The oxygen isotopic composition of a banded coral from the western equatorial Indian Ocean provides a 150-year-long history of the relation between the El Niiio-Southern Oscillation (ENSO) phenomenon and the Asian monsoon. Interannual cycles in the coral time series were found to correlate with Pacific coral and instrumental climate records, suggesting a consistent linkage across ocean basins, despite the changing frequency and amplitude of the ENSO. However, decadal variability that is characteristic of the monsoon system also dominates the coral record, which implies important interactions between tropical and midlatitude climate variability. One prominent manifestation of this interaction is the strong amplitude modulation of the quasi-biennial cycle.Compilations of sea surface temperature (SST) anomalies demonstrate that the ENS0 phenomenon is equally important in both the Indian and Pacific oceans, which implies a climate connection over much of Earth (I ). However, the different basinal geometry and the seasonally reversing monsoon currents demand a separate set of physical processes governing interannual temperature variability in the Indian Ocean (2), and, therefore, the degree of coupling between the ocean basins-a central issue for predictability-is not necessarily fixed. In fact, the tendency for this coupling to evolve may explain why the relation between the ENS0 and the Asian monsoon has been the subject of one of the longest standing debates in climatology (3).Previous analyses have variously concluded either that the Asian monsoon affects the ENS0 or that the converse is true (4). Each of these conclusions allows the possibility for different large-scale feedbacks [for example, the former view implies that Eurasian snow cover could influence tropical climate (31. Yet, resolution of these dynamical effects with the instrumental record alone is compromised by the fact that continuous reliable observations rarely span more than three decades-a period areuablv alreadv modified bv human activitgs. %us, it isdifficult to predict whether the relations determined from modem observations would apply in basic climate states different from those of recent years. Furthermore, the long records that do exist (for example, the India Monsoon Rainfall Index) may not be the most appropriate measures of the large-scale Asian monsoon phenomenon (3). Defining the interaction between the ENS0 and the Asian monsoon requires not only an extended record but also a globally relevant index.

Worldwide, policymakers, health system managers, practitioners and researchers struggle to use evidence to improve policy and practice. There is growing recognition that this challenge relates to the complex systems in which we work. The corresponding increase in complexity-related discourse remains primarily at a theoretical level. This paper moves the discussion to a practical level, proposing actions that can be taken to implement evidence successfully in complex systems. Key to success is working with, rather than trying to simplify or control, complexity. The integrated actions relate to co-producing knowledge, establishing shared goals and measures, enabling leadership, ensuring adequate resourcing, contributing to the science of knowledge-to-action, and communicating strategically.

Abstract.While instrumental and proxy-based climate records describe significant decadal-scale climate variability throughout the tropical Pacific, Indian, and Atlantic Oceans, the processes responsible for these

No abstract