Various conceptual and modeling approaches address relevant questions, for instance, one in thermal biology on how thermal reaction norms (performance curves) of organisms indicate limits to thermal tolerance (Angilletta 2009). However, these approaches usually do not identify the underlying physiological and biochemical mechanisms that are operative at various levels of biological organization. Recent efforts to understand climate sensitivity of marine ectotherms in the field led to the development of the concept of oxygen and capacity limited thermal tolerance (OCLTT) as a complex framework linking such levels from ecosystem to organism, tissue cell and molecular effects, and the effects of various drivers (Pörtner 2002, 2010, 2012). Essentially, the OCLTT concept aims to identify the mechanisms setting the shape and positioning of thermal reaction norms of species and their life stages on the temperature scale. The concept focuses on temperature as the key driving force in climate change impacts on biota, through temperature means, extremes, its changing variability as well as its interactions with other drivers. Here, we emphasize its ability to integrate thermal responses across levels of biological organization.

Including other drivers thus means assessing how and why such reaction norms respond to their effect. In the terrestrial realm, temperature interacts with water availability, atmospheric CO ₂, and nutrient levels. Ongoing distribution shifts of organisms are largely determined by the temperature trends in both the ocean (Poloczanska et al. 2013) and terrestrial climate; however, the capacity of organisms to move or restrain geographical barriers may restrain organisms to follow moving climate zones (Settele et al. 2014). Marine paleo-records suggest strong interference of temperature-induced changes with other drivers such as expanding hypoxia or ocean acidification (Pörtner et al. 2005). These effects combine with those of changing ocean currents and enhanced stratification regimes of warming oceans, which cause changes in nutrient availability and thus productivity of large ocean areas and gyres.

In addition to identifying the specific vulnerabilities and their variability within and between species, a key issue in understanding ecosystem shifts involves addressing how biotic interactions are modulated by climate, for example, through shifts along temperature-dependent performance curves of interacting organisms (Pörtner and Farrell 2008; Storch et al. 2014). As interactions involve organisms across all domains an overarching understanding of reaction norms is needed. It may build on the principle hypothesis that sublethal tolerance limits to temperature (influenced by other stressors) are set at the level of biological organization with the highest complexity. The associated limiting process(es) then involve(s) coordination of the largest number of body and cell compartments. In animals, the OCLTT concept proposes that the functional capacity of systems supplying and using oxygen sustains the aerobic performance capacity of the organism (Pörtner 2002; Pörtner and Farrell 2008) and becomes limiting at high temperature extremes. This hypothesis is supported by laboratory and field observations of, for example, declining growth and abundance in benthic eelpout, due to oxygen-dependent capacity limitations under extreme summer temperatures (Pörtner and Knust 2007). Declining exercise performance in migrating salmon sensitive to warming (Eliason et al. 2011) also reflects a key limiting role of cardiocirculatory capacity in covering demand.

