Back in May of 2015, I wrote this article, Understanding water footprint numbers, after spending a month corresponding with the organization generating these numbers: WaterFootprint.org. Prior to this correspondence and writing this article (now posted on my blog), like many others I was a bit mystified as to how the global water footprint numbers always bantered about in the press were derived.
Waterfootprint.org was very helpful. They sent me research papers to read and responded to all of my questions. As the below excerpts from our written correspondence will show, most people who cite or critique water footprint methodology don’t fully understand how this methodology is done. For example, the global numbers bantered about are provided solely to “raise awareness” of water usage. They’re not really intended to be used as tools for assessing sustainability of water use. Why? These global numbers don’t really provide any information about local conditions especially in regards to environmental impacts. The global numbers are just global averages that may or may not be relevant for any given place on the globe. As further explained in the excerpts below, to generate water footprints, a grid is superimposed on land areas of the planet, and local data for each portion (grid sector) of the map is interpolated. From this interpolated data, water balance equations are used to derive water footprint numbers at the local level. So more specific local data can be derived from this GIS mapping process by looking at only the data for specific areas of the map.
As my final comment from the excerpts below details, the real shortcoming in this modeling process is that many parameters are not or cannot be properly accounted for in the mathematical equations. Many of these parameters, that are either averaged or not factored in at all, really change the environmental outcomes. For example soil type, synthetic nitrogen use are averaged over wide areas plus soil organic matter isn’t accounted for at all. Neither are polycrop systems or different methods of irrigation or the different depths of aquifers.
Since droughts and floods are as much a function of how much water infiltrates and is retained by soil and refill aquifers, instead of just how much rain falls from the sky, water balance equations that don’t or can’t account for such parameters like soil health, even when applied at a more local level, don’t really tell the full story of water usage. (More soil organic matter retains a lot more water. Per the NRCS, an one percent increase in soil organic matter over an acre of land may hold to as much as an additional 27,000 gallons of water).
Moreover, land use isn’t interchangeable, but appropriateness of land use also isn’t accounted for in any of the math. So the best or most appropriate use of green or rain water that falls isn’t reflected in any of the outputs. Plus not all blue or ground water is the same. (Remember too from my prior article, blue water is what’s critical…not green water. Green water is primarily rain water. Green water falls regardless of the land use and in many footprints is most of the water…for example, see graphics for beef just below). A shallow aquifer that recharges from rainfall or flood irrigation is a lot different than a pumped deep fossil aquifer that won’t refill for thousands of years. Again, water footprints, especially global averages, when bantered about aren’t very useful when it comes to ascertaining environmental impacts.
So in other words such water footprints, derived from these water balance equations, are really crude and flawed tools.
ME: living with drought in California, water here is a major issue. In doing some research to better try to understand water foot print numbers, I needed some clarifications as to how some of these numbers are derived for certain crops, in particular almonds. So here are some questions:
- When calculating the footprint for an almond does the WFP# include the three or so years it took to grow the tree before the tree produces almonds?
- Is this true of any other type of tree for fruits or nuts, that is does the time (lifecycle) of used water include the years and months it took to grow the tree before the produces fruit or nuts?
- When calculating the footprint for an almond does the number account for different methods of irrigation for example drip irrigation versus other forms of irrigation?
- What about the impact of salinity in soil decreasing yields? If fewer almonds are produced due to salinity, wouldn’t the average number per almond be higher with lower yields?
- In general, are any impacts of the soil’s capacity to retain green water factored in the math of deriving water foot print numbers since rain water is better utilized with better soil infiltration?
ME: Thank you very much for your fast response. Reading through the link you sent me, I don’t fully understand some of the language. Specifically some of these terms:
- grid-based approach with a spatial-resolution of 5 by 5 arc minute
- crop production with a spatial-resolution of 30 by 30 arc minute
- The grid-based dynamic water balance model
What exactly is a grid based approached? Is it some sort of grid overlay applied to a region for a certain amount of land? Do you have a detailed example for a crop of an application of this “grid-based water dynamic balance model”? I found an example for the CROPWAT 8.0 model. Also do they actually physically measure soil moisture or is this derived from already available data for different crops?
WFP: Arc minute grid is a term used in Geographical Information Systems (GIS), AutoCad and other spatial analysis tools. The numbers, like 5 by 5 or 30 by 30, represent the resolution of the spatial information created (grid cells’ dimensions). For convenience the 5 and 30 arc-minute grids are usually called as 10 km and 50 km grids, respectively, although that’s not exactly/technically true, since the size of grid cell changes with latitude (these values are true for the Equator). So, in practical terms, these numbers refer to the spatial resolution of the data. Within each grid cell you will only have one value for the information you are computing/analyzing and which can be climacteric, soli, crop, etc. The resolutions are chosen according to the available data, and preference is given to higher resolutions. Higher resolutions provide more accurate data when one zooms in a particular area.
