Water footprint: a tool for climate adaptation?
28/02/12 15:42
Today, a major new article on water footprinting by Hoekstra and Mekonnen published in PNAS as an open-source PDF. Hoekstra is the most famous proponent of water footprint analyses, and he’s responsible for publicizing one of the most influential ideas in the past decade or two about water.
If you’ve not been exposed to WF, the idea is fairly simple. Water can be both “real” (i.e., actual water in precipitation, rivers, and so on) as well as “virtual” (i.e., embedded in products where water is used to make or grow those products, such as wheat, cotton in clothing, beer, petroleum [from the refining process], electricity [from hydropower], and iron and steel).
Perhaps the classic WF analysis looks at how much water is “embedded” within coffee. One estimate suggests about 140 liters of water make a single cup of coffee.
A WF analysis can be used to provide many kinds of information. In today’s paper, Hoekstra and Mekonnen use about ten years of data to develop a WF analysis of global trade of real and virtual water between countries. Some countries are net exporters of water. Some are net importers. Because water is so often hidden, countries may be exposing themselves to insecure water conditions by exporting significant amounts of water.
There is some interesting subtlety in this paper and in other WF analyses. For instance, water can also be characterized into different categories. In the same paper, water is also categorized as “blue” (surface and groundwater sources of water such as aquifers and lakes), “green” (precipitation), or “grey” (polluted water, such as from industrial or agricultural effluent). These are heuristic and thoughtful categories.
The authors argue that this approach is relevant for those involved in global trade — consumers, corporations, and policymakers concerned about water scarcity. In other places, Hoekstra has also argued persuasively that global trade and the global scale of climate change both force water managers and decision makers to think beyond the scale of particular river basins. In fact, the amount of water moving around means that global trade represents a massive virtual inter-basin transfer. There is much merit in these arguments and I find these arguments insightful.
As with other papers by Hoekstra, today’s PNAS paper will be widely cited and is a very useful contribution to thinking about water security, global flows of resources, and trade relationships, but is the analysis connected to climate change?
The short answer is not yet. In today’s paper, the authors do not claim that their analysis is relevant to looking at either climate impacts or climate adaptation. However, other writings about WF make claims about using WF as a tool for adaptation. There are several weaknesses in this argument:
These thoughts are not really a criticism of WF, and certainly not a criticism of Hoekstra’s thinking over the past decade. WF is a powerful tool that has been extremely useful. My own thinking about water has been deeply influenced by WF. However, I am also concerned about facile applications of this tool to climate adaptation and long-term water resources management. Not all water problems have a climate or climate change component, particularly for corporate audiences, which are often focused on non-climate timescales.
Ultimately, the WF movement is good to point out that water usage is essentially a choice. A country such as Costa Rica may choose to export large quantities of its water as coffee. But that choice also reflects the rich water reserves within Costa Rica. Hopefully WF serves as a tool for more effectively representing risk and vulnerability as a moving, shifting target, with climate and climate change as contributors to those evolving vulnerabilities.
Perhaps the classic WF analysis looks at how much water is “embedded” within coffee. One estimate suggests about 140 liters of water make a single cup of coffee.
A WF analysis can be used to provide many kinds of information. In today’s paper, Hoekstra and Mekonnen use about ten years of data to develop a WF analysis of global trade of real and virtual water between countries. Some countries are net exporters of water. Some are net importers. Because water is so often hidden, countries may be exposing themselves to insecure water conditions by exporting significant amounts of water.
There is some interesting subtlety in this paper and in other WF analyses. For instance, water can also be characterized into different categories. In the same paper, water is also categorized as “blue” (surface and groundwater sources of water such as aquifers and lakes), “green” (precipitation), or “grey” (polluted water, such as from industrial or agricultural effluent). These are heuristic and thoughtful categories.
The authors argue that this approach is relevant for those involved in global trade — consumers, corporations, and policymakers concerned about water scarcity. In other places, Hoekstra has also argued persuasively that global trade and the global scale of climate change both force water managers and decision makers to think beyond the scale of particular river basins. In fact, the amount of water moving around means that global trade represents a massive virtual inter-basin transfer. There is much merit in these arguments and I find these arguments insightful.
As with other papers by Hoekstra, today’s PNAS paper will be widely cited and is a very useful contribution to thinking about water security, global flows of resources, and trade relationships, but is the analysis connected to climate change?
