What is the difference between ‘water withdrawal’ and ‘water consumption’, and why do we need to know?

What is the difference between ‘water withdrawal’ and ‘water consumption’, and why do we need to know?

Post by Inge de Graaf, postdoctoral researcher at the University of Freiburg, Environmental Hydrological Systems group


Last week I had to teach my first class in global hydrology. When I showed the global trend on increasing demands and withdrawals (see Figure) I needed to explain the different terms as sometimes the term “water use” gets, well, misused.

The term “water use” often fails to adequately describe what happens to the water. So I told the students; if you see or hear to term ‘water use’ always ask yourself what’s actually being said. The term is often used for water withdrawals or water consumption, and it’s important to understand the difference.

Water withdrawal describes the total amount of water withdrawn from a surface water or groundwater source. Measurements of this withdrawn water help evaluate demands from domestic, industrial and agricultural users.

Water consumption is the portion of the withdrawn water permanently lost from its source. This water is no longer available because it evaporated, got transpired or used by plants, or was consumed by people or livestock. Irrigation is by far the largest water consumer. Globally irrigated agriculture accounts for 70% of the total water used and almost 50% is lost either by evaporation or transpiration.

Understanding both water withdrawal and consumption is critical to properly evaluate water stress. Measurements of water withdrawal indicate the level of competition and dependence on water resources. Water consumption estimates help to quantify the impact of water withdrawals on downstream availabilities and are essential to evaluate water shortage and scarcity. For example, most water used by households is not consumed and flows back as return flow and can be reused further downstream. However, water is rarely returned to watershed after being used by households or industry without changing the water quality, increasing water stress levels.

Already more than 1.4 billion people live in areas where the withdrawal of water exceeds recharge rates. In the coming decades global population is expected to increase from 7.3 billion now, to 9.7 billion by 2050 (UN estimate). This growth, along with rising incomes in developing countries, is driving up global food demands. With food production estimated to increase by at least 60% (FAO estimate), predicting water withdrawal and consumption is critically important for identifying areas that are at risk of water scarcity and where water use is unsustainable and competition amongst users exist.

Global trend I showed in my class, published in Wada et al (2016).


Wada, Y., I. E. M. de Graaf, and L. P. H. van Beek (2016), High-resolution modeling of human and climate impacts on global water resources, J. Adv. Model. Earth Syst., 8, 735–763, doi:10.1002/2015MS000618.


How prehistoric water pit stops may have driven human evolution

How prehistoric water pit stops may have driven human evolution

Post by Matthew Robert Bennett, Bournemouth University and Mark O Cuthbert, Cardiff University

Our ancient ancestors seem to have survived some pretty harsh arid spells in East Africa’s Rift Valley over five million years. Quite how they kept going has long been a mystery, given the lack of water to drink. Now, new research shows that they may have been able to survive on a small networks of springs.

The study from our inter-disciplinary research team, published in Nature Communications, illustrates that groundwater springs may have been far more important as a driver of human evolution in Africa than previously thought.

Great rift valley.Redgeographics, CC BY-SA

The study focuses on water in the Rift Valley. This area – a continuous geographic trench that runs from Ethiopia to Mozambique – is also known as the “cradle of humanity”.

Here, our ancestors evolved over a period of about five million years. Throughout this time, rainfall was affected by the African monsoon, which strengthened and weakened on a 23,000-year cycle. During intense periods of aridity, monsoon rains would have been light and drinking water in short supply. So how did our ancestors survive such extremes?

Previously, scientists had assumed that the evolution and dispersal of our ancestors in the region was solely dependent on climate shifts changing patterns of vegetation (food) and water (rivers and lakes). However, the details are blurry – especially when it comes to the role of groundwater (springs).

We decided to find out just how important springs were. Our starting point was to identify springs in the region to map how groundwater distribution varies with climate. We are not talking about small, babbling springs here, but large outflows of groundwater. These are buffered against climate change as their distribution is controlled by geology – the underlying rocks can store rainwater and transfer it slowly to the springs.

