Getting back (finally) to my most recent topic I have some numbers to throw out, in order to assess whether relying on rainwater for a domestic water supply is technically feasibly in a semi-arid or arid climate. Tucson would be considered semi-arid because our average rainfall is roughly 12 inches per year (although this year we have only had about 6-7 inches, which makes us arid at this time), but we are classified as arid because of our high evapotranspiration rate.
There are many skeptics who say that rainwater cannot be relied on as a water source in places like this because rain is too unreliable. Our annual rainfall typically comes in two seasons - winter and summer. Our winter rains tend to be gentle, slow rains that might occur several times a month, amounting to about 5 inches on average during the period December to April. Our summer rains, on the other hand, are known as monsoon rains because they result from a seasonal wind shift in summer, and often come in torrents. We typically receive between 5 and 7 inches of rain in the summer and it's not uncommon for the bulk of that rain to arrive in 3 or 4 rainfall events during the months of July and August. In between those seasons we might typically go for 2 or 3 months with little or no rain.
With that type of rainfall pattern, obviously, the key elements of rainwater harvesting will be capture area and storage. If you have a sufficiently large surface area from which to capture water and sufficiently large storage to hold onto that water during long dry periods rainwater becomes a feasible water supply.
Here are a few basic calculations of available water:
10" of rain falling on 2000 sq. ft. of roof surface will yield roughly 12,500 gallons of water.
10" of rain falling on 3000 sq. ft. of roof surface will yield about 18,700 gallons of water.
12" of rain falling on 2000 sq. ft. of surface yields about 15,000 gallons, and
12" of rain falling on 3000 sq. ft. of surface yields about 22,400 gallons of water.
If you take those numbers and average them out over the course of the year you come up with a range between 34 and 61 gallons per day. Obviously you are not going to maintain a home, yard, and pool by collecting rainwater unless you have a very large surface from which to collect the water. But it's perfectly reasonable for two people to survive on 61 gallons of water per day for indoor uses such as cooking, cleaning, bathing, and drinking.
But how do you make rainwater suitable for drinking? That is the tricky part, some of which I will try to address in a post on the regulatory limitations on use of rainwater for water supply. But in a general sense, you must install a home water treatment system to make this water suitable for consumption. This ranges from selecting proper roofing material that won't leach chemicals into the water falling on it, to engineering the collection system so the first flush of water coming off the roof is bypassed (to limit the bird poop in your water supply), to a system of filters and treatment technologies that will ensure no harmful bacteria or other nasties in your water. This is the primary annual cost of this type of water supply - the energy and maintenance of the treatment system. Most of the other costs are upfront when the collection and storage are installed.
This site includes some helpful information on what is necessary for making rainwater suitable for potable uses.
Is this type of system suitable for the average homeowner? Clearly not. But there are people out there who are willing to invest the money, time, and effort to get off the grid, or off the pipes in this case, and whose lifestyle allows them to live on only the amount of water necessary for basic, indoor human needs, without all the extraneous uses of high-quality drinking water many of us find necessary for our quality of life.
More when I find the time, including the aforementioned regulatory analysis of rainwater harvesting