Wednesday, October 28, 2009

Cambrian Explosion: This Time Its The Calcium That Did It

Geological processes and evolution - 4

A short article in Science Daily describes research that suggests that a buildup of calcium in the oceans in the early Cambrian provided the necessary trigger for the Cambrian "explosion" - the pronounced expansion and diversification of multicellular groups of animals. Some of the best examples of this evolutionary radiation is preserved in deposits like the Burgess Shale and the Chengjiang strata.

Here is the one sentence summary-

The researchers succeeded to show that the massive and sudden surge in the calcium concentration of the Cambrian seawater -- that is believed to be the result of volcanically active mid-ocean ridges -- not only initiated the buildup of calcified shells, but was also mandatory for the aggregation and stabilization of multicellular sponge structures. This allows, on the other hand, to formulate a novel theory where the geologically induced increase of marine calcium might be the key for understanding the Cambrian Explosion of Life.

There is deja vu when I read another ultimate causative explanation for the Cambrian explosion. Over the years you could write a similar sentence but substitute the word calcium with the sudden increase in oxygen, the warming of the earth following the thawing of the snowball earth, the increase in shallow shelf areas following marine transgression, all proposed as a one point explanations of this evolutionary phase in earth history.

Reading the press release you get a sense that the Cambrian explosion coincided with the origin of multicelluarity. Take this sentence:

However, the causes of its origin have been the subject of debate for decades, and the question of what was the trigger for the single cell microorganisms to assemble and organize into multicellular organisms has remained unanswered until now.

Its important to make a distinction here between the origin of a system and its subsequent diversification. Groups like fungi, plants and animals have evolved multicellularity independently of each other.  Whatever triggered single celled animals to assemble into multicellular ones (metazoans), that transition did not happen in the Cambrian but much before in the late Proterozoic maybe as much as 50 -100 million years before.

The original study recognizes this point but it gets lost in the press release.

There is a good and improving fossil record of metazoans from the late Proterozoic Ediacaran fauna and other types of fossil preservation that indicates that  molecular mechanisms for cell to cell signaling and cell adhesion must have evolved well before the Cambrian. Multicellularity in animals originated long before the Cambrian explosion occurred and long before the rise of calcium in Cambrian sea-water.

The figure below summarizes the current fossil record of the evolution of metazoans. The Cambrian "explosion" corresponds with the Chengjiang fauna.

 Source: On The Origin of Phyla

Explanations of the origin of metazoans and subsequent evolutionary radiations like the Cambrian explosion involve long and intertwined causal chains. A rise in oxygen by late Proterozoic may have favored larger body size and one solution to achieve this large size was to aggregate into colonies of cells which eventually became integrated as one organism. The marine transgressions by late-Proterozoic early Cambrian expanded available shallow marine shelf areas and created large and diverse ecologic niches for evolutionary diversification to take place. The evolution of predation would have set forth selective pressures for skeletonization.  Many organisms would have taken opportunistic advantage of the rise of calcium in sea-water to boost skeletal production. Likewise calcium may have provided stability for even larger masses of cells to aggregate.

There were many geological and ecological factors at play feeding of each other that were responsible for one of the major transitions in the history of life. Insisting on just one cause as the most important to the exclusion of others is too simplistic.

Saturday, October 24, 2009

The Latest Numbers On Arctic Oil and Gas Potential

Some interesting readings I came across on hydrocarbon resources and challenges over the last few days. has published a report by the Energy Information Administration on the latest estimates of undiscovered technically recoverable oil and gas resources of the Arctic basins.

And the number is large - about 400 billion barrels of oil equivalent or about 20 odd percent of the world's undiscovered resources. This is an estimate based on occurrence of geologically favorable conditions in various Arctic basins. These are resources or potential. They will be or rather a fraction of these will be added to reserves only when someone drills and more directly estimates the amount of hydrocarbons that can be economically recovered.

We won't be running out of oil and gas or other fossil fuels like coal soon,  geologically, and BP's chief executive officer Tony Hayward thinks that in 2030 fossil fuels will still be meeting 80% of the world's energy needs.  He explains his position in a speech given at the Oil and Money conference, London. The title of the conference conjures up images of greedy petro-oligarchs doing everything to maintain a vice-like grip on the supremacy of fossil fuels, but the changeover to renewables won't be easy given the enormous gap that exists between the contributions from fossil fuels and renewables to our energy mix. India for example generates 70% of its electricity from coal and has plans to build plenty of coal fired power plants in the near future. The contribution of solar and wind to power generation in India is currently negligible. The U.S generates about 46% of power from coal and the contribution of non-hydro renewables to power generation is just 3%.

