New user?

Rainwater harvesting

Implementation of a large scale rainwater harvesting project using the roof of Tabora cathedral in Tanzania

The cathedral has a large roof and using average rainfall data between 1894 and 1990 from and a run-off co-efficient of 0.85 around 500'000 litres could be collected in a year. Using a rationing method this gives a storage requirement of approx 200'000 litres to supply around 1360 litres per day. Given the significant size of this project and in want of more information regarding the performance of the guttering in heavy rain the goal is perhaps to provide 100’000 litres of storage. The cathedral has 8 down pipes so current thinking is between multiple below ground plastic tanks at all 8 locations with a moveable pump or one or two very large below ground/partially below ground ferrocement tanks with appropriate guttering. So here are some initial questions which I would be grateful for your advice on. 1. What is your opinion on the two schemes? 2. Do you have any specific construction advice given that the area may experience some seismic tremors on an irregular basis? 3. Can you suggest available and reliable pumps for each scheme? 4. Is a mutely-tank/moveable pump design a viable one, have similar designs been done, what might access to the individual tanks look like?

edit retag flag offensive close merge delete

4 Answers

Sort by » oldest newest most liked

I would do anything to avoid pumps – they need care, attention and money – they usually lack all three despite everyone’s best intentions and breakdown permanently – to everyone’s frustration. I would suggest therefore ground level tanks with lockable taps to fill buckets. May not be aesthetically so pleasing but practical arguments are strong. Number of tanks? I think one would have to see the site and do some rough calculations of cost and also think about how water is to be used. I am doing the same for my garage and house roofs here in Switzerland and am still thinking about all the advantages and disadvantages of various combinations. Possibility of earthquakes is an argument for multiple tanks – less impact if one breaks and not so likely that they all break. Perhaps site them a diameter away from the cathedral walls? You can make special collars which fit into the down pipes at the same level as the maximum water level in the tanks. They collect most of the water when the tank is not full, but then automatically let the water pass on down the pipe when they are – no moving parts. Can send sketch if required. The classic reference on rain water harvesting and ferro-cement tanks are:

Rainwater Harvesting: The collection of rainfall and runoff in rural areas Arnold Pacey and Adrian Cullis This book emphasizes the importance of social, economic, and environmental considerations when planning and implementing projects. For rural development workers, it aims to fill the gap in existing literature on the gathering and storage of water. 224 pages, 155x234mm, Paperback, ISBN 9780946688227

And Ferrocement Water Tanks and their Construction Simon Watt Describes how cylindrical water storage tanks of up to 150 cubic metre capacity can be built using wire-reinforced cement-mortar. Covers design and planning, costs, standard, recommended and alternative construction methods, and other information. 120 pages, 147x209mm, Paperback, ISBN 9780903031516

Regards, Tim Foster

edit flag offensive delete publish link more

My own feel is that plastic tanks might be more earthquake proof than ferrocement but more expensive? I would go for 8 tanks rather than 1 large one and would be tempted to use lockable manual diaphragm pumps, one on each tank if you put them into the ground or as Tim says site the tanks above ground and use lockable taps. Tim's book recommendations are sound - I used the first in DRC to produce rainwater tanks for a rural hospital. You have to ensure some filtration system to get rid of the first flush after a dry spell which would contaminate the water. Regards, Toby Gould

edit flag offensive delete publish link more

I would say that while you may expect to receive 500,000 litres in a year you should also consider the rainfall pattern with respect to the usage pattern before sizing your tanks. Some parts of the country have two rainy seasons and others only get one, I'm not sure what the situation is at Tabora in the middle of the generally arid central plain.

You will need to use a pump anyway to raise water from an underground tank so you'll need to look at the construction practicalities and costings for all options before you decide whether to go for one, eight, or some other number of tanks. You need to consider where they want to pump water to as well as where the down-pipes are.

For roof water harvesting you should look at some mechanism on the down-pipes to remove the dirt and debris in the first flush of rain before diverting water into the tanks. The Centre for Alternative Technology ( may be able to provide appropriate information sheets or publications. With a couple of exceptions I have only seen plastic tanks used above ground. I have seen onion shaped cisterns dug out of solid rock in the West Bank and rectangular concrete lined ones in softer sediments. The entrance should be covered with a lockable man-hole to prevent unauthorised use of water and there should be a lip to prevent surface run-off from getting in. A good supplier of a variety of pumps in Tanzania is Davis and Shirtliff (

You may wish to get in touch with the Rainwater Harvesting Association of Tanzania (RHAT) which is a network of a variety of people interested in the subject. I do not have a contact in Tanzania but you can probably get in touch through the Southern and Eastern Africa Rainwater Network (SEARNET) at or the Greater Horn of Africa Rainwater Partnership in Nairobi.


Martin Ager

edit flag offensive delete publish link more

You could consider connecting multiple tanks to a wet well sump located below the level of your tanks. This would allow a single pump, the wet well being supplied by gravity, connected to your intended distribution system. This would allow either a submersible or surface mounted pump to be used. It is difficult to comment on the cost effectiveness of these options without knowing what kind of piping network would be required to connect, but a single pumping station would probably cheaper and certainly less complicated to run and maintain. It would also be fairly simple matter to provide extra storage to the network after it had been established. I would suggest you check out the IUCN website on rainwater harvesting, you will be able to have a look at appropriate guttering and associated collection efficiencies. Commercial pump suppliers, such as Davis and Shirliff (Offices in Tanzania websites will generally have performance data on their pumps for download. Regards Cody

edit flag offensive delete publish link more
Login/Signup to Answer
Question Tools
1 follower
Public thread

This thread is public, all members of KnowledgePoint can read this page.

2014-01-13 08:17:52 -0500
608 times
Last updated:
Jan 13 '14