In the area where I live in sunny South Africa it has been a little too sunny for the past two years. We have been experiencing a drought. As such the reservoirs that supply water to our town are at very low levels. This has resulted in water restrictions being enforced by the local municipality. No topping up of swimming pools, washing cars or irrigation of any nature. I like doing all these things and therefore decided to install a reasonably well engineered rainwater harvesting system.

The main purpose of this instructable is not to give a step by step breakdown of the installation of my specific system but rather a breakdown of my design process and how it ties in with my intended water use.Obviously I include the construction part as well.

Just like I found many ideas from many different builds all over youtube and this website I hope others can use this in the same way. I had to extract ideas that were useful to my project from probably a hundred different videos and other sources. I got so much good information from all over, so as a way of trying to contribute to the knowledge base I have created this instructable with the hope that it will be useful to others in some small way.

Materials needed

  • Rainfall data for your area
  • Understanding of your catchment area (your roof and other hard surfaces)
  • Clear objectives of what you want to use your harvested water for
  • 110 mm PVC drainpipe (amount is obviously site dependent)
  • Assorted fittings for the above
    • 90 degree bends
    • T-pieces
    • Connectors
    • Drain gulley heads
  • 32 mm HDPE pipe, enough of
  • Assorted pressure couplings
    • T pieces
    • 90 degree bends
    • I recommend an arrangement with your hardware store to take a whole lot of these of all different sorts where you can return what you didn't use for cash back. It is very time consuming to plan it all out 100% before going to the shop
  • Valves
  • Taps
  • Appropriately sized pump
  • Tanks
  • Cement, stone and sand
  • Basic tools
    • Spade
    • Axe to cut tree roots
    • Hacksaw
    • Soapy water to help with fittings (in my mind this will become "sticky" with time, if not I am in for a bad time!)
    • Tape Measure
  • Strong back for digging lots of trenches

Step 1: Determine Your Rainwater Harvesting Potential (and the Specifics of My Project)

This step is quite simple and is repeated on every instructable and website related to capturing rainwater.

In simple terms your harvesting potential is the surface area you have available to collect from (typically a roof) multiplied by your annual precipitation. 100 m2 roof in an area that receives 1 000 mm of rain per year ==> 100 m3 of harvesting potential.

For my project I have about 300 m2 of tiled roof and 100 m2 of paved area that is available to collect from. The mean annual rainfall where I live is 1200 mm per year. Thus my collection potential is 520 m3 per year. This is the theoretical maximum water that you can harvest and use per year. As you will see later on designing a system that will actually deliver close to that theoretical number can be very expensive and impractical.

Once you have you your harvesting potential unfortunately you have to start subtracting from it. For this you will need to know:

1) The type of surface you are collecting from OR

2) How your rainfall is distributed in terms of event sizes

Collection surfaces have different run-off factors. From looking around extensively (wouldn't call it research as such) it appears that corrugated iron surfaces are best in terms of heaving a high run-off factor(0.9) and thatch is bad (0.5). My house has a tiled roof for which 0.8 seems to be a good number to use. I would recommend that if you don't have a lot of rainfall data you make use of factors.

I am in the fortunate position to have a lot of rainfall data for my town. I have also been observing my gutters during rain-events. I must admit my conclusion is not super scientific but I have found that for rain-events of less than 3 mm there is virtually no water coming from the gutters. I therefore remove all rain-event smaller than 3 mm from my data set and subtract 3 mm from all the remaining events to determine my site specific annual run-off to be 944 mm per year. Thus my effective factor is 0.79 which is slightly lower than the recommended factor but quite close. If you have many small events the factors will be optimistic and if you have fewer large events the factors will be pessimistic.

Step 2: Determine What Is the Right Amount Storage for You

How you determine this is going to be very dependent on your site and water uses. To get a good idea of how much tank volume you need you really need to have access to some good data. If you do not have access to good data you will have to fill in the blanks with estimates. Nothing wrong with that, just know the limitations and be aware of how wrong you might be. If you are happy with your degree of wrongness and are willing to live with the consequences no big deal.

My intended water uses:

1) Watering the garden

2) Filling up the pool

3) Washing cars

These are clearly not very critical uses as I intend to rely on municipal water for drinking, cooking etc. If rainwater was my only water source the following thought process would obviously be very different. My view was to get close to the most economical system and not one that could supply my drinking water year round.

Storage size determination for my site.

To determine the right amount of storage I chose to find the point of diminishing returns. To do this I once again consulted my trusty rainfall data. See graph. The analysis that follow does not allow for multi day events. If it rains for four days 20 mm of rain every day my tanks would be full and i would lose the last three days of opportunity. The annual usage is thus still on the optimistic side. I am sure one can calculate this but I felt that for a DIY system upon which my life does not depend it was unnecessary. Basically i didn't feel like faffing with excel anymore and was quite happy to file it under acceptable wrongness.