Thermal ranges are principally wider in isolated molecules, organelles, or cells than in the intact organism. Molecular or organellar functions thus become thermally constrained earlier when integrated into the organism. The latter provides regulatory feedback on the use and expression of genome, transcriptome, and proteome such that these patterns follow the shape of the thermal reaction norm (Windisch et al. 2014), reflecting the mechanistic integration of molecular into whole organism functioning. Functional scope of the organism becomes constrained by limiting biotic interactions (e.g., limited food availability, competition, predatory pressure), which demand energy and reduce the energy excess supporting other life-sustaining performances and thus fitness. Such effects constrain the fundamental niche to the realized niche at ecosystem level and shrink the temperature-dependent geographical distribution of multicellular organisms such as animals and plants (Pörtner et al. 2010). These links and interdepencies across organizational levels, accessible through OCLTT, are only just emerging and need further investigation (Pörtner 2012). We are not aware of alternative concepts equally powerful in integrating molecular to whole organism and ecosystem functioning under climate change. We also suggest that recent attempts to prove or disprove the OCLTT framework have not been successful when using reductionist approaches outside ecological or evolutionary context. OCLTT may be common in animals using a convective oxygen supply system (thus possibly excluding adult insects). It has been proposed that more "classic" ways of interpreting the results of reductionist experiments (e.g., by Clark et al. 2013; Gräns et al. 2014; and Wang et al. 2014) do not capture relevant aspects and contrast the more integrative, ecosystem-oriented, and evolutionary ways of interpretation in accordance with OCLTT (cf. Farrell 2013; Pörtner and Giomi 2013; Pörtner 2014). For example, the term "capacity" is used differently by the OCLTT sceptics than by those who apply OCLTT. The former asks whether the circulatory system if pushed to its limits is able to provide sufficient oxygen to the organism until lethal temperatures (Gräns et al. 2014; Wang et al. 2014). The latter emphasize the interdependence of capacity and cost (e.g., of mitochondria, or cardiocirculation) and therefore variable maintenance costs and associated functional scope. Such cost is relatively low within the optimum but becomes overproportionally high toward the edge of the thermal tolerance range, then constraining functional scope (the percent increment with warming is highest in low capacity systems such as in polar or winter stenotherms, e.g., Pörtner, 2002, Wittmann et al. 2008). In this way, the progressive increase in thermal limitation is captured from the earliest onset of constraints to lethal temperatures. OCLTT might be viewed as an early evolutionary principle in animals that has been modified in various climate zones (e.g., polar areas, Pörtner et al. 2012) or during the evolution of air breathing (Giomi et al. 2015). This emphasizes the need for the parallel development of theoretical and experimental approaches, especially as experiments cannot resolve all facets of complex phenomena. Current theories of evolutionary biology illustrate these requirements (e.g., Angilletta 2009).

Lower complexity levels in most unicellular organisms thus support thermal limits higher than in animals. While the mechanisms causing limitations remain largely unexplored in this group, the complexity hypothesis leads to specific testable predictions: In archaea and bacteria, highest complexity levels may be found in molecular and/or membrane complexes with, for example, metabolic functions; in unicellular eukarya (heterotrophs), earliest functional limits may arise during coordination of mitochondria and cytosol. Eukaryotic phytoplankton with maximum thermal limits similarly low as in animals, possibly experience earliest limits during additional coordination of chloroplasts with mitochondria and the cytosol. This high cellular complexity may be an overarching constraint setting heat tolerance in plants (Pörtner 2002). Within each domain, however, individual species and life stages display differential thermal ranges reflecting specialization on temperature regimes, habitat characteristics, and mode of life. Heat limits can be modified by acclimation or adaptation at temperatures below domain-specific limits, but acclimation or evolutionary adaptation cannot overcome the specific limits of the domain as complexity cannot be changed. The OCLTT concept can serve as a role model for studying these principles across organism domains as needed for comprehensively understanding ecosystem level changes. We also suggest that OCLTT may in the end and from an evolutionary point of view, be a key concept able to explain the limited thermal tolerance of metazoans compared to other organisms.

From the point of view that the complexity principle underlying thermal specialization holds for all organisms in an ecosystem and defines differential, species-specific sensitivities, such specialization will also shape biotic interactions of coexisting species. Additional stressors such as ocean acidification may affect performances of interacting species differently due to divergent physiological ranges, optima, and sensitivities. This will change their fitness in predator prey or competitive interactions. Changes in performance thus produce changes in behaviors and phenologies, and thereby the balance and synchronization of trophic levels and species community structure (Pörtner and Farrell 2008; Bozinovic et al. 2013b).

Further joint studies by physiologists and ecologists should thus look at species interactions across domains and climate zones, as well as in various, aquatic and terrestrial ecosystems. Overarching mechanistic frameworks, such as OCLTT for animals, need to be developed for all organisms and complemented, for example, under scenarios of warming, hypoxia, and acidification in the ocean, or under drought and changing temperature means, extremes, and variability on land.