When the Rivers Run Dry by Fred Pearce

When the Rivers Run Dry is a kind of modern, global Cadillac Desert, looking at present and future water issues around the world.  I think in the end it was too ambitious, looking at too many individual situations superficially, without going into the details on how they came to be the way they are (which Cadillac Desert was able to do, since it focused only on the American West), and also without drawing enough insightful generalizations from the many different cases the author studied.  It ended up feeling mostly like a dreary litany of mistakes painstakingly repeated in nation after nation, decade after decade, apparently without any learning going on.  Often these projects were funded by the World Bank and other international “aid” organizations, or by powerful central governments.  In both cases, the motivations often turned out to be short sighted and political or financial and had little to do with good engineering, productive agriculture, fisheries, or long term stability.

The most incredible river in the book has to be the Tonle Sap, a tributary of the Mekong in Cambodia.  The Mekong floods so high in the monsoon season that the flow of the Tonle Sap reverses direction, filling up an enormous shallow lake of the same name, which serves as a massive fish breeding ground, and a highly productive ecosystem.  There are whole floating villages on the lake with populations of more than 10,000 people.  This annual cycle will be destroyed shortly by dams high on the Mekong in China/Tibet which will moderate the river’s annual floods to the point where the Tonle Sap may no longer reverse its flow.

The tales of water in Israel and Palestine were painful.  The entire Jordan river is diverted to Israel at the moment, at a cost (mostly in pumping) that rivals the cost of producing desalinated seawater.  Israel has usurped virtually all the rights to the aquifer lying beneath the West Bank, and wants the international community to pay for a desalinization plant on the coast, to provide water to replace it (which of course they would still have the power to cut off if they chose to).

The tens of thousands of wells in Bangladesh and eastern India, poisoned with fluorine and arsenic.  The south to north water transfer project in China, which rivals in scale the plans which we thankfully aborted to re-direct the Columbia to SoCal.  100 km long tunnels through bedrock.  Egypt defiantly building the Aswan High Dam right where the greatest amount of water would be lost to evaporation, instead of high in the mountains of Ethiopia (and threatening to wage war if any diversion takes place beyond their borders) but at the same time trying to drain the marshes in Sudan, using as their excuse that too much water (1/3 of what’s lost from Lake Nasser) evaporates there.  Salt and pesticide filled toxic dust storms in the playa which now lies where the Aral Sea was only 30 years ago, all in the name of cheap cotton, which will stop growing in another couple of decades because of the salinized soils.  Flood controls dams which seem only to control moderate floods, while making really great floods horrific: there’s no flood like the one that comes from a burst dam (just ask the giant ground sloths of the pacific northwest).

Since digging into the detailed histories and narratives of all these disasters would have taken a 20 volume series that nobody but policy wonks would have read, really what I would have liked is something more like Jared Diamond’s Collapse, in which the author does a lot of research, and then makes generalizations, and uses particular case studies to support those generalizations, laying out what the basic problems and dynamics are, what the potential solutions look like, and what the most imposing barriers to implementation are.

In outline form I think such a book would have looked something like this:

Basic problems:

• Salinization of soils: when water is lost from soil through evaporation, or when the solubility of minerals in ground water decreases, salts are deposited.  This isn’t really a problem with rain, because rainwater doesn’t have a lot of dissolved minerals in it.  It can be very bad with aquifer water, and some river water (like the Colorado).  This means, unfortunately, that all irrigated agriculture has a finite lifetime.  If we’re smart, we’ll try and get as much food as possible out of that finite lifetime, and that means using as little water as we can get away with per unit food produced.
• Evaporative loss: Open canals and reservoirs in dry regions waste vast quantities of water through evaporation.  Lake Mead, for instance, loses about a cubic kilometer of water each year to the sky (and produces about 4 billion kWh of electricity… which means 70 gallons of water per kWh!).
• Dam-based flood control: Dams, it turns out, don’t work for flood control.  For irrigation and power generation purposes, dam operators have incentives to keep reservoirs as full as possible.  This allows them to successfully control most floods, which are moderate.  However, serious floods and full reservoirs in combination end up being catastrophic.  In order to protect dams from failure when the water overtops them, operators have to release the flood waters all at once, faster than they would have flowed naturally.  And sometimes, the dams fail anyway.  In any case, dams generally end up only controlling moderate floods, and making bad floods much, much worse, especially when those downstream have been told by the government that the floodplain is now safe, and lulled into a sense of security by the lack of moderate floods.  This another instance of “picking up nickels in front of steamrollers”.
• Accelerated channel cutting: Often channelization of rivers for the purpose of flood control ends up increasing flow rates, and preventing the deposition of silt in the bed, which results in increased channel cutting.  This can create hydrological problems downstream when, for instance, the bed of the river ends up below the level of the fields it’s supposed to be irrigating.
• Aquifer depletion: Since the introduction of cheap and efficient pumps, the world has gone on an aquifer-based irrigation spree, often drawing down fossil water supplies deposited long ago, even during the last ice age.  We are “mining” water.  This ends badly, and quickly, when the water runs out.
• Loss of alpine glaciers: The storage and flow buffering capacity of alpine glaciers (esp. in the Himalayas, other high central Asian mountains, and the Andes) are utilized, free of charge, by between one and two billion people.  These glaciers are rapidly melting, and thus temporarily increasing river flows.  This encourages greater (and more inefficient) water use.  When the glaciers are gone, many regions will alternate between floods and drought, which is lousy for civilization.  Building dams will probably be the “solution”.  They will silt up, and exacerbate the worst of the floods, while reducing the amount of water actually available for use, through evaporative loss.
• Water rights disputes: Interjurisdiction transfers of fresh water, either downstream in rivers, or via border-spanning lakes and aquifers, are going to be increasingly serious points of contention.  Basically, it sucks to live downstream.  This is an inherently human (legal, political) problem, not a natural one.  As water supplies become more scarce, and demand continues to increase, this will get “interesting”.
• Shadow trade in “virtual” water: Vast quantities of water are shipped across borders, mainly in the form of agricultural commodities, especially grains and cotton.  There are water importing and water exporting nations.  Except nobody thinks about it that way, because all of the water is embodied in a product.  For instance, a one pound steak, produced using feed lots and grain-fed cattle in the US, requires 1,500 gallons of water to produce.  Water projects often end up subsidizing these industries to a breathtaking degree, but there are enough degrees of separation between the players involved that it is often not a political issue the public can be interested in.  This is another human problem.