Unfortunately there was no clear inflection point on the graph as rainfall event sizes seems to vary a lot for my area. However at eight cubic meters of installed storage the angle of the graph flattens out. It also so happens that I have a very nice spot where I can fit eight cubes of storage and didn't really have appetite to install more than that. My system allows me to capture 20 mm of run-off (23mm rain event) if the tanks are empty. And can yield up to 220 m3 per year. For those keeping track we are now at a theoretical efficiency of (220/530) 41.5%.

To get my system effcience close to 79% (limited due to run-off) I would need to install 40 m3 of tanks which is just stupid in my case.

Step 3: Determining My Layout

Once you have your storage capacity you might want to consider the layout that you want. If you are lucky there is an obvious solution for your property. I have quite a big property and had many options available to me. I will discuss two main ones.

1) Spread out tank option


- Tanks close to main collection surface (roof)

- Tanks close to electrical supply for pump

- Minimal digging up of the yard for laying new pipes

- Easy to install in stages


- Elevation differences around the house would require sealed tank lids and stand pipes if i wanted all tanks to act as a single storage. Which I do.

- Ugly piping requirements on the outside of the house

- Would require sumps and pumps to capture water from paved areas

- Difficult to size sump pumps as rate of rainfall varies significantly

- Generally difficult to make use of 100% of my collection surfaces without unsightly modifications to gutters

2) Cluster tank option


- Rainwater can gravity feed to collection

- easy to make use of 100% of collection surfaces

- Can hide tanks from street view and the view from inside the house


- Lots of digging up my garden

- Far from electrical supply

- Far from existing sprinkler manifold (I would not consider moving this as it consists of nine zones.... just no!)

- Have to find a solution to deal with sand washing in from paved areas.


A reasonably diligent cost estimate comparison between the two systems showed the two options to be similarly priced. The long piping distances of the cluster tanks system was offset by the cost of sump pumps. For the electrical supply it turned out the the biggest cost component was the labor of connection and the connections themselves much more than the distance of supply so in terms of electricity there wasn't a big difference. I found that I could make the system about 50% of the cost if I relaxed on my condition to capture run-off from my entire roof. I did not want to do that. I repeated the calculations from step two for this reduced system, estimated the total cost and found that the payback for the smaller system would be slightly longer than for the bigger system.

I selected the "cluster tank option" mainly because of the minimal visual impact it would have on my property. I didn't feel like having a bunch of scattered tanks around my house. The cost .... short term destroyed lawn to avoid long term ugly tanks everywhere.

Step 4: Decide on the Features You Want in Your System

For those that have been looking at rainwater harvesting systems for a while you will notice that there are many typical features of these systems that are absent in mine. The main reason for this is that I have a very clear set of uses for the water. Certain other design features in my system also compensate for some of the missing features

Some notable omissions

1) First flush diverters

I am capturing water from my paved areas as well. It would be very difficult to use a first flush system while maintaining 100% gravity feed to my tanks. I could have installed something on some of my downspouts but my feeling is that if it is not installed in 100% of the system I might as well not use it at all.

Some of the function of these will be performed by the 110mm PVC pipes that connect to the tanks. They relatively large diameter and include vertical sections. Any sand and sinking bits will remain in the lowest portions of the pipes as the water velocity in the pipes will not be sufficient to carry sand over into the tanks. I have installed an inpection hatch at the lowest point in the pipeline where I can periodically clear out sand.

Small "floaty" particals will make it into the tanks. Each tank has about 5 cm of dead volume where these can accumulated and be cleaned out every once in a while. If they make it into the pump it is also not a big deal as they should be able to pass easily through the irrigation system.

2) Gutter guards

Filter for leaves and stuff just before the tank inlet. I will just have to have the discipline to clean it regularly. I am not good with that.

3) Post tank filtering an UV

I am not planning to use for drinking water. For my other uses a bit of bits doesn't bother me.

Features I do have

  • 100% of my catchment area utilized
  • Gravity feed to tanks
  • Pumped distribution
  • Coarse screening
  • Each tank can individually be emptied for cleaning
  • Connection to my automated irrigation system
  • Overflow drainage

Step 5: The Tank Stand. More Accurately a Tank Sump in My Case.

Up to deciding to go with the cluster tank i have estimated the fall on my yard by eye. This is a bad idea, and I just happened to have luck on my side. Luckily when I borrowed my Dad's dumpy level it turned out the fall was quite a bit more than i thought. I used this information when I sourced the tanks for my system. I want the water to gravity feed into the tanks. Thus if the tanks are too high it would either not work or I would have to dig them down quite a bit.

The tanks I ended up buying are 1.8m high. Although I allowed for a some digging in the option evaluations it turned out I didn't really have to to make the system work. I decided to do it anyway for two reasons:

1) Hides the tanks better

2) I would have more diving head to get the water into the tanks which means that I am less likely to be piping capacity constrained in heavy rainfall events. Although with 110 mm pipes I didn't foresee a problem. Mostly the hiding thing then.


