Hidrológiai Közlöny, 2021 (101. évfolyam)
2021 / Különszám
116 Hidrológiai Közlöny 2021. 101. évf. különszám This is true despite the backdrop of many more localized disagreements about water (Pacific Institute 2020). CONCLUSIONS This brief analysis has demonstrated the feasibility of combining existing data sets and an integrative framework to generate a geography of the sources for potential disputes and conflicts over fresh water. While our aim has not been to predict water-derived conflicts per se, our initial assessment indicates a general pattern of agreement with an observed conflicts database across a continental domain that includes much of Europe and parts of the Middle East. We reason that this preliminary success can be attributed, in part, to capturing an essential geography that links threat producing biogeophysical factors generated upstream to affected downstream populations, and as modulated by additional societal indicator data (i.e. transboundary nature of the basin in question; presence/absence of fragile states). While our initial tests have focused on national-scale indicators of politically fragile states and transboundary effects, additional indicators such as sub-national conflicts, asymmetries in economic power, and the control or release of upstream resources to downstream users could have similarly amplifying impacts on the potential conflict geography. Additionally, cross-sectoral issues could also weight heavily, for example with irrigated water use competing with the cooling water requirements of thermoelectric power production (Miara et al. 2017) or with urban water supplies that require water in both sufficient quantity and quality, unencumbered by aggravated pollution loads associated with food production (McDonald et al. 2016). The technical capacity to manage water, if not the policy dimensions of avoiding conflict, are also an important component of the issue at hand, but is woefully inadequate in many parts of the world (Wehn de Motalvo and Alaerts 2013). Without decision-makers and practitioners who can identify and act on the multi-dimensional aspects of conflict avoidance, it will be difficult to see how cooperation will emerge spontaneously. An important development on the horizon involves the use of more complex statistical models and the use of machine learning to better develop this capability. Recent work by Kuzma et al. (2020) exemplifies such an approach and focuses on the use of a random forest model to predict (as opposed to explain) such tendencies. The authors caution against using a proliferation of predictor variables. This can, in theory, improve model performance but risks elevating the levels of input covariance while simultaneously reducing the capacity to understand, interpret, and formulate policy around such models. Overall, the researchers indicate that it is difficult to pinpoint water as the controlling factor in conflict generation, but that it can be a contributing or amplifying factor where there are existing stressors on the human-environment system (Gleich and Iceland 2018). In the Central European domain, an excellent example of this limitation involves the Gabcikovo- Nagymaros conflict, which was not driven by the shortage of water or biogeophysical stresses themselves, but historical disputes between Slovakia and Hungary, where the Danube turned out to be the point of conflict. If there were no river between these countries the flashpoint would likely have been over some other precipitating factor. Nevertheless, the linkage of machine learning to the network topological approaches that we advocate here could help to decipher the causal links between water and other resource stresses, geopolitical precursors, and the emergence of actual conflict. Our metrics are suitable for developing hypotheses to better understand the origins and potential severity of water conflict. For example, given a threshold probability, we could identify when water issues themselves are the genesis of a conflict P(W|C) rather than the trigger of conflicts when there are underlying tensions (P(C|W). Using this information, we then could hypothesize that P(W) and P(C) are independent (that is P(W|C) = P(W) and P(C|W) = P(C)). Demonstrating that these conditional probabilities are not independent, we would reject the hypothesis and then explore the driving biogeophysical and human dimension factors contributing to this result. We thus view the merger of such approaches to be promising, given the multi-dimensional nature of the interactions, which cause humans to move toward (or away from conflict) and the inescapable geography of upstreamdownstream linkages that constitute a large proportion of the global water resource base on which the development agenda must rely. Despite whatever variables one might choose to analyse in the context of the elusive capacity to predict water conflicts it is important to imagine how the continued pressures on the resource could be attenuated by innovative engineering solutions as well as improved environmental stewardship. On the engineering side, the transition toward higher levels of water productivity (Flörke et al. 2013, Gleich and Palaniappan 2010), that is the de-intensification of water use through efficiency gains, demand management, and reuse is a minimum entry point for lessening the pressure on an already oversubscribed resource in many parts of the world. An equally, and arguably potentially more critical policy objective is to improve the management of the ecosystems, as when using upland watersheds to provide clean drinking water or wetlands to enhance flood attenuation, thus reducing the need for investments in costly infrastructure. With the continued globally significant loss and degradation of ecosystems (Venter et al. 2016), comes losses in important ecosystem services, which encumber society with enormous costs (Costanza et al. 2014). Valuing natural capital and nature-based solutions and incorporating them into water resource systems (WWAP 2018, Browder et al. 2019) will help support long-term, sustainable water security in the face of climate change and variability, population growth, land use change and agricultural intensification, anticipated multi-sectoral economic growth, and inter-sectoral competition for water. Such strategies for incorporating environmental system protection and rehabilitation are therefore a part of the formula for water security and thus conflict avoidance. These must be adopted beyond the actions of governments and civil society alone, specifically by developing business
