This paper synthesizes important elements from case studies presented in its companion paper (Quigley et al. in Environ Syst Decis, 2019, https://doi.org/10.1007/s10669-019-09728-0) to define mutual and distinct characteristics, and to develop a more holistic understanding of how earth science was used to support diverse examples of decision-making. We identify a suite of 28 different science actions used within the case studies that are classified as pertaining to (i) evidence acquisition and analysis, (ii) provision of science to target audience, or (iii) enhancing future science provision and utility. Sample action pathways provide empirically evidenced, albeit simplified, examples of how scientists may contribute to the progression of science through complex decision-making frameworks. Decision trees with multiple scientific and non-scientific inputs are presented based on empirical evidence and theory to provide scientists and decision-makers with simplified examples of complex multi-step decision-making processes under conditions of risk and uncertainty. Evidence for nonlinear engagement between decision-makers and science providers is presented, including non-traditional approaches such as provision of unsolicited science through the media and stakeholders. Examples of scientifically informed, precautionary decision-making with adaptive capacity, even where economically favourable decision alternatives exist, are provided. We undertake a self-elicitation exercise of case studies to derive values and uncertainties for % scientific agreement amongst utilized inputs and % uptake of potentially relevant and available science. We observe a tendency towards increased scientific uptake with increasing scientific agreement, but this is not ubiquitous; politically affected decisions and/or complex multi-decision scenarios under time pressure complicate this relationship. An increasing need for decision-making expediency that is not met by increased availability of relevant science evidence may rely on expert judgement, based on incomplete knowledge that is manifested as large uncertainties in defining a singular value for scientific agreement and uptake. We encourage scientists to further document their experiences using the science-action classification scheme provided herein to stimulate further comparative analyses of this nature.