To keep the sand slopes from falling into the tank stand area a small retaining wall was constructed. We left access for piping and built in provision for drainage.

Tank site drainage:

Given the sunken tank design drainage would be crucial for:

1) Rainwater falling directly in the tank stand

2) Tank overflows when the design overflow is overloaded

3) Any water emanating from draining small portions of a tanks or pipes for maintenance

This is something I overlooked initially. Luckily not too expensive in the end. I decided to install two subsoil drains.

I used special fabric to wrap building rubble in and had 50mm pvc pipes leading from the site into the subsoil drainage channel. Also included a large overflow sump for those really big problems. I did no calcs on the size requirements. These we just eyeballed. Hope it works OK otherwise my pump motor and controller will be gonners next time we get 100mm of rain! They are probably undersized.

Step 6: Installing the Underground PVC Pipes

This is hard work but quite simple.

1) Dig trenches (do what is legally required before you start doing this, where I live it is not a big issue but I do believe in other countries some administration and permissions are required). Spade length depth I have found is good.

2) Try to have you pipe as a single fall to avoid sediment collecting in places you can't reach. Or make them reachable for cleaning

3) Put pipes in trenches and connect using the connectors. I found it very difficult without soapy water to help the joints slip into place. I hope that the joints don't stay slippery forever.

4) Do your best to share trenches between pipes, for instance I use the PVC pipe trench to connect my sprinkler manifold to the pump with HDPE pipe.

Not much to say here. It just has to be done.

NB NB For my system to work all the collector sites have to be higher than the inlet of the tank (hence the sump). I used a dumpy level to survey the elevations in my yard. Based on this info I determined that to get sufficient head (to deal with quick large rain events) between my collectors and the tanks I would have to drop the tank elevation by 600mm from ground level. Also in certain areas around my house I will have to add up to a meter elevation to the collectors( I fastened these to the wall with brackets, not in photo). With all these I have about 1.1 m of head between collectors and tanks.

In a 40 m pipe length of 110 mm pipe this will give enough flow to cope with most rain events. Very roughly 1.1m head is about 10 kpa pressure. For a flow rate of 13 l/s the head loss is 10 kpa per 40m. Thus if the system is asked to handle more than 13 l/s the collectors will overflow. What is this in mm/hour for me?

1) 13 l/s = 46.8m3 per hour

2) 46.8 m3/ 400 m2 = 0.117 m = 117mm of rain in an hour

3) This is enough because my tanks will fill in 15 minutes and it is about a 1 in 20 year storm for the area where I live.

4) Frankly, when this happens I have bigger problems than overflowing collectors.

Step 7: Install Your Tanks

I am using four 2200 L tanks. They are connected with 32 mm HDPE pipe at the bottom. Each tank can be individually closed. This is simple

1) Make a tank stand suitable to your property (there are many ideas out there)

2) Put your tanks on there

3) Make sure they don't blow away when empty

4) Connect them up so they can behave as one volume

5) I strongly recommend putting a valve for each tank mostly for maintenance

6) Connect some sort of overflow arrangement. Lots of times I believe you can get away with no overflow, my tanks are in a sump so I need overflow and lots of drainage for when the overflow is undersized otherwise my pump will flood.

7) Make your final connections between the PVC collector pipes and your tanks.

Step 8: Test Your Collectors

Put some water in the collectors and experience satisfaction as water flows into the tanks!

Step 9: Install Your Pump and Connect to Irrigation

I installed a lather large 9 zone irrigation system about 2 years ago (2014). The system required a pump. I am using this pump to distribute water from the rainwater tanks. I am leaving my irrigation manifold where it is and pumping back to it from the tanks. The pump is a bit over sized for the zones and they used to be slightly over pressured resulting in excessive misting from the sprinkler heads. I hope that the additional 40 m of pumping distance and 4 m head differential will absorb some of that pressure to alleviate my misting problem.

I am using HDPE as this will be my mainline and will always be at about 4,5 bar of pressure. Again using pressure couplings. I installed a few taps along the mainline for convenience. I will use these to fill up the pool and run hosepipe for washing cars. I also installed a pressure tank to reduce the pump cycling on/off for small amounts of water (it is apparently rated for 100 cycles per hour continuous duty but I'm not buying it. The pump was too cheap for me to believe it!).

You should now have a functioning rainwater harvesting system. Enjoy and grow some tasty veggies!

As I improve and install finishing touches on the system I plan on updating this and posting some additional photos. Looking forward to some comments and opinions!

<p>Kudos mate, very well done, you got my vote! Hope you get rain soon - last I heard Kroonstad was still suffering badly :(</p>
<p>This it's a great rainwater system. I really want to make one of these for my garden.</p>

About This Instructable



Bio: Father, husband, engineer, hole digger, woodworker, tinkerer, aspiring triathlete
More by Bechtel82: Dust Collector remote control for $25.50 Rainwater Harvesting and Distribution 
Add instructable to: