| 英文摘要: | Global models highlight that environmental change in marine ecosystems is caused by multiple stressors. Now a study puts these projections into a biogeographical framework suitable for integration with wider biological understanding and more robust impact assessment. Marine ecosystems supply irreplaceable economic and cultural value through coastal and maritime communities, livelihoods, tourism and more than 8% of global animal protein through capture alone1. Under human influence, these systems are undergoing changes in temperature, acidity, light, nutrients, and other environmental stressors with unknown impacts on living marine resources (LMRs). One of the greatest challenges in representing marine ecosystems is reconciling the vast experiential insight of observational oceanographers and marine biologists (on physiology, biodiversity and ecological function) with the idealized mathematical descriptions embedded in Earth System Models (ESMs). The diverse physiological and ecological sensitivities of LMRs and potential for threshold and multiplicative responses necessitate an integrative approach to robustly characterize the impacts of projected biogeochemical change on these resources. Writing in Nature Climate Change, Philip Boyd and colleagues2 report an end-to-end statistical approach for province-level grouping of changes in multiple ocean biogeochemical stressors and subsequent regional application of direct biological observations for more robust assessments of ecosystem change. Current ESMs project regionally intense environmental change across multiple stressors3, 4. The challenges in interpreting how such changes will impact LMRs are manifold — ranging from biogeochemical response characterization to establishment of the significance of such multivariate changes to LMRs. The biogeochemical challenges include representation of the mean state, internal variability, and anthropogenic response in a self-consistent mathematical representation of ocean biogeographical provinces5, which must account for features of circulation, seasonal stratification, ice cover, nitrate and iron limitation all characteristic of the real ocean. This diversity of interactions implies that different subsets of factors will have outsized or negligible influences depending on the oceanic regions and LMR, and that some model biases may lead to severe regional mischaracterizations. At their best, current generation ESMs represent biome-level ecosystem biogeochemistry rather than local species-level biodiversity. At their worst, they misrepresent major spatial or temporal features of biomes and project widely divergent regional changes3, 4. Thus, even as ESM representation of idealized diatoms or coccolithophores provides some insight into functional biogeochemical diversity, the representations do not suit most needs of LMR managers. While the quest to represent complex Earth system interactions and sensitivities with sufficient physical and biodiversity to meet the full spectrum of needs is critical and continuing6, careful interpretation of current generation model response in relation to the richness of biological understanding available in field and laboratory studies can also shed some light. This interdisciplinary work requires a kind of 'roadmap' to engage the full intellect of the marine biological community across the scope and capabilities of current generation ESMs, while quantitatively incorporating expert judgment to account for observational and model limitations. Boyd et al.2 take an innovative approach to build and use such a roadmap (Fig. 1), through application of a statistical tool (principle component analysis with factor rotation) to cluster biogeochemical impacts on 15 stressors into 6 global patterns and assess their relative contributions over 14 biogeographical provinces. This biome-level grouping of systems among the potential biogeochemical stressors facilitates comprehensive assessment of regional change. The fundamental advantage of this approach is in maximizing the information on ESM biome-level biogeochemical change in which the authors have more confidence, while minimizing reliance on model-dependent regional phytoplankton physiology and biodiversity representations in which they have less confidence. Boyd et al. then exercise this roadmap with a suite of observationally motivated interpretations of these biome-level changes through key phytoplankton taxa including coccolithophores, diatoms, and cyanobacteria. Their results are generally grim news for diatoms, but highlight a potentially growing niche for coccolithophores, despite decreasing rates of calcification.
|