If fossil fuels are going to be an important if slowly declining part of our energy mix for some time to come then Geoffrey Styles of Energy Outlook argues that we rethink our reluctance to pursue low emissions strategies like carbon capture and sequestration (CCS). He writes on not just the technological and economic challenges facing CCS but also a public backlash against it based on  - he thinks - a lack of education among the public about geological principles and the efficacy and safety of CCS.

Plenty to think about as the Copenhagen climate change summit nears.

Wednesday, October 21, 2009

How Much Oil Underneath?

A friend asked me a few days ago why the Karnataka government were giving two different estimates for iron ore potential in the state. One file says that the iron ore reserves are about 3,447 million tonnes, while another department file says that the state has about 9,000 million tonnes of iron ore resources.

I replied that assuming the Karnataka government is using the words as they are commonly used in industry, reserves are the fraction of resources that can be economically exploited at any given time. Neither is a static quantity. Both will change given new finds, technological breakthroughs that enable recovery of previously out of bound deposits and the economic and political climate.

Coincidentally Scientific American has an article in their recent issue on the current reserves of oil and how that quantity is changing with new finds and new technology. They give one example:

When Kern River Oil Field was discovered in 1899, analysts thought that only 10 percent of its unusually viscous crude could be recovered. In 1942, after more than four decades of modest production, the field was estimated to still hold 54 million barrels of recoverable oil, a fraction of the 278 million barrels already recovered. “In the next 44 years, it produced not 54 [million barrels] but 736 million barrels, and it had another 970 million barrels remaining,” energy guru Morris Adelman noted in 1995. But even this estimate proved wrong. In November 2007 U.S. oil giant Chevron, by then the field’s operator, announced that cumulative production had reached two billion barrels. Today Kern River still puts out nearly 80,000 barrels per day, and the state of California estimates its remaining reserves to be about 627 million barrels.

This story will apply to a wide range of mineral/oil deposits all over the world.  Famously the United States has produced a cumulative 200 billion barrels of oil from reserves that never at any one time exceeded 40 billion barrels.

In the many responses to the Scientific American article I thought this one from JR Wakefield stood out as it explains what peak oil means from different perspectives:

Its not about how much oil is in the ground, it's how fast you can get it out and at what net energy.  This thus article is highly misleading.   Here are the Five Horsemen of Peak Oil:

1) Geological Peak. That is the point where we have consumed half the oil in the ground. So far we have consumed a trillion barrels. Estimates of remaining oil range, but the number appears to be 3 trillion barrels remaining in the ground. So we are not at geological peak. Hence skeptics of peak oil use this for their arguments, like this SA article does.

2) Flow Rate Peak. That's the point at which you cannot extract the oil fast enough to meet demand. This is especially so with old fields in decline (which is a fact) and new fields which have difficult geology (like this one).  The flow rate from them does not keep up with decline, nor keep up with growing demand. The article failed to mention that North Sea is all in terminal decline and the UK has to now import oil. Indonesia peaked years ago and has to import oil forcing them out of OPEC.  The Cantarell field in Mexico, the third largest in the world, and the US's 3rd import source, was producing 2.3mb/day at it's height.  Today it's 560kb/day with a 41% drop from last year.  WE ARE AT FLOW RATE PEAK NOW.

3) Geopolitical Peak. That's when exporting countries, due to their own growing demand, decide not to sell their oil abroad any longer but decide to keep what's in the ground for their own future domestic needs. So far only the US does this, but expect other countries to soon follow that.

4) ERoEI peak. This is the point at which it takes as many joules to extract the oil than you get from the oil extracted. That is, one barrel in to get 1 barrel out. Conventional wells in the 1960s were 100:1. That has dropped to about 25:1 today. Aging fields and new unconventional fields have very low ERoEI. The tar sands in Alberta for example is less than 6:1. Our entire society is based on the NET energy, not what's extractable. Calculations show that we will reach over all break even in oil extraction between 2020 and 2030. Once that is reached it basically means we have completely run out of oil.

