Does Mitigation Reduce Net Global Impacts of Climate Change on Water Resources?
Arnell, N.W., van Vuuren, D.P. and Isaac, M. 2011. The implications of climate policy for the impacts of climate change on global water resources. Global Environmental Change 21: 592-603.
Specifically, based on modeled estimates of river run-off, the study calculates annual water available for each of 1163 watersheds into which it partitions the world's land area. It then characterizes a watershed as water-stressed if the annual amount of water available per capita (i.e., the water stress index, WSI) is below 1000 m3 per capita per year or if the ratio of annual withdrawals to available water resources (i.e., water resources vulnerability indicator, WRVI) exceeds 0.4. The analysis uses four Atmosphere-Ocean General Circulation Models (AOGCMs). To calculate impact, it estimates the population that (a) becomes water stressed and (b) no longer experiences water stress. However, it does not calculate the net change in the population experiencing stress, arguing that (a) in areas that experience a relative increase in water availability, it may not be possible to store the extra water through a dry season, or the extra water may occur during flood events, and (b) the populations experiencing increase and decrease are not the same.
All scenarios assume that global population will increase to 8.8 billion in 2050 and stabilize at 9.1 billion before 2100. Per van Vuuren et al. (2011), the scenarios also assume that economic growth will follow the IPCC's B2 scenario till 2050; however, the write-up is vague regarding the subsequent rate of economic growth (van Vuuren, 2011, p. 576).
River runoff is simulated using a global hydrological model. However, there is no quantitative discussion of validation results for this model or of the performance of any of the AOGCMs using empirical data that were not used in model development. Also, the analysis ignores autonomous adaptations that ought to occur due to economic development and secular technological change (Goklany 2007) under all scenarios. This is contrary to the IPCC's methodological guidelines for conducting impacts analyses (Carter et al., 2007, p. 136, footnote 2). More importantly, this means that the results overestimate negative impacts while simultaneously underestimating positive impacts (Goklany 2009), that is, the increase in population experiencing water stress is overestimated while the population no longer experiencing water stress is underestimated.
According to the study, in 2100, the global population experiencing water stress would be 3.9 billion using WSI, and 4.1 billion using WRVI in the absence of climate change. Results for the increase and decrease in the population experiencing stress (relative to the "no climate change" condition) are shown in the following table for 2020, 2050, 2080 and 2100. Despite the study's reservations, the net change in the population experiencing stress is also calculated here for three reasons. First, as noted, the study overestimates the population increases experiencing water stress while underestimating the decreases. Even if one accepts that there is an asymmetry in consequences for human welfare from equivalent increases and decreases in the water-stressed population, this overstating of net negative impacts would, arguably, more than compensate for that asymmetry. Second, the study's authors have made a value judgment and assumed that readers subscribe to that judgment. Third, the mitigation scenarios would themselves reduce water availability in some areas, while increasing it in others, and that too needs to be accounted for in order to evaluate the utility of mitigation.
The table indicates that the net increase in the water-stressed population in 2100 (for example) would be dominated by non-climate change factors than climate change (3.9 billion or 4.1 billion vs. a maximum of 1.3 billion or 0.8 billion, using WSI or WRVI respectively). It also shows that depending on the scenario and AOGCM employed, climate change may or may not increase the net water-stressed population through 2100 (relative to "no climate change"). Similarly, even after mitigation, the net water-stressed population may or may not be higher relative to the "no climate change" case. Equally importantly, mitigation may actually increase the net water-stressed population over the unmitigated climate change scenario. These instances are highlighted on the table. For instance, for 2100, see the HadCM2 and CSIRO2 cases under the WSI criterion and all but the CGCM11 case under the WRVI criterion.
Thus, this paper shows that, based on water resource impacts, there is no compelling case for a mitigation policy. If anything, it suggests that adaptation policies would be much more beneficial because it would enable watersheds that would experience an increase in water availability to capture those benefits, while enabling others to cope with any negative impacts, as noted by others previously (Goklany 2009).
Carter, T.R., Jones, R.N., Lu, X., et al. 2007. New assessment methods and the characterisation of future conditions. In: Intergovernmental Panel on Climate Change. 2007. Climate Change 2007. Impacts, adaptation, and vulnerability. Cambridge, UK: Cambridge University Press.
Goklany, I.M. 2009. Is Climate Change the "Defining Challenge of Our Age"? Energy & Environment 20: 279-302.
van Vuuren, D. P., Isaac, M., Kundzewicz, Z. W., Arnell, N., Barker, T. Criqui, P., Berkhout, F., Hilderink, H., Hinkel, J., Hof , A., Kitous, A., Kram, T., Mechler, R., and Scrieciu, S. (2011). The use of scenarios as the basis for combined assessment of climate change mitigation and adaptation. Global Environmental Change 21: 575-591.