The lakes of the African Rift Valley.SeaWiFS Project, NASA/Goddard Space Flight Center, and ORBIMAGE

We figured that our ancestors could have stayed close to such groundwater in dry times – playing a greater part in their survival than previously thought. When the climate got increasingly wet, groundwater levels would have risen and made springs more plentiful – feeding smaller rivers and leading to lakes becoming less saline. At this point, our ancestors would have roamed across the landscape free of concerns about water.

Life and death decisions

To test this idea, we embarked on a computer experiment. If the springs and water bodies are thought of as the rest stops, or service stations, then the linkages between can be modelled by computers. Our model was based on what decisions individuals would have taken to survive – and what collective behaviours could have emerged from thousands of such decisions.

Individuals were give a simple task: to find a new source of water within three days of travel. Three days is the time that a modern human and, by inference, our ancestors could go without drinking water. The harder and rougher the terrain, the shorter the distance one can travel in those vital three days.

We used the present landscape and existing water springs to map potential routes. The detailed location of springs may have changed over time but the principles hold. If our agent failed to find water within three days, he or she would die. In this way we could map out the migration pathways between different water sources as they varied through 23,000-year climate cycles. The map shows that there were indeed small networks of springs available even during the driest of intervals. These would have been vital for the survival or our ancestors.

The model also reveals movement patterns that are somewhat counter-intuitive. One would assume that the easiest route would be along the north to south axis of the rift valley. In this way, hominins could stay at the bottom of the valley rather than crossing the high rift walls. But the model suggests that in intermediate states between wet and dry, groups of people may have preferred to go from east to west across the rift valley. This is because springs on the rift floor and sides link to large rivers on the rift flanks. This is important as it helps explain how our ancestors spread away from the rift valley. Indeed, what we are beginning to see is a network of walking highways that develop as our ancestors moved across Africa.

Mapping human migration.

Human movement allows the flow of gossip, know-how and genes. Even in modern times, the water-cooler is often the fount of all knowledge and the start of many budding friendships. The same may have been true in ancient Africa and the patterns of mobility and their variability through a climate cycle will have had a profound impact on breeding and technology.

This suggests that population growth, genetics, implications for survival and dispersal of human life across Africa can all potentially be predicted and modelled using water as the key – helping us to uncover human history. The next step will be to compare our model of human movement with real archaeological evidence of how humans actually moved when the climate changed.

So next time you complain about not finding your favourite brand of bottled spring water in the shop, spare a thought for our ancestors who may died in their quest to find a rare, secluded spring in the arid African landscape.

The ConversationThis research was carried out in partnership with our colleagues Tom Gleeson, Sally Reynolds, Adrian Newton, Cormac McCormack and Gail Ashley.

Matthew Robert Bennett, Professor of Environmental and Geographical Sciences, Bournemouth University and Mark O Cuthbert, Research Fellow in Groundwater Science, Cardiff University

This article was originally published on The Conversation. Read the original article.

Is highway de-icing ‘a-salting’ our aquifers?

Is highway de-icing ‘a-salting’ our aquifers?

Post by Mark Cuthbert, Cardiff University, and Michael Rivett, GroundH20 plus Ltd; University of Strathclyde.

If you live in a cold climate, have you ever wondered where all the de-icing salt (or ‘grit’ as we call it in the UK) that gets spread on the roads in winter time ends up, aside from that accumulating salty grime that coats your car? As you might expect, most of the salt gets washed off the highways as the salt has the desired effect of melting the ice, or carried away by rain. This salty ‘runoff’ ends up in streams nearby via pipes which drain the highway. However, that is not the end of the story…

We studied a major highway intersection on the edge of UK’s second largest city, Birmingham (more than a million people), to see how much of the salt spread on the highway ends up in groundwater. Our interest stems from concerns of the national regulator to understand not only how much salt ends up in streams, but also the potential long-term build-up of salt in underlying groundwater resources that are pumped for public and private water supply. Although the origin of salt in streams and lakes is relatively well studied, the pathways by which salt moves from highways to groundwater are poorly understood and quantified.

The various ways salt could be getting in to the groundwater are shown in Figure 1 and a bit of detective work was required to find out what was going on.

Figure 1*

This involved dressing in silly (and warm) clothes, and pulling our equipment around on a sledge, which was a lot of fun (see photos below).

The authors, Mark (left) and Mike (right), enjoying some urban winter fieldwork.