5) economic peak.  This is the point where the economy cannot tollerate high oil prices and plunges the world into a recession, like this one which was caused by $140.barrel oil.

The challenge before us in terms of combating global warming is to ensure a transition from hydrocarbon energy to low emissions renewable energy much much before we start running out of the stuff not just in geological terms - which is not going to occur anytime soon - but by the other measures like Flow Rate and ERoEI as well. Likely even the Flow Rate Peak and the ERoEI peak are not constant but will keep shifting as more efficient ways of extracting oil and other hydrocarbon resources are discovered.

Monday, October 19, 2009

The Probability Of Evolutionary Pathways

Joe Thornton of the University of Oregon and the Howard Hughes Medical Institute writes these thoughts about how to understand and appreciate the probability of evolution following one specific pathway among many possible:

Consider the future: there are countless possible that could emerge from our present state, making the probability of the one that actually does evolve extraordinarily  low.  Does this mean that the future state that will ultimately emerge is  impossible?  Obviously not.  To say that our present biology did not evolve  deterministically means simply that other states could have evolved instead; it does not imply that it did not evolve.

Consider your own life history as an analogy.  We can all look back at the road  we have traveled and identify chance events that had profound effects on how our lives turned out.  “If the movie I wanted to see that night when I was 25 hadn’t been sold out,  I never would have gone to that party at my friend’s house, where I met my future spouse….”  Everyone can tell a story like this.  The probability of the life we actually lead is extraordinarily small.  That obviously doesn’t mean that its historical unfolding was impossible.

That we inhabit an improbably reality requires a divine explanation only if we, like Behe, take the teleological view that this is the only reality that could exist.  But if we recognize that the present is one of  many possibilities, then there is no difficulty reconciling the nature of  evolutionary processes with the complexity of biological forms. As history unfolds, potential pathways to different futures are constantly opening and  closing. Darwinian processes are entirely adequate to move living forms  along these pathways to a remarkable realization – but just one realization out of many others that could have, but didn’t, take place.

Beautifully written.

The entire article is worth reading as it explains how complex traits with interlocking components can evolve through a combination of natural selection and random genetic drift acting on adaptive mutations and neutral intermediates. And this is not just theory. Joe Thornton leads a team that does experiments to show how this can happen. The article encapsulates modern evolutionary thinking about the evolution of complexity quite well.

That it is a devastating put down of the arguments made by the Intelligent Design community makes it especially pleasurable to read. 

Friday, October 16, 2009

A 20 Million Year History Of Atmospheric CO2

From Brave Blue Words I found out a day late that yesterday was Blog Action Day for Climate Change. Bloggers all over the world are writing about various aspects of climate change.

Being a geologist I want to point to a study that reconstructs atmospheric CO2 levels as far back as the Miocene  - a 20 million year history. Atmospheric CO2 levels have been reconstructed with some confidence for the last 800,000 years or so using gas bubbles trapped in the Antarctic ice sheets. Before that the data was thin.

Aradhna Tripati and colleagues have used the boron to calcium ratio in foraminifera shells to calculate ancient CO2 levels. As atmospheric CO2 increases some of it diffuses into the ocean increasing the dissolved CO2 content of sea-water. That in turn reduces the amount of boron that is incorporated in a growing calcium carbonate shell of the foraminifer individual. The variation in the boron to calcium ratio over time as recorded in foraminfera fossils of different ages should tell us something about transitions in CO2 levels.

The scientists first validated their calculations using the 800 K record of CO2 trapped in ice. They compared their results with those obtained by the direct measurement of CO2 trapped in gas bubbles. It was a good match. The scientists calculate that the uncertainty in their results is about 14 parts per million.

Their results show that there is a close coupling between CO2 levels, sea-level and temperature over the last 20 million years.  In the middle Miocene (~ 20 ma) CO2 levels were about 400 ppm - comparable to modern levels - and that sea-levels at that time were 30 -40 meters higher than today (geologists estimate this using distribution of ancient shorelines), with temperatures about 3-6 deg C higher (using geochemical proxies like the oxygen isotope composition of shells which depend partly on temperature of the water from which they precipitate). Decreases in CO2 levels in the later part of Miocene and Pliocene were synchronous with major episodes of cooling and glacial expansion.

