Improving the community-temperature index as a climate change indicator
September
Improving the community-temperature index as a climate change indicator
Diana Bowler 0 1
Katrin BoÈ hning-Gaese 0 1
0 Senckenberg Biodiversity and Climate Research Centre , Frankfurt am Main, Germany , 2 Institute of Ecology, Evolution & Diversity, Goethe University Frankfurt , Frankfurt am Main , Germany
1 Editor: Jan Geert Hiddink, Bangor University , UNITED KINGDOM
Climate change indicators are tools to assess, visualize and communicate the impacts of climate change on species and communities. Indicators that can be applied to different taxa are particularly useful because they allow comparative analysis to identify which kinds of species are being more affected. A general prediction, supported by empirical data, is that the abundance of warm-adapted species should increase over time, relative to the cooladapted ones within communities, under increasing ambient temperatures. The community temperature index (CTI) is a community weighted mean of species' temperature preferences and has been used as an indicator to summarize this temporal shift. The CTI has the advantages of being a simple and generalizable indicator; however, a core problem is that temporal trends in the CTI may not only reflect changes in temperature. This is because species' temperature preferences often covary with other species attributes, and these other attributes may affect species response to other environmental drivers. Here, we propose a novel model-based approach that separates the effects of temperature preference from the effects of other species attributes on species' abundances and subsequently on the CTI. Using long-term population data of breeding birds in Denmark and demersal marine fish in the southeastern North Sea, we find differences in CTI trends with the original approach and our model-based approach, which may affect interpretation of climate change impacts. We suggest that our method can be used to test the robustness of CTI trends to the possible effects of other drivers of change, apart from climate change.
Introduction
Monitoring species response to climate change is an important challenge for ecologists and
conservation managers. Generalized approaches that can be applied to population census data
regardless of taxonomic groups are particularly useful as climate change indicators [
1
]. Such
indicators allow for standardized comparisons of how different taxonomic groups and
communities in different locations are responding. As ambient temperatures increase, the relative
performance of warm-adapted species is predicted to exceed the relative performance of
cooladapted ones [
2, 3
]. This community change can be summarized by the community
temperature index (CTI) [3], which is a community weighted meanÐa commonly applied community
summary variable [
4
]±of species' temperature preferences. This change in community
composition can be used as a generalized indicator of the realized impacts of temperature change on
communities [
5
].
ªWarmingº or ªthermophilizationº of biological communities has been already recognized
as one of the main fingerprints of climate change [
1, 6
]. This indicator is potentially more
sensitive than other climate change indicators because it is based on changes in species'
population abundances [
5, 7
] and may be detectable anywhere within species' ranges and not only at
range limits. Positive trends in the CTI over past decades have been observed in bird and
butterfly communities in Europe [
5, 8, 9
], birds in North America [
10
], and plants in Europe [
11
]
and the tropics [
12
], although not consistently across all populations that have been examined
[
13
]. Additionally, variation in the strength of CTI temporal trends among habitats have
suggested interactions between climate change and land use change [
6, 14, 15
].
The CTI has similar advantages to related trait-based approaches that use the differential
responses of species according to their traits to detect the impact of an environment driver.
For instance, declines of species known to be sensitive to nutrient-rich conditions have
indicated increased pollution. Such approaches have the advantage of naturally integrating the
components of environmental change, at the relevant spatial and temporal scales, that are
most important to the organism. Additionally, as a simple indicator to calculate and explain, it
has a value in the communication of climate change researchÐrepresenting a simple
ªbiothermometerº±making it especially popular among bird and butterfly conservation
organizations [
16
].
Despite these advantages, it has been recognized that interpretation of the CTI is not
straightforward [
17, 18
]. Species' temperature preferences often covary with other species
attributes, particularly habitat preference [19]. Because of such covariation, it can be difficult to
interpret the ecological significance of trends in the CTI, and specifically whether they are
really caused by c (...truncated)