Discussion Session 4: How can we best use natural gradients (latitudinal, environmental) to study the impact of climate stressors on individuals and communities?

Chair: Piero Calosi. Rapporteur: Sophie McCoy.

The comparison of phenotypes of species, populations or strains living under different environmental conditions (e.g. tropical vs. temperate environments) has been historically used to inform our understanding of how adaptation through natural selection has shaped biodiversity patterns as we know them today, as well as how differences in the physiology, life-history, development, morphology and behaviour of extant organisms has come to be (Darwin 1859, Fox 1936). These comparative studies have often been conducted not only in divergent habitats (i.e., environmental extremes) but have also exploited the fact that different taxa are often found living along environmental or latitudinal clines (Spicer & Gaston 1999, Chown et al. 2004) and attribute the observed phenotypic variation to organisms’ capacity for adaptation to different abiotic (but also biotic) conditions, effectively exchanging space for time. This approach is defined as the Synchronous Approach (see Kinnison & Hendry 2001, Reusch and Wood 2007).

During the last century, valuable information underlying our understanding of communities and assemblages ecological evolution (e.g. Kellerman et al. 2009, Calosi et al 2010), mechanisms for physiological and molecular adaptation (Somero 2010 - Fig. 1,2), and sensitivity to the global change (Stillman 2003, Deutsch et al. 2008, Dillon et al. 2010, Sunday et al. 2012) has come from macroecological and macrophysiological studies (Gaston et al 2009, Bozinovic et al 2011). These investigations have largely employed a correlative approach (Gaston et al. 2009, Bozinovic et al. 2011), which does not come without limitations (e.g. collinearity, spatial autocorrelation). To overcome the issues related to correlative approaches, a variety of analytical solutions (including carrying out model selection exercises – for example Calosi et al. 2010, Byrne et al. 2013) are employed. Furthermore, the increased utilization of phylogenetically corrected analyses has allowed the evolutionary history of the taxa investigated to be taken into account (e.g. Garland et al. 1992, 2005, Calosi et al. 2010, Byrne et al. 2013). On the other hand, transplant experiments and laboratory natural selection experiments have been used to validate evidence from study conducted using taxa from environmental gradients (Huey an Rosenzweig 2009).

All of these approaches are complementary, and can all be used in isolation or in combination to validate observed patterns in nature. What they cannot do is to substitute the investigation of adaptation to natural gradients, given that a great deal of difference exists between ‘adaptation to laboratory controlled conditions’ and ‘adaptation to a natural complex environment’ (see Huey and Rosenzweig 2009), thus including indirect effects (e.g. interactions) (Bibby et al. 2007, Dashfield et al. 2008, Godbold and Solan 2009, Munday et al. 2009, Briffa et al. 2012, Kroeker et al. 2012, Sunday et al. 2012, Godbold and Calosi 2013, Laverock et al. 2013, Russell et al. 2013).

Other intrinsic limitations of comparing taxa (and communities) living along natural gradients are that: i) the amount of time that different taxa have been separated is generally unknown, although genetic studies may help overcome (at least in part) this issue, ii) often gradients (in particular latitudinal gradients) are habitat mosaics rather than a linear succession of habitats (see Helmuth et al. 2005). All in all, this approach can provide valuable indications on the relative vulnerability, levels of phenotypic plasticity (and thus phenotypic buffering ability) and even potential for further adaptation in target species, populations or strains.

Changes in biotic systems in correspondence with changes in environmental conditions can also be investigated along temporal gradients (the Asynchronous Approach, see Reusch and Wood 2007). One may make use of existing time series (Southward et al. 1998, Mieszkowska et al. 2007, Wootton et al. 2008, Wootton 2010), investigate the fossil records (Roy et al. 2001, 2009, Belanger et al. 2012) and revive resistance state to characterize the responses of past genotypes/phenotypes (such as diatoms cysts, Härnström et al. 2011). Regarding the utilization of the fossil records, as far we acknowledge the limitation of working at different time scales than the characterization of extant genotypes/phenotypes, it can still be considered a valuable tool. The issue of scaling applies also to models, which cannot be extrapolated outside the spatial/temporal scale of the observation and experiment data used to parameterize them.