Its important to establish that historical connection to answer doubts expressed on what exact impact would increasing levels of CO2 have on climate and sea-level. Many climate change doubters are not happy with computer simulations and models of CO2 increase and climate change. This study shows that CO2 has been a strong driver and amplifier of climate change in the deep geological past. History is also a guide and often a reliable one.

Go here for the press release.A minor quibble. The press release calls the shells used by the scientists as belonging to single celled marine algae. Foraminifera are not algae. They are protists.

Wednesday, October 14, 2009

A Government Scientist Speaks Out On Flood Management

You don't hear many Indian Government scientists opining about science, policy, the environment, civic criticism...well you get the picture.

There are put it politely.

But Chetan Pandit of the Central Water Commission has broken the shackles and in the October issue of Current Science comes out sparring strongly against - in his opinion - the woolly headed arguments of the environmental community  and the media on the subject of big dams and flood control.

He puts forth several fallacies and myths regarding dams and flood control and argues very well that big dams have served as effective flood managers.

And I had to chuckle at this:

We Indians seem to be particularly susceptible to this ‘romancing with the past’. Everything, be it water management, or agriculture, Ayurveda, mathematics, literature, astrology, etc., we like to believe that in India all learning had reached its peak in some distant past, and the best thing for us to do is, to continue to do what our ancestor’s did. And this is given a lofty name ‘wisdom of the centuries’.

You may be inclined to view big dams as destroyers of forests and biodiversity and so on but Chetan Pandit has made some good points about the intellectual environment in which debates on these issues take place and are presented to the public in India. One of the most important points he makes is that many people who get involved in criticizing government projects are numerically challenged. Arguments are high on the emotional quotient but there is little quantitative analysis of data to go with it.

I am sure not all environmentalists are so mathematically naive, nor should every developmental project be reduced to just numbers. But this is a government scientists view. When presented with a  rare opportunity to speak out he doesn't restraint himself.

Read the article here.

Update: Chetan Pandit via an email to me wants to clarify:

I never said that those criticizing government projects are “numerically challenged”, or mathematically na├»ve, etc. Even a high school student would be able to import the data on flood affected area into a spread sheet and plot a graph to see whether or not there is any increasing trend. What happens is, if they do that their argument will collapse, the numbers do not favour them, they know it, and therefore they have to willfully ignore the numbers. Which is why I wouldn’t describe the environment in which debate takes place as “intellectual”. It is pseudo-intellectual.

Chetan Pandit

Sure, he didn't use the words numerically challenged and mathematically naive. Those are my words but he does imply this. For example he writes about noted anti-dam activist Shripad Dharmadhikary -

In his exhaustive critique of the Bhakra dam, noted anti-dam activist Shripad Dharmadhikary writes, ‘Even after the Sutluj flows were augmented by the transfer of Beas water into the Bhakra reservoir, the reservoir has not filled up in most of the years’4. This was probably intended as a critical comment on filling of the Bhakra reservoir. But its implication, which probably escaped Dharmadhikary, is – Bhakra is very successful in flood control. In the years, the dam did not even fill; it is obvious that all the floods were absorbed 100%. And this, despite transfer of a substantial quantity of water from Beas to Sutlej through the Beas–Sutlej link.

But its implication, which probably escaped Dharmadhikary,...

So I read this as meaning Dharmadhikary did not understand the topic well enough. It certainly doesn't read as meaning that Dharmadhikary churned out the numbers, didn't like what he saw and then ignored the finding.

Monday, October 12, 2009

Dinosaur Eggs And Some Stratigraphic Thoughts

The discovery of dinosaur eggs from Cretaceous fluvial sediments near the village of Ariyalur in the state of Tamil Nadu , South India is getting lots of press cover and here. Hundreds have been found in clusters of about 8 over an area of about 2 sq km. Looks likely to be a nesting site. On a sadder note I read in the Times of India a few days ago that there has been no protection given to the site by the government despite requests from the scientists. Locals are already taking away the fossils eggs and disturbing and damaging the site in the process. What a shame!

There also has been some silly press coverage calling this a Jurassic treasure trove, a  holdover of the popular link between anything dinosaur and the word in Jurassic Park the movie. This particular south Indian sedimentary basin does not have Jurassic sediments. It contains an Early Cretaceous to Early Cenozoic sequence.

