Ecosystems of Climate Services in Latin America: examples from Guatemala and Chile

Climate services ecosystems are dynamic complex network of institutions, agents, information, knowledge, products and services functioning as a unit, at any (or across multiple) spatial or temporal scale(s), with the objective of supporting decision-making to enhance the resilience to a changing climate, and support countries and institutions to achieve their adaptation and mitigation goals while optimising available resources (González Romero et al., BAMS, sub-judice). 

Similarly to natural ecosystems, climate services ecosystems are self-contained and self-adjusting, having the ability to adapt in response to changes in the network or system (MilleniumEcosystemAssessment, 2003). Ideally, the more interconnected and interdependent the elements of the climate service ecosystem are, the higher the value and resilience of its network (e.g. Sawai 2013; Watts and Strogatz 1998).

Here, ecosystems of climate services in Guatemala and Chile are analysed and contrasted, using the González Romero et al (BAMS, sub-judice) approach, based on Dynamic Causal (Bayesian) Network Theory. Bayesian Network Analysis here aims at infering probabilities under changing conditions, like changes induced by programs, policies, budgets, institutional changes or any other external interventions. The causal assumptions that can be inferred from Bayesian Network Analysis identify relationships that remain invariant when external conditions change, allowing for the assessment of these changes, predictions of plausible scenarios and evaluation of counter-factuals and testable scenarios (e.g. Pearl, 2009). The assessment of causality derived from a Bayesian Network Analysis implies that the influence of one event onto another is stable and autonomous, so the change in one of them would necessarily result on a change in the linked event.

This research also involves an assessment of how optimal each network of climate services is, including potential improvements, and also an intercomparison -whenever possible- between the two sample ecosystems.