Irrigation Systems: Structured versus On-demand System Design

In case some of you think that all I am interested in is Vetiver, then you are wrong!! Last week I was visiting my old friend and Canadian colleague Brian Albinson, one of the best, experienced, and practical irrigation engineers that I have had the privilege to know and work with. We worked together in India in the 1980’s, and at that time he did much to get irrigation managers and engineers to see the advantages of the “structured system” of irrigation that had functioned very effectively in northern India and Pakistan since the 1870s. The “structured system” is all about sharing scarce irrigation water between farmers in a reliable fashion with minimum management.

The alternative,  the “on-demand system”, attempts to give farmers water when ever they need it. This system works in highly developed countries, but needs high level management, complex and costly equipment, and in times of water scarcity (an ever increasing problem at this time of climate change) regularly leads to disorder and vandalism. The “on-demand system” is often advocated by well trained engineering consultants, most with  little field experience in the day today management of irrigation schemes, and who often have vested interests in the more complex and more costly approaches.

Now read the following:

From Brian Albinson

Structured vs: Demand Systems

The late Leslie Shanan’s favorite epigram – “a structured system bus service or a demand system taxi service”. The analogy is pertinent but not exact, a well designed structured system can take you to nearly all the destinations served by a demand based system.

But we must pause for a reality check. The core of the problem is in the social, political and cultural differences between serving farms of 0.2 ha (Indonesia average) and 500 ha (typical of the developed World ), a ratio of well over 2000 to 1. Successful management concepts in one environment does not automatically transfer to another. In no other field is such a wide range of parameters so ignored. The dividing line where demand based irrigation can be sustained is dependent on a complex mix of the technical and society mores involved. Each new proposed project should be evaluated, in particular past irrigation performance should be given much weight.

The profession is polarized on the issues. The basic problem is a fundamental lack of reality experience – hardly any current irrigation engineer from the Western world has any experience of actually managing a small holder irrigation system in the developing world. I do not mean ‘adviser’, ‘consultant’, ‘expert’ or as visiting missionaries or NGOs. I mean farmers knocking at your door (where the buck stops) pleading – ‘my paddy has no water’. It is a different environment out there, the interplay of social, cultural, economic and technical issues is almost beyond the comprehension of anyone who has not actually experienced it.

Somewhere between an average holding of 0.2 and 500 ha demand based systems will not be sustainable. There will be a gray range where demand based systems may succeed. Many practitioners now accept this. There are still some who while accepting the concept is the best choice for current projects are concerned that 50 years down the road there will be a requirement for individual farmer demand based supply for some undefined exotic crops. This is a fallacy, the basic oil seed, grain and fibre crop proportion will not change, nor will their irrigation needs. Nor will the size of holding change, in fact under most cultures holding size will decrease.

Surely the only argument can be at what level of holding size do we (a) mandate a structured system; or (b) review the feasibility of a demand system?

The World Bank has shown the way. In India, Narmada, the largest new irrigation system in the World (1.6 M ha) is designed as a Structured System. Moreover Structured Systems used in millions of ha in N.W. India for the past 120 years have proved to be the most successful in India.

Perhaps some comments on how we got to the present unsatisfactory position would be in order. The huge expansion in irrigation after WWII was financed largely by the developed Western countries. Quite naturally the theory and technological practices developed and used highly successfully in the West became textbook norms which were transferred without hesitation to the developing countries of SE Asia and Equatorial Africa. The result was disastrous, the record is replete with failure. For 50 years the futile pursuit of the Holy Grail of demand irrigation – a hose pipe for every farmer – has been pursued with a single minded fanaticism. Western consultants, eager to please their home technical lobby simply preached more of the same – more controls, more computers, more automatic controls, more controlled volume canals more complexity, ignoring all the evidence to the contrary. The aid receivers, at first misled into believing they were getting the red hot cutting edge of technology, settled for as much funding as could be squeezed out of the donors. The poor farmers simply vanished into a technological fog which promoted the worst forms of exploitation.

The obsession with control and regulation had another unfortunate side effect. The profession completely ignored the question of system hydraulic stability. The fundamental equations governing system stability for most common nodal arrangements were published by Mahboob and Gulhati in the Punjab Engineering Congress in 1941. For nearly 50 years they were ignored. (The theory covering all possible nodal cases was published internally in my paper in the World Bank in the late 1980s and used, with permission, by Leslie Shanan in his book “Agricultural Water Management in Developing Countries”; Elsevier 1992). As a result of this fundamental ignorance we see in Indonesia, consultants replacing canal stop log controls with radial gates which, together with the Romijn weir offtakes, form the most unstable hydraulic combination possible ( “The Dilemmas of Water Division”, Lucas Horst, IIMI, 1998). Any network of canals, unless all the nodes are F=1 is unstable, even though the network may deliver a designed distribution pattern at a given Q, and change of Q will produce instability. Horst was the first to point out this previously completely ignored hydraulic fact, and should be given full credit. The result is a continual desperate fiddling with gates to try and get some measure of equitable service in a system which is to begin with conceptually unworkable.

It was in the early 1980s that I began to wonder what we, as a profession, were doing wrong. I had been working in the developing world for some 30 years and could not help but notice the appalling record of non performance in demand based small holder irrigation systems. Systems typically served the top third most of the time, the middle third part of the time and the bottom third hardly ever. In S.W. India (demand based) I asked a tail end farmer when he last had an irrigation supply and he called his father – it was around 4 decades ago. Solutions attempted followed the Western preconception of a hose pipe for every farmer to water his petunias and included ever more sophistication, more controls, more measuring structures, more computers, more training, more foreign education tours to California (where cell phone equipped ditch riders were exceedingly puzzling – but Disneyland entertaining) and, of course, more consultants. Hundreds of millions of dollars have been spent on “Technical Assistance” in the second half of the 20th Century to 3rd World irrigation and I defy anyone to show me a single demand based project that works as advertised. The downward spiral grinds on. No one seems to mind.

In 1983 I met Leslie Shanan in India and saw the warabandi system in action. It lifted the scales from mine eyes. I coined the term ‘structured system’ in the mid 1980s because I realized that wari (turn) bandh (fixed) was only the end feature of a system in which the internal structure is at least as important. And in the case of paddy, the fixed turn is not applicable to the field to field method – which the structured concept can serve perfectly well – see ‘the special case of paddy’ (IWMI Research Report #58). Leslie Shanan agreed with me and uses the term ‘structured system’ in his epic paper.

My own personal contribution to the technology, apart from word coining, is the discovery that a pipe outlet, whether a ‘free’ off take or overshot controlled, if carefully designed can yield a useful range of proportionality. This is contrary to what the textbooks say and is very important for the design of the proportional portion of structured systems. The design requires a simple computer program which I have called ‘CFO’ – canals, flumes and offtakes.

Of course there are some who sincerely believe demand systems are possible irrespective of scale. And there will always be vested interests who profit from these opaque highly technical systems. But sound research on the effects of scale on system performance will help in resolving these problems.

All the lending agencies, both bilateral and international should require countries borrowing for new, or major rehabilitation of irrigation projects to show cause why structured systems should not be used. All of us must passionately desire to redirect the awful exploitation of farmers and eliminate the poverty which condemns 3rd World rural humanity to a Stone Age existence. Structured Systems will not solve all the problems but it forms a foundation providing a basis of hope to build on – something which does not exist with the demand approach.


For those who want to learn more you should read: Fundamentals of Smallholder Irrigation: The Structured System Concept. by B. Albinson and C. J. Perry. International Water Management Institute, Research Report 58