Reader Comments

Post a new comment on this article

Reply to: Schwilk & Keeley (2012), "A plant distribution shift: temperature, drought or past disturbance?"

Posted by MGoulden on 26 Jul 2012 at 18:59 GMT

We have strong reservations about Schwilk & Keeley’s logic and conclusions, and feel their assertions need to be addressed.

We discuss our concerns in a pdf reply posted by PLOS. Please see:
http://www.plosone.org/at...

Additionally, we discuss further concerns in a paper in JGR-Biogeosciences. Please see:
Fellows, A. W., and M. L. Goulden. Rapid vegetation redistribution in Southern California during the early 2000s drought. J. Geophys. Res., doi:10.1029/2012JG002044, in press.
http://www.agu.org/journa...



The pdf reply posted by PLOS focuses on four issues:

(1) S&K’s field observations focused on stem density observations (# stems/area) whereas K&G investigated plant cover (crown area/area). S&K assume recent trends in C. greggii plant stem density are directly equivalent to plant cover. This assumption ignores the basis of ecological self-thinning, where stem density and cover are decoupled as mean plant size increases.

(2) S&K provide one figure that can be directly compared with K&G’s cover measurements. S&K Fig. 2B reports the altitudinal patterns of live and dead plant cover. This figure both confirms K&G's core observation for C. greggii, and contradicts S&K's causal interpretation.

(3) S&K's hypothesized patterns of postfire recovery dynamics are contradicted by previous work on C. greggii (Zammit & Zedler 1993; Z&Z).

(4) S&K’s attribution of K&G’s core observation to a complex mosaic of fire histories is highly speculative: it requires an improbable series of hypothesized past events. S&K provide no observational basis for extrapolating their observations on one species and a 489-m elevation range to the 10 species and 2,314-m range investigated by K&G.

We welcome follow-up data collection of the type conducted by S&K, which may shed light on our research findings and cause us to revise our interpretation. We also welcome commentary and constructive criticism, especially if it is carried out as an open dialogue. However we have serious concerns about S&K's logic and conclusions (e.g., issues 1 and 4), as well as their selective consideration of past and current evidence (issues 2 and 3).



If needed, please cite this reply as: Goulden ML, Kelly AE (2012) Reply to: Schwilk & Keeley (2012), "A plant distribution shift: temperature, drought or past disturbance?". http://www.plosone.org/ar...

No competing interests declared.

Reply to Goulden and Kelly's reply

firepulaski replied to MGoulden on 16 Aug 2012 at 20:28 GMT

Goulden & Kelly (G&K)'s response is pure smoke and mirrors. The primary problem is that they lack an adequate understanding of the fire-prone ecosystems they are studying. In our report we did not 'hypothesize' a complex fire history, rather we demonstrated it. In these crown fire ecosystems ages of live obligate seeding shrubs such as Ceanothus greggii are perhaps the only reliable means of determining timing of the last fire. It reflects stand age for this entire community and thus is relevant to all of the species reported on Kelly & Goulden (2008) (K&G). Interpreting K&G's original data as evidence of climate-induced change is dependent on the assumption of similar fire histories along the elevational gradient. That assumption is clearly wrong. Our study of thinning rates was presented as a demonstration of how fire alone could generate patterns similar to what K&G found but regardless, the basis of their study lacks a proper foundation to conclude as some have that they demonstrated climate-induced changes in plant distribution.

There are two issues, 1) we demonstrated different fire histories based on live stems --- this applies to the communities and to all the species in the communities and calls into question any interpretation of climate induced change, and 2) we attempted to show how fire might produce patterns K&G report. However, regardless of whether or not they agree with the second point, they cannot dispute the fact that climate change induced effects are dependent on similar fire histories along the gradient.

G&K's response focused only on the second issue of the possible self-thinning mechanism for an apparent shift in one species. To respond in more detail to the points they raise:

1) We measured both cover and density. Yes, they are not equivalent, but are highly correlated in this species. We measured density in addition to cover because doing so provides additional information: with density, we can determine population sizes associated with dead stems (initial cohort sizes) and therefore determine past per capita rates of mortality. Mortality is a discrete event in time and therefore provides important information: knowing when a plant died can provide some indication of why it died. Our time series goes back much further than the initial Zabriskie's 1979 data that K+G use, so is therefore informative as to whether any differences in mortality rate amongst sites is a response to "recent climate change." The pattern we report is not consistent with recent change because site-specific differences go back to stand establishment, and the mortality time series shows a pattern consistent with self-thinning. This is consistent with the influence of past disturbance history in determining community composition in fire-prone ecosystems.

2) No, the Fig 2 (Schwilk and Keeley 2012) shows that different elevations differ in both initial density post-fire and in total mortality since stand establishment, but neither pattern is monotonic with elevation (G+K lump elevations together in their response figure 2). Yes, there are differences in background mortality rate across elevations: some sites are more favorable and the elevations above 1585 m seem most favorable. But there is no evidence of a recent divergence in mortality patterns that would be consistent with recent climate change. Mortality patterns over time within a site are most strongly related to stand age.

Our responses above cover G&K's points 3 and 4 as well.

We suggest that in the future climate change scientists make a greater effort to collaborate with ecologists knowledgeable about the ecosystems they are studying, particularly in fire-prone landscapes.

Dylan Schwilk and Jon Keeley

Competing interests declared: We are authors of the 2012 PLoS One article to which G+K refer.