The need for a (re)new(ed) synthesis

The investigation of taxa living under different environmental conditions must however transcend the artificial barriers we have created between Ecology and Physiology, as well as between ‘single species functional studies’ and ‘community studies’ (see for example Gaston et al. 2009, Bozinovic et al. 2011).  A new synthesis among the fields of Macroecology, Macrophysiology, Macro- and Micro-Evolution is much needed, if we are to understand within the context of climate change:

  • the consequences of local adaptation,
  • potential shifts in communities functional responses (and thus changes in ecosystem services),
  • emerging properties (surprises!) only observable at a macro-scale (space and time) level (Chown et al. 2004).

 

This information is essential to help predicting the likely responses of complex global changes on extant marine diversity, and provide information to stake holders and policy makers needing to set management guidelines for natural resources conservation and sustainable exploitation.

A synthesis among Ecology, Physiology and Evolution can help us address questions of both specific and general nature, while always placing them within a stronger theoretical context. It can, in fact, exploit both the predictive value of functional ecosystem responses, as well as provide a mechanistic interpretation of the observed patterns (i.e., it is important to know where we are going (patterns), as well as to know how we are getting there (mechanisms)), whilst making no major assumptions. Large-scale studies can be used, for example, in defining the relative importance of ocean acidification when it occurs (as it will, see IPCC 2013) in combination with other stressors (e.g. warming and de-oxygenation) using model selection exercises (see also above) or pathway analyses. In addition, even just within the context of ocean acidification investigated in isolation, model selection can help identify whether different populations, species, or communities living in different oceanic (climatic) regions are sensitive to different drivers associated with ocean acidification (i.e., pCO2, pH, carbonate and aragonite saturation, see for example Byrne et al. 2013). These exercises of parameter estimation will be extremely useful for a more accurate parameterization of mathematical models.

Future directions and urgent issues

Whilst to date we possess a relatively good understanding of the relationship between temperature and taxa diversity, we cannot assume these trends hold true for other drivers. There is, therefore, a need to specifically investigate taxa sensitivity to ocean acidification along latitudinal and environmental gradients (but only to pCO2/pH), as the existing adaptation to a given factor gradient (e.g. temperature) could influence populations and species sensitivity to varying pCO2/pH levels. We have so far very few examples of studies which show us the ‘shape’ of biotic systems’ plastic and adaptive responses to elevated pCO2 / low pH (e.g. Maas et al. 2012, Miller et al. 2012, Calosi et al. 2013). Thus we identify the need to identify gradients (in pH, as well as in pH variability) which can be used to investigate the issues listed above. This can definitively include: i) shallow and deeper waters, ii) CO2 vents (Fig. 3), iii) upwelling areas, and iv) along bathymetric gradients. Eventually gradients can be ‘created’ in the field with current technological advances (e.g. the utilization of gas/temperature manipulation experimental systems, such as FOCE, e.g. CP-FOCE Kline et al. 2013).

Our workshop has also identified a number of issues which need to be urgently addressed:

  • we need to better explore existing observational time series, as well as explore the idea to implement new ones, purposely for the investigation on ongoing global changes and ocean acidification in particular
  • integrate field observations on marine communities along natural gradients to field and laboratory manipulation experiments conducted using representative organisms and assemblages from along such gradients
  • map our existing knowledge in environmental conditions and biological responses to identify: i) habitats and regions currently underrepresented, and ii) the ‘ideal’ gradients we should exploit to test specific questions (e.g. on pH levels, pH variability levels, saturation status levels, etc)
  • firmly move away from the ‘two species’, ‘two locations’, ‘two regions’ studies, as whilst more onerous, a good level of replication (6-10 points in space or time) is possible and in any case necessary to be achieved to obtain reliable data;
  • collect environmental and genetic data for a better interpretation of the results obtained (i.e., we may be dealing with an environmental and/or genetic mosaic)
  • investigate further the hypothesis that biotic communities from more variable systems are going to be more resilient
  • contextualize studies within species and communities capacity for acclimatisation and adaptation, so far largely ignored within the context of ocean acidification (but see for example Pistevos et al. 2010, Sunday et al. 2010, Lohbeck et al. 2012, Miller et al. 2012, Calosi et al. 2013, Dupont et al. 2013, Benner et al. 2013, Tatters et al. 2013)
  • need to create both regional and international level teams/communities of scientists dedicated to answer the point above.

 

DEFRA Logo NERC Logo DECC Logo LWEC Logo