The grid-based water dynamic balance model inputs for each crop are climate data – such as precipitation, temperatures, potential evapotranspiration – and soil characteristics data. The model will then compute the daily soil water balance and calculates crop water requirements, actual crop water use and actual yields.
WFP: By the way… why are you digging into this information? Just curious in case you feel like sharing.
ME: In short, because the numbers are frequently referred to, I’m trying to understand what the numbers actually mean and how they are derived plus whether or not the numbers are a good guide for assessing the actual environmental impact. Most people just use them and don’t understand the basis behind the numbers derivation this is especially true of the media and advocacy groups. There are a lot of parameters involved, and, in general, people oversimplify without looking at the local conditions or specific methods of production of crops or animals.
WFP: I couldn’t agree more with you that numbers can be deceiving when oversimplified. When we hear about a certain product water footprint, the two main questions, to understand its sustainability are:
- Where is it produced? Is it produced where water resources are abundant and enough for all needs, including the environment ones? Or is it produced where water scarcity and/or water pollution levels are beyond acceptable safe limits?
- Is the water use efficient for its production? Or is it being produced using (consumption + pollution) more water than what would be required if good practices were in place?
The answers to these questions are what allow us to understand if that product water footprint is sustainable or not. A product water footprint is sustainable if it is produced in rivers basins without water scarcity and pollution problems and if water for its production is used in an efficient way.
Our data on global and country level products’ water footprint, are meant to raise awareness about the hidden water use, and make businesses, governments and consumers aware that water is limited resource and we are overusing it. These data on water footprint of products helps consumers understanding the impact of their daily choices; businesses that they need to understand the context they and their supply chain are operating and manage water resources properly and in order to ensure their long-term sustainability and minimize risks; and governments to understand that water policies and regulations need to be crosscutting through all sectors.
ME: Yes a lot of what you write is correct, but water foot print numbers (especially global ones) also include a lot of assumptions (soil composition/ nitrate use/methods/non-mixed systems, etc) and don’t necessarily seem to reflect environmental impact especially in a place like California where a lot of non-arable land has been converted to arable land via blue water diversion or pumping. In these cases where diverse grasslands with cattle have been converted to mono-crop land, bio-diverse grasses/forbs requiring only rainfall (no synthetic inputs or irrigation) to produce food (pastured meat with high foot print numbers per your stats) have been replaced with mono crops including nut trees requiring a lot of irrigation, pesticides and fertilizers. In assessing the environmental impact for these crops on converted lands, looking at where the blue water comes from and how it is applied also is important. When wells are drilled too shallow they add too many nitrates on the soils, when drilled to deep, salinity and selenium may become an issue. Salinity also reduces yields, and thus theoretically should increase foot print numbers (I don’t see how these issues are factored into your current numbers). The irrigation type matters too. Drip , micro-sprinklers, sprinklers and flood irrigation all have different ramifications regarding the amount of water applied. That’s why I asked about whether your numbers accounted for irrigation type. Per your reply, your data currently doesn’t include this information.
Furthermore when the crop land isn’t converted, and cattle is raised on the grasslands without synthetic nitrates, the dispersed cattle’s manure builds the soil improving water infiltration rates via increased soil microbial activity meaning any rainfall that does fall is more effectively utilized. So grey water will also obviously be further reduced. Where converted land is tilled, soil biology is disrupted, leading to compaction, less water absorbed and more run off with polluted waters particularly where NPK’s have been applied. Uncovered tilled soils (no mulch or cover crops) also get hotter which also kills soil microbes including methanotrophs which again leads to compaction AND reduces the soils ability to oxidize methane. Tillage also releases soil carbon into the atmosphere. So crop land with and without diverted blue water can have a tremendous adverse environmental impact upon water use even if the crops grown have lower water footprint numbers (especially where more blue water is required).
Now don’t get me wrong, the GIS models of water use interpolating data are interesting tools. But most people who cite the global numbers in the media or to further their food ideologies have zero understanding of how these foot print numbers are actually derived or what assumptions are made. If people understood the mapping process, it seems they could go into the models, put in more specific data highlighting the parameters that effect a specific location and get a number more reflective of that given place.
Applying the numbers universally to me just seems like a gross oversimplification and abstraction. Comparing numbers of different products also seems problematic too especially in the case noted above where non-arable land is converted to arable land or where non-arable land remains non-arable or where land is developed. it seems appropriate land use should factor into your equations. Unless, I’m mistaken appropriate land doesn’t seem to be applied. Again remember not all land is appropriate for crops especially industrial crops, so the most effective use of green water in those places may be pastured meat production instead of vegetables, fruits or nuts. Here the water used for grasses is a lot less than water used for feed crops, and even though the life of the respective heifer or steer is longer to get to finish weight , the environmental impact of the water used is smaller and more appropriate.