Our main field activities were fortuitously timed to include the severe winters of 2009-10 and 2012-13. We collected information about the amount of salt spread on the roads and measured the salt concentrations in the local streams, as well as in the shallow and deep groundwater wells in the area around the highway network.

7 km of motorway drained to the studied stream. Over the winter of 2012-13 we estimated that around 510 tonnes of de-icing salt was applied to the highway and major road network across the catchment. Most of that washed off the road via drains when it next rained or as the snow and ice melted, and ended up in the stream which flows under the highways. Some of it ended up on the roadside verges – this was not quantified, but would likewise eventually leach into the underlying groundwater over time.

We estimated about 12% of the de-icing salt, that’s around 63 tonnes over the 2012-13 winter, leaked through the bottom of the stream channel into the groundwater in the sandstone aquifer beneath, and may pose a risk to groundwater supplies in the area (see “9” in Figure 1, and results in Figure 2). While increases in groundwater chloride concentrations observed to date have been modest, and fortunately remain far below the drinking water standards, the steady year-upon-year build-up of salt in groundwater remains a concern. This is especially the case as the UK’s de-icing salt applications reached record levels in the recent severe winters and add to a potential de-icing salt legacy in our aquifers that has accumulated now over some 50 plus years.

Figure 2*

So how concerned should we be that so much salt is flowing down these streams and into the groundwater around our highways? There isn’t yet a simple answer to that but, with a global trend towards increasing urbanisation, this is an important area of ongoing and future research.


Further details of our published research to date on the Birmingham case study may be found at:

Rivett, M.O.,  Cuthbert, M.O., Gamble R., Connon, L.E., Pearson, A., Shepley, M.G., Davis, J., 2016. Highway deicing salt dynamic runoff to surface water and subsequent infiltration to groundwater during severe UK winters.  Science of the Total Environment 565, 324-338. http://dx.doi.org/10.1016/j.scitotenv.2016.04.095 (*figures reproduced with permission) [Read More]

How did our planet get its water?

How did our planet get its water?

Post by WaterUnderground contributors Elco Luijendijk and Stefan Peters from  the University of Göttingen, in Germany.

After my first ever scientific presentation, someone in the audience asked a question that caught me off guard: “Where does the groundwater come from?”.  “Ehm, from rainfall”, I answered. The answer seemed obvious at the time. However, we did not realize at the time that this is actually a profound question in hydrogeology, and one that is rarely addressed in hydrology textbooks and courses: “How did our planet get its water?”. To find out how far science has come to answering this question I (EL) joined up with a geochemist and meteorite expert (SP) to write this blog post.

We are lucky to live on a planet of which ~71% of the surface is covered with water, located mostly in rivers, lakes, glaciers and oceans at the surface and as groundwater in the shallow subsurface. Liquid water sustains life on our planet and seems to play a critical role in plate tectonics. And incidentally, it also to gives hydrogeologists something to study. Liquid water is so important in sustaining life, that the search for life on other planets in our solar system or beyond always focuses first on finding planets with liquid water.

Not only do we have abundant liquid water, we seem to have just the right amount. Compared to our direct planetary neighbors, Mars and Venus, we are extremely lucky. On the surface of Mars, at present, water mainly occurs as ice, whereas tiny amounts of water vapor are present in the Martian atmosphere. Venus also has minute amounts of water vapor in the atmosphere, but its blazingly hot surface is entirely devoid of water. In contrast to Mars and Venus, some objects in the solar system that are further away actually have too much water. Take for instance Enceladus, a moon of the planet Saturn, at which an icy crust overlies a 10 km deep water ocean. The amount of water on Enceladus is so large that it causes a wobble in the rotation of this moon, which is one of the reasons why this large volume of water was discovered in the first place. Clearly Enceladus is great for ice-skating, but probably not for sustaining land-based life similar to humans.

Figure 1: From left to right, Venus, Earth and Mars. Which one would you like to live on? Source: ESA (link) .

So how did we on Earth get so lucky?