Potential solutions:

• Stop using flood irrigation. Flood irrigation is a long slow train wreck.  It will eventually end agriculture as we know it, all over the world.  It is wasteful, in that a large quantity of water is lost to evaporation and only a small proportion ends up going to the crops (though it does in some cases end up recharging aquifers… unintentionally), More seriously, it brings salt.  Vast, civilization ending quantities of salt, in direct proportion to the water lost to evaporation.  All irrigation using transported water that has been in contact with the earth does this, but flood irrigation is the worst.  It is also the cheapest (when water is free) and most widely used method.  This is terrifying, because it cannot be undone.  All of the other problems can potentially be worked around, or recovered from (excepting perhaps the loss of alpine glaciers), but the salt will never leave.  We are squandering the agricultural productivity of the world’s arable land by using tens or hundreds of times more water than is necessary to irrigate our crops, and depositing tens or hundreds of times more salt per crop yield than we have to.  This is the kind of thing that will make future generations will look back and just shake their head.  We need to transition to drip irrigation, and that probably means charging more for water, and using more human labor and less mechanization.
• Stop building dams. They create more problems than they solve overall.  This might best start with the World Bank.  The only dams and reservoirs we should consider keeping are the ones that are deep, narrow, and at relatively high altitude, minimizing evaporative loss, and maximizing power production.
• Store and transport water underground. Surface transport and storage is a huge waste of water.  We should store our collected water in the aquifers we are depleting, and transport it underground, in covered canals, like the qanats developed by the Persians.  This will help with the salt problems too, by decreasing the concentrating effect of evaporation en route and in storage.
• Harvest rainwater locally. Rainwater can be harvested for local irrigation, domestic use, and aquifer recharge, at scale, using ponds, check dams, cisterns, and diverting flood waters into wells.  This also helps with salt (rain is essentially distilled water), and flood control.  Permeable pavement and urban swales can help cities deal with runoff in ways other than just sending all that fresh water to the sea: in LA, for instance, if we captured all the rain that fell on pavement, it would satisfy half the city’s current water demands.  If we landscaped intelligently and made moderate efforts to reduce domestic usage, that would mean we wouldn’t need any external water! Even desert agriculture is possible using harvested rainwater, by using runoff from an area larger than that which is cultivated (100,000 gallons of rain falls on each acre of the Negev in a year)
• Give rivers more space to flood. We like to think that we can control rivers and their flooding.  It turns out that this is largely a fantasy.  We can change the flood dynamics, either exacerbating or moderating them, but rivers will continue flooding.  They need the space to do so.  We need to work around it.  Floodplain agriculture is fine.  Floodplain condos, not so much.
• Re-locate population and agriculture, not rivers. Don’t move the Yangtze to N. China.  Move people south.  Or better yet, reduce the number of people.  “Food grows where water flows” is true, but doesn’t mean you should move the flow.  Move the farms.
• Use human and animal waste as irrigation and fertilizer. Conveniently, most cities produce a constant flow of nutrient and water laden sewage, right next to where a bunch of people need food to eat.  Instead of wasting that water and fertilizer, it should be actively and conscienciously used as safely as possible.