**6. Applying the model**

New ideas and innovations often struggle to gain traction particularly when benefits around them are obscure or disputed both rightly or wrongly. Cognisant of this fact, two examples illustrate plausible applications of the proposed model in both climate change mitigation and climate change adaptation. Before presenting the examples, a brief recap of some tenets of the biomimicry-policymaking nexus that informs the Mjimba-Sibanda dynamic policy model may be in order.

First, it is important to remember that biomimicry applies to three main typologies: mimicking form, process or system. Mimicking form entails emulating (not exact copying) a particular shape in order to achieve a particular function. Second, mimicking a system requires understanding of dynamics of complex interactions. In nature, an ecosystem illustrates such complex system. In natural ecosystems, the concept of waste does not apply because everything is either raw material or food for another and everything is recyclable in a closed resource loop. In such systems, there is co-evolution and co-development. The analogous human concept is the circular economy [38]. The Kalundborg Industrial Park in Denmark illustrates this concept through the co-location of complementary industries that exchange resources that include water, heat, gas, fertiliser and fly ash [39]. Here, we take the view of policy development as being analogous to co-evolving mutualistic relationships in a thriving ecosystem. In that regard, policy development has to seek greater benefits of systems and not components of systems. Let us now revert to the examples of two cases relating to climate change mitigation and adaptation.

Carbon sequestration is a plausible intervention of climate change mitigation. Forests such as the Amazon forests are significant and natural sequestration arenas. However, since the 1970s, over 18% of the Amazon rainforests have been destroyed mainly for agricultural, timber logging and mining, among other activities. This has removed a significant carbon sink. In addition, these forests provide habitat for about a quarter of the world's terrestrial species and account for about 15% of terrestrial photosynthesis, whose by-product is the oxygen that humans and other animals breathe. Part of climate change mitigation seeks to retard and halt the further destruction of these forests. AI can help understand the relationships among various parameters such as rainfall, humidity, wind, temperature and floods. In addition, AI can also project changes in the acreage of critical forests projecting how these changes affect ecosystem services such as carbon sequestration. The relative changes in these parameters can be used for policy modelling seeking to enact automatically more stringent policy and legislation prescripts that can reduce the rate of depletion and promote regeneration of natural forests. The optimisation of natural systems such as the hydrological cycle, natural runoff, percolation and evaporation rates could provide benchmarks for what ideal conditions policy may seek to foster. An advantage of applying the proposed model in such a scenario is that the policy relies on both current and predicted possible conditions based on current deforestation rates. In addition, the generation of stringent conditions relies less on human judgement with AI generating the interrelations among species and components and, most important, making specific policy adjustments seeking to halt or mitigate present and future hazards.

Regarding climate change adaptation, machine intelligence such as AI and machine learning can predict possible future scenarios including the timing of their manifestation. Big Data fed into machine learning could help predict, for example, the areas with the likelihood of coastal flooding associated with climate change-related sea level rises. A dynamic policy could prescribe future actions such as land rezoning to stop further construction or the introduction of new building codes in areas with highest exposure to natural hazards and high probability of such risks manifesting. Similarly, the policy could define insurance models and levels of disaster preparedness triggered automatically should sea levels reach specified thresholds. This would enable various groups such as residents, investors, insurance providers and emergency services such as the police, hospitals and disaster units to be better prepared.

At first glance, the two examples may appear far-fetched and impractical. However, the plethora of late but well-intended and sometimes incoherent policies in many disciplines suggest that a systemic and self-adjusting policy regime is ideal to deal with the dynamics and nuances of issues such climate change. This holds in majority of policies that relate to climate change adaptation. Traditionally, these have often appeared after adverse events have occurred, instead of manifesting pre-the event to minimise damage. Even in the mitigation drive, the various actions arguably react to events that are avoidable, that is, driving reforestation instead of avoiding deforestation. The Mjimba-Sibanda dynamic policy model seeks to avoid this by taking a proactive policy approach to managing climate change. The approach takes the 'evidence-based' policymaking position by using large amounts of data to change both timeously and appropriately the relevant policy standards, codes and other parameters. The advantages of such dynamic policies are available elsewhere. Although the advantages do not necessarily use machine learning and do not draw from biomimicry, they nevertheless apply in this argument.

One important example of dynamic policy is in Kenya. Following the disputed 2007 elections, the feuding parties in the country eventually agreed to a negotiated settlement that culminated in a Government of National Unity (GNU) that comprised representatives of the various political parties. The political parties designed reforms for a more democratic political dispensation. Due to the prevailing mistrust at the time, the parties agreed that the implementation of the new Constitution, which came into effect in 2010, would include self-executing mechanisms. For example, certain provisions that required the President to ratify Bills by a specific date were set to become the responsibility of the Chief Justice if the President did not act accordingly. Similarly, if the Chief Justice reneged in signing the same within a defined period, the said provisions could automatically become law. Compelled by these conditions, on 27 August 2011 the then President signed 15 out of 27 Bills that were to meet the 1-year deadline [40]. Elsewhere, and using computer technology, the development of the blockchain-based smart contracts phenomenon offers interesting cases for the proposed Mjimba-Sibanda model. A blockchain is a distributed data structure replicated and shared among the members of a network [41]. Smart contracts are instruments that coded to automatically execute when certain criteria are met [41]. Merging the blockchain and smart contracts innovations gives rise to decentralised self-executing and self-enforcing contracts. Similarly, the proposed Mjimba-Sibanda model envisages future-oriented self-executing mechanisms aiming to manage climate change. Its criteria for executing changes will be a continuous computation of biological, climatic, physical and other data to generate policy that drives best practice concerning both climate change mitigation and adaptation.