It turns out that this depends on which scientist you ask. There are two theories:
Theory 1: The major building blocks of the Earth contained water from the start. This water then accumulated at the surface of our planet (by “degassing” from the mantle) and formed the oceans and the hydrosphere.
Theory 2: The major building blocks of the Earth were bone dry, and most of the water was delivered by comets and water-rich asteroids some time after most of our planet’s mass had formed by accretion.
So far, scientists do not have reasons to discard either of these theories, but there are two important arguments in favor of water being delivered after most of the planet had already formed:

Earth formed in a hot region of the solar system from which molecules with “low” condensation temperatures such as water had largely been removed before planetary accretion started (Albarède, 2009). Secondly, the ratio of heavy to light water in Earth’s oceans is similar to that of water in some comets and asteroids (Hartogh et al., 2011). Although you may not have noticed this when you last opened your water tap, a very small fraction (0.016 %) of the water on our planet is heavy, because it contains an extra neutron. The similarity in heavy water composition between asteroids and comets and Earth’s oceans does not prove that water on Earth was delivered by comets, but it certainly is consistent with this scenario. To make matters more complicated, however, the recent European space agency mission Rosetta to the water-rich comet 67P/Churyumov–Gerasimenko found that it has a very different ratio of heavy to light water than our oceans, which certainly complicates the debate.

Figure 2 Comet 67P/Churyumov-Gerasimenko losing water (and dust) as it gets closer to the sun. Source: ESA

Interestingly, neither theory can directly explain why our direct planetary neighbors, Mars and Venus, are so dry compared to Earth. So is it possible that these planets once were similar to Earth, and contained more water in their early days than that they do now?

Due to the high surface temperatures at Venus, any liquid water near the surface would immediately evaporate and diffuse into the atmosphere of the planet as a gas. We know that due to the lack of a protective magnetic field on Venus, solar winds continuously erode the atmosphere of the planet. If Venus had abundant water in the past, such erosion by solar winds would therefore have effectively stripped water from the planet’s atmosphere. Similar to Venus, Mars also does not have a protective magnetic field, but the temperatures and pressures at the Martian surface are significantly lower than at Venus’ surface, allowing water to be present at the surface as ice. In fact, Mars may have had a denser atmosphere in the past that allowed liquid water to be present at the surface. Nowadays, erosional features such as channels are the dry witnesses that water indeed once occurred as a liquid on the surface of the planet.

Figure 3. Dry channels (in inverted relief) in the Eberswalde delta on Mars as seen by NASA’s Mars Global Surveyor (link)

As a summary, we have an idea on why our planet was lucky enough to keep large amounts of water compared to Venus and Mars. However, do we know how our planet got its water in the first place? Unfortunately we are still not sure. There is hope though: we keep getting closer to the answer thanks to recent research on the composition of water on our planet and comets and asteroids in the solar system. So stay tuned, there’s a good chance that science will be able to answer this question in the coming years…

Hartogh, P. et al. (2011), Ocean-like water in the Jupiter-family comet 103P/Hartley 2, Nature, 478(7368), 218–220.
Albarède, F. (2009). Volatile accretion history of the terrestrial planets and dynamic implications. Nature, 461(7268), 1227-1233.

Crop kites

Crop kites

Post by WaterUnderground contributor Mikhail Smilovic. Mikhail is a PhD  candidate in the Department of Civil Engineering at McGill University, in Quebec.

Crops use water for photosynthesis, absorbing nutrients, and transpiration, or the plant-equivalent of sweating. A crop may experience water-stress if the soil surrounding the roots is not adequately wet, and this stress will affect the crop differently depending on the crop’s stage of growth. Irrigation is the watering of plants to ultimately avoid such water-stress.

Non-irrigated crops are more vulnerable to intervals of dry and hot weather, and the increasing unpredictability of a changing climate will further complicate other crop management tools, such as choosing different cultivars (the particular variety of crop, some which may deal with certain stresses in an improved way) or changing planting dates.

Irrigated crops do not experience water stress (they may in fact experience water stress under a non-perfect irrigation system, but forgive this for now), but the water is necessarily derived from somewhere else. This somewhere else may also experience water withdrawals from municipalities, industry, and other agriculture. The source of water may be underground, or water from a river, lake, or spring, but a connection between both underground and surface waters shares with us that water removed from a system somewhere will have a response somewhere. This somewhere may very well be an ecosystem. Irrigation may also be costly related to the abstraction, transportation, and on-farm distribution.