What caught my eye was that the sedimentary layer containing the eggs were capped by a volcanic layer. The scientists from Periyar University, Salem, seem to think that this volcanism represents the Deccan volcanic activity dated to around 65 million years ago and that the field relationship between the volcanic layer and the underlying sediment could suggest that this particular volcanic event may have killed or damaged those eggs.

I don't know enough details about the deposit to answer that with certainty but I did have some random thoughts on event deposits i.e layers deposited almost instantaneously and how geologists use such deposits to ascertain the age of the associated sediments. In this case they suggest that sedimentation and the volcanism took place in quick succession and by quick I mean volcanism took place immediately after the dinosaurs laid those eggs....before those eggs hatched.

If a sediment unit is capped by a volcanic layer that is dated to say 65 ma (million years old) then that would mean that the sediment cannot be younger than 65 ma. But does the relationship mean that the sediment too is 65 ma? And what does it mean when you say the sediment is 65 ma. How much 65 ma plus minus ...years, or uncertainty is there in a calculation like this? 

Lets say we get lucky and that the sediment layer is sandwiched between two volcanic layers that can be dated with radiometric methods. Let's say both layers give a date of 65 ma. What does that tell you? Radiometric dates that old come with a sizable uncertainty on the order of hundred thousand years or so. That means the date of the overlying layer might come out as say 65 ma with an uncertainty of 300 thousand years. The date of the layer underlying the sediment may come out say 65.2 ma with an uncertainty of 300 thousand years. The two dates are statistically unresolvable. So, even if the sediment is sandwiched between two volcanic layers that indicate the same statistical date there could still be a time lag of tens to a hundred thousand years or so between the sediment being deposited and the volcanic activity.

Fossils are not much use either for this purpose.  Fossils can eliminate the possibility that the events took place in quick succession if the sediment contains fossils which are obviously much older that 65 ma. But again fossil species have temporal ranges of a hundred thousand years or more. Even if the sediment underlying the volcanic layer contains fossils that are indicative in this case of the latest Maastrichtian age (close to 65 ma) there will be an uncertainly of tens of thousands of years and so their presence won't resolve events taking place on smaller time scales.

Geologists have then to rely on the detailed physical relationship within and between the sediment and the volcanic material.

Preserved sedimentary layers are mostly time averaged deposits. That means that material at the bottom of a particular bed is not necessarily older than the material in the upper part of the same layer. During deposition waves and currents keep reworking the same bundle of sediment. Animals may burrow into it and churn up sediment. Material at the bottom part may get transported to the upper part of the layer. The layers thus becomes time-averaged. Organisms who have lived at different times through that depositional episode are all distributed randomly - with respect to their age - throughout the deposit.

A good recent example of a time averaged deposit is the sediment unit that contains the remains of Ardipithecus ramidus the early hominin found in Ethiopia. Scientists lucked out there and realized that the sediment is sandwiched between two volcanic layers dated to about 4.4 ma. Each layer differed in their radiometric age by about 30 thousand years with an uncertainty of about 75 thousand years. Based on this information and the internal characters of the layers the scientists concluded that the geological unit was a time averaged deposit representing a few thousand years of deposition.

This layer containing the dinosaur eggs does not have the characters of a time averaged deposit at least not one representing hundreds or thousands of years. The eggs were found in clusters of 7-8 eggs per nest and these nests are found in successive layers of the sequence. These two indicators suggest that the sediment was not disturbed much. Rather the setting, fluvial floodplains, would have meant that the eggs were buried rapidly and entombed during seasonal floods, a sort of an event deposit just like volcanic eruptions. That may have been the main cause of the eggs not hatching. The layers immediately underlying the volcanic cap may thus represent sedimentation taking place over a few tens of years, each nest bearing layer essentially preserving or freezing ecosystem conditions as they existed at that time. There is the possibility that the uppermost layer containing eggs represents maybe the last egg laying season before the eruption.

Volcanism when it occurred would have sealed the deposit from further damage from the elements. How close in time was that event to the last egg laying season? Again the scientists will have to look closely at the relationships. Are the eggs in the uppermost layers caked with volcanic ash? Do they show signs of being baked or cracked due to heat, a kind of a Cretaceous hard boiled feast?

Detailed sedimentology and stratigraphy will provide more clarity than absolute radiometric dating and fossil ranges when posed with questions of this nature.

Saturday, October 10, 2009

Its Hard To Classify That Archaeopteryx

How birdlike was Archaeopteryx?