The short answer is not yet. In today’s paper, the authors do not claim that their analysis is relevant to looking at either climate impacts or climate adaptation. However, other writings about WF make claims about using WF as a tool for adaptation. There are several weaknesses in this argument:
- The underlying data for today’s paper and other WF analyses I have seen have a strong assumption of stationarity. That is, they assume that the past will be a useful and accurate guide to emerging conditions — that the water cycle itself is not in a period of rapid evolution from a period of climate change. Over short time scales of less than a decade, this may be a productive assumption. But it is less useful at longer timescales (>10 years). The stationarity problem is a common one in water and there is a lot going on to work around this issue. It does not undermine WF for all applications, but for institutions, infrastructure, and ecosystems that will be managed for longer than a decade, relying on WF could lead to dangerous vulnerabilities.
- Many “applications” of climate change to WF rely on taking projected climate model data for particular temperature and precipitation variables from some future point and applying them without adjustment or qualification current usage patterns. As has been discussed on this site many times, climate model projections have low confidence and credibility. For many applications, they are not appropriate for water management do not represent the state of the art. Climate models are essentially hypotheses for testing particular sets of assumptions (called “scenarios”) about how global economic and policy trends will evolve in coming decades. They are very useful when used as they were designed, but they were not designed for climate change adaptation projects, particularly not at local or even national scales (unless your country is the size of China, Canada, or Russia). Model data is not even consistent across much multiple scenarios, much less across the two-dozen odd models and all of their individual scenarios. Water projections are especially weak among the climate variables considered. This is not a criticism of Hoekstra’s work or today’s paper; no allegations are made in connection to this issue by Hoekstra. But many applications of WF do use make claims about climate adaptation using this approach.
- Most WF applications do not look at water consumption within basins or watersheds. The Hoekstra/Mekonnen paper, for instance, looks at a 5x5 degree grid globally, but liquid water on the earth’s surface is organized into basins rather than rectangles; water flows downhill rather than along longitude and latitude coordinates or even along political boundaries. In many cases, these drainages may even cross national borders, complicating coherent water management. Dividing the earth’s surface into arbitrary units may obscure where there are problems with water scarcity and overabundance. With all due respect, basins are probably still the most critical unit of water management, even when we need to think at other scales too. Certainly for looking at economic and ecological vulnerabilities, the basin is extremely important for accurately assessing impacts.
- The temporal scale of assessment in most applications of WF does not match how humans and other species (much less ecosystems) “experience” water, which is much closer to either a normalized season or to extreme events, such as droughts and floods. A growing body of research shows that humans and other species have responded to most freshwater according its normal seasonal variation of highs and lows, which is usually called the natural flow regime or water seasonality. Freshwater conservationists call this the “master variable” for ecosystems, since so many ecological processes are influenced by the timing of this variable (e.g., see an annual hydrograph here). Moreover, the natural flow regime essentially reflects ambient climate, and it is extremely sensitive to changes in climate. In contrast, many economists and policymakers prefer to discuss water at an annual scale — the average amount of water in a particular place annually. This makes the math simple, but it obscures both the variation (the frequency and severity of extreme events or “climate variability”) and the sub-annual scales that determine what species can live in a particular place, what livelihoods or industries may rely on seasonal timing, etc. And since climate change is altering the timing of precipitation, there can be very important vulnerabilities that are connected to shifts in timing. A small change in the spring flood may mean that fish do not receive a single to spawn, while a slight shift in the timing of the annual monsoon may mean that farmers plant or harvest their crops at the wrong time. Neither of these changes are likely to show up as significant shifts in annual precipitation, but they are absolutely critical to what happens where — and when.
- Lastly, climate change is also making some places wetter rather than just drier. Often a WF analysis tends to focus on mean water scarcity conditions rather than too much water (or water at the wrong time — see above). Many regions are experiencing both too much and too little water, with an explosion in the number of both types of extremes.
These thoughts are not really a criticism of WF, and certainly not a criticism of Hoekstra’s thinking over the past decade. WF is a powerful tool that has been extremely useful. My own thinking about water has been deeply influenced by WF. However, I am also concerned about facile applications of this tool to climate adaptation and long-term water resources management. Not all water problems have a climate or climate change component, particularly for corporate audiences, which are often focused on non-climate timescales.
Ultimately, the WF movement is good to point out that water usage is essentially a choice. A country such as Costa Rica may choose to export large quantities of its water as coffee. But that choice also reflects the rich water reserves within Costa Rica. Hopefully WF serves as a tool for more effectively representing risk and vulnerability as a moving, shifting target, with climate and climate change as contributors to those evolving vulnerabilities.
blog comments powered by Disqus