Between non-irrigated and irrigated is a curious place where we can increase the resiliency of our agricultural systems to periods of drought and heat with limited irrigation, while allowing crops to experience well-timed water stress. Agricultural productivity or yield is determined as the amount of crop produced per area of land, say 3 tons/hectare for wheat. When water is a limiting factor, we would be sensible to also consider water productivity, that is the ratio of crop yield and water use, or, the amount of crop produced per drop of water. The practices of limited irrigation, also known as supplemental or deficit irrigation, makes an effort to increase this water productivity.

This space in-between non-irrigated and irrigated, however, has been often poorly explored or simplified. Crop kites is a novel tool to determine and quantify the potential agricultural and water productivity associated with different irrigation practices. This is important for regions interested in shifting investments into or away from irrigation, as well as for researchers interested in evaluating limited irrigation practices as initiatives to establish food and water security, both currently and with changing climates.

A first thought might be, if a crop uses three quarters of the water than it would under ideal conditions, does the crop produce three quarters as much as the crop under ideal conditions? In fact, the answer depends very much on when this water is used.

Let us take the example of winter wheat in northern Africa. Winter wheat can be broadly characterized into five different growth stages. We can illustrate water use throughout the season with the following figure:

Water use is represented by the bottom blue colour, and the associated deficit is represented with the upper orange colour – the top line of the shape is the amount of water the crop would use under ideal conditions on the associated day. This example shows a 0, 10, 20, 30, and 40% deficit occurring in stages 1 to 5 respectively, representing a 78% water use across the entire growing season as compared to ideal conditions. Understanding both the amount of water used and when the water was used, we are able to determine the associated yield, for this example, we reach 68% of potential yield.

Now, what if we were to simulate the yield using all reasonable water uses and all reasonable distributions of the timing of this water use? The resulting shape is our crop kite, with each point associated with a water use distributed throughout the growing season in a particular way:


This shape illustrates the incredible range of yields associated with each water use; for example, 80% of potential water use relates to between ~20 and 90% of potential crop yield.

Water distributed through canals are often delivered according to a schedule, and not necessarily related to growth-stage sensitivities or actual weather. From the crop kite we can derive estimates on how the crop yield will be affected by adopting certain irrigation schedules. We elaborate on this with three examples: S1) water use is distributed to optimize yield, S2) the deficit is distributed evenly across all growth stages, S3) water is used preferentially for the earlier growth stages. The resulting crop-water production functions are illustrated in the following figure:


Although the first schedule optimizing for crop yield may be in line with the motivations of the irrigating farmer, it is often an unreasonable assumption for farmers delivered water according to predetermined schedules, but may be appropriate for farmers irrigating with a privately owned well. Evaluating the potential of supplemental irrigation necessitates estimating the ability of farmers to manage both the amount and timing of irrigation applications. Otherwise, non-reasonable assumptions may be used to evaluate and over promise estimates for agricultural production, with the fault not in the practice of limited irrigation, but in the criteria used to evaluate the system.

Crop kites demonstrate the wide range of water use-crop yield relationships, and can be used to evaluate the potential of limited irrigation to shift both food and water security.


Mikhail Smilovic is a PhD candidate at McGill University and the University of Victoria . Mikhail’s work investigates the interplay between foot security, water resources, and energy, and evaluating and integrating initiatives that increase agricultural production while reducing demands on water resources.

Limits to global groundwater use

Limits to global groundwater use

Post by WaterUnderground contributor Inge de Graaf. Inge is a postdoc fellow at Colorado School of Mines, in the USA.

Groundwater is the world’s most important source of freshwater. It supplies 2 billion people with drinking water and is used for irrigation of the largest share of the world’s food supply.

However, in many regions around the world, groundwater reserves are depleting as the resource is being pumped faster than it is being renewed by rain infiltrating through the soil. Additionally, in many cases, we are still clueless about how long we can keep drawing down these water reserves before groundwater depletion will have devastating impacts on environmental and socio-economic systems. Indeed, these devastating effects are already being observed.