Beak - Yes
Feathers - Yes
Wishbone - Yes

Growth pattern of blood vessels - Not quite there yet

By comparing the structure of the bone of Archaeopteryx with modern birds and fossil dinosaurs, researchers found out the growth pattern of blood vessels in Archaeopteryx was more like dinosaurs than modern birds. Archaeopteryx individuals seem to have taken a longer time to mature than modern birds do.

Archaeopteryx lived about 150 million years ago and is the earliest bird-like creature to have been found yet.  A younger bird-like fossil -Ichthyornis dispar - dated about 94 million years ago shows several characteristics of quick growth, giving us an idea of the timing of the various changes in morphology and physiology that were taking place within the Maniraptor dinosaurian clade as they evolved  characteristics that are recognized as bird-like.

By the way this is not some earth shattering find. The basic relationship that  small carnivorous dinosaurs evolved into birds still holds. But the study highlights how difficult categorizing a creature which is a composite of ancestral and derived traits can be.

It doesn't matter in the larger view if Archaeopteryx is classified as a dinosaur or a bird. The value is in demonstrating that evolution of a complex suite of characters takes place in fits and starts, some features evolve earlier than others preserving the historical trajectories that major transitions in evolution have taken.

Go here for the lighter version.

Tuesday, October 6, 2009

Physicist Extraordinaire- Science Friday On A New Book On Paul Dirac

Science Friday has a really interesting talk with Graham Farmelo on his new book on Paul Dirac - The Strangest Man: The Hidden Life of Paul Dirac, Mystic of the Atom.

What's my connection to this great man? Later in his career Paul Dirac joined Florida State University, Tallahassee which was my Alma mater too. I walked often past his statue on the way to the Dirac Science Library. 

Farmelo gives a good account of Dirac in that talk as a reticent personality, who avoided people...but not unaware of the importance of knowing the right people....he did make sure he was standing right behind Einstein in a group photo at a conference.

Meanwhile from my book shelf in one of the best science autobiographies I have read- Sir Fred Hoyle's vibrant and entertaining Home Is Where The Wind Blows- I found the following nuggets about Hoyle's interaction with Dirac:

Shortly after inveigling me into the secretaryship of the Delta-Squared V Club, Maurice Pryce left Cambridge to take up a lectureship at Liverpool University. If I was to retain my student status with the Inland Revenue, another research supervisor was therefore needed. Pryce suggested Dirac, so I went to Dirac and explained the situation. Although normally he didn't accept  students, Dirac broke his rule on this occasion because he simply couldn't resist the circular counterlogic of a supervisor who didn't want a research student who didn't want a supervisor.


More than any other person I have known, Dirac raised the meaning of words and syntax to a level of precision that was mathematical in its accuracy. He had nothing at all of the irritating habit of attempting to read hidden significance  into your remarks. He paid everybody the compliment of supposing they knew exactly what they were saying, a compliment he sometimes took to extreme lengths.

how great was Dirac..? truly great minds he is posthumously productive...  says Farmelo.

Sunday, October 4, 2009

Indian Groundwater Extraction May Be Contributing To Sea Level Rise

I don't know what to make of this calculation which I picked up in a New Scientist story. A few weeks ago there was a study using NASA's Grace satellite measurements that showed an increase in groundwater extraction from North Indian aquifers.

A second study on these satellite measurements asserts that the groundwater loss amounts to about 54 cubic km per year over a time period of 2002 and 2008. A lot of this extracted groundwater ends up in the sea and could be contributing to raising sea-levels by 0.16 millimeters every year, about 5% of the total sea level rise. That is about the same as contributed by runoff from melting Alaskan glaciers the authors conclude.

I don't have access to the full paper so I don't know the details of the calculation but here is what the scientists have to take into account:

Part of the groundwater extracted will be taken up by plants and remain there over the life of the plant and make its way into the food chain.
Part of it will be lost through the plants through evapo-transpiration.
Part of the water will remain in soil adhering to clay and sand particles.
Part of it will make its way back to the aquifer.
Part of it will make its way through the soil to local streams and eventually to the sea.
Part of it will be lost by direct evaporation. That evaporated water (and the water lost by evapo-transpiration) will fall as rain and part of it will infiltrate as groundwater and part of it will be surficial runoff into streams and eventually into the sea.

Just giving you  something to think about what happens to extracted groundwater.