009 (Changes in a tropical forest)

Additional currencies

Population vs assemblage

Trade-offs between currencies

Assemblage-level trends

For any given currency, I see 2 questions: - Random walk vs regulated - Trending vs no slope - Maybe these fit into a 2x2 table? - Nonzero slope + not stationary - Nonzero slope + regulated - Zero slope + not regulated - Zero slope + regulated

Population vs assemblage

• Tallying population slopes fails to account for relative abudnances. If rare species are stable but abundant species decline, the overwhelming signal is decline.
• The best tally of slopes is of population slopes.

Currency tradeoffs

Gleanings

Trends/slope

I'm inclined to focus first on the trend slope: In particular, do we see consistent declines in abundance, energy, or both?

Headlines from BioTIME studies might lead us to expect no slope for abundance, but those studies do not actually amount to an expectation of no slope. Species richness and populations have no consistent slopes, but the relative abundances of populations as well as colonizations and extinctions introduce wiggle room between tallying the slopes of individual populations and the overall trend. I find it curious that they have not published on the overall slope. More generalized concerns about defaunation and declines in population sizes would point towards an expectation for general declines.

Energy use may be a more direct metric of ecological function. Total energy use is often assumed to be fixed or not changing directionally over time, especially in theoretical contexts. However, shifts in resource availability, total abundance, and/or the size spectrum can induce shifts in total energy use that may or may not correspond to trends in total abundance. The implication from concerns over defaunation and size-biased extinctions again points towards expecting declines in total energy use.

These currencies may co-vary but they may trade off with each other. An important follow-up question to documenting the prevailing overall trends in the two currencies is to ask what are the prevailing relationships between trends in total abundance and total energy use? In particular, do we tend to see compensatory trade-offs between energy and abundance leading to constant total energy use (or pointing in that direction)? When one is stable, does the other tend to be stable?

Once we have a sense of the prevailing trends in aggregate, we can also begin to dig into what are the smaller-scale behaviors getting us there? When slopes are 0, is it because most of the populations are not changing, or is there a large amount of change cancelling itself out? When slopes are not zero, are most populations moving in a similar way, or are very abundant species overwhelming the dynamics we'd see if we treated populations equally regardless of abundance?

Regulation

I am less clear on what to take away from testing for stationarity and trend stationarity (and especially testing for these as interchangeable). As I understand it these tests show whether a variable tends to return to a mean or a trend following perturbation, or just keep drifting from wherever it lands - whether it is a drifty random walk or something with a systematic mean/trend.

This seems kind of abstractly interesting but not as directly relevant to questions about trade-offs or compensatory fluctuations?

Unless it's something like: a nonsignificant slope + a stationary time series is different in some way from a nonsignificant slope + nonstationarity? But I don't think that really tells you the difference between "not allowed to change" and "not changing systematically".

diazrenata/BBStrends documentation built on March 10, 2021, 11:23 p.m.