The most direct effect of groundwater depletion is the decline in groundwater levels. As a direct impact, groundwater-pumping cost will increase, so too will the cost of well replacement and the cost of deepening wells. One of the indirect consequences of declining water levels is land subsidence, which is the gradual sinking of the surface. In many coastal and delta cities, increased flooding results in damages totaling billions of dollars per year. Next to this, declining groundwater levels lead to a decrease in groundwater discharge to rivers, wetlands, and lakes, resulting in rivers running dry, wetlands that are no longer sustained, and groundwater-dependent ecosystems that are harmed.

Over the past decades, global groundwater demands have more than doubled. These demands will continue to increase due to population growth and climate change.

The increase in demands and the aforementioned negative effects of groundwater depletion raise the urgent question: at what time in future are the limits to global groundwater use reached? This is when and where groundwater levels drop to a level where groundwater becomes unattainable for abstraction, or that groundwater baseflows no longer sustain river discharges.

In my PhD research, I predicted where and when we will reach these limits of groundwater consumption worldwide. I defended my dissertation last year April at Utrecht University, in the Netherlands.

Where and when are the limits reached?

Results show that many large aquifer systems are already highly depleted, especially for intensively irrigated areas in dryer regions of the world, like India, Pakistan, Mid West USA, and Mexico (see Figure 1). New areas experiencing groundwater depletion will develop in the near future, such as Eastern Europe and Africa. Future predictions show that some areas, like the Central Valley, and the High Plains Aquifer, partly recover when more recharge will becomes available. Notwithstanding, environmental groundwater demands will increase as to buffer more irregular streamflow occurrences due to climate change.

Figure 1: Estimated groundwater depletion (1960-2010) in [m], masked for aquifer areas, and zooms for hotspot regions, which are the intensively irrigated regions of the world.

In 2010, about 20% if the world population lived in groundwater depleted regions, where groundwater dropped below the economical exploitable limit. As a rule of thumb: the economic limit is reached when groundwater becomes unattainable for a local farmer, which is approximately when the water level drops to 100 m below the surface. In 2050, 26% to 36% of the world’s population will live in areas where the economic exploitable limit is reached (see Figure 2). Evidently, this persistence and increasing level of groundwater stress will impair local development and generate tension within the global socio-economic system.

Figure 2: First time that groundwater falls below the 100m limit.


Global-scale simulations

To answer my main question, I studied the effects of groundwater abstractions on river low flows and groundwater levels worldwide, as well as which trends in river low flow frequency and groundwater level change can be attributed to groundwater abstractions.

I used a newly developed physically based surface water-groundwater model to simulate i.a. river flows, lateral groundwater flow, and groundwater-surface water interactions at a high resolution (approx. 10×10 km) at the global scale. Total water demands were estimated and account for agricultural, industrial, and domestic demands. I simulated groundwater and surface water abstractions based on the availability of the resource, making the estimate reliable for future projections under climate change and for data-poor regions where we do not know how much groundwater or surface water is abstracted. Next, I developed a global-scale groundwater model. I estimated alluvial aquifer thickness worldwide, as no data at the global scale is available (see Figure 3). Aquifer thickness is one of the parameters you need to estimate groundwater flow and storage.

Figure 3: Estimated alluvial aquifer thickness. White areas are mountain regions, where no aquifers are simulated.

Simulations were done for the recent past and near future (1960-2050) and the results include maps and trends of groundwater heads, groundwater fluctuations, and river discharges.

In conclusion, most of our water reserves are hidden underground and most of our groundwater abstractions rates exceed groundwater renewing rates, leading to depletion. The growing demand and the expected climate change bring our groundwater reserves under mounting pressure. More than two-thirds of all abstracted groundwater is used for food production. Every year the world’s population is growing by 83 million people.

Improving our knowledge about how much water we can use in the near future while avoiding negative environmental and socio-economic impacts is therefore extremely important. A study like this contributes to the knowledge gap and can help guide towards sustainable water use worldwide to overcome potential political water conflicts and reduce potential socio-economic friction, as well as to secure future food production.

Want to read more? Check out the recent AGU press release or if you have more time… read my papers on dynamic water allocation (click here), development of a global groundwater model (click here, or here), or read my PhD thesis (here).


Author Inge de Graaf receiving her PhD degree from her advisor, professor Marc Bierkens (at Utrecht University, Netherlands). Note Tom Gleeson’s bald head in the lower left…


Get every new post on this blog delivered to your Inbox.

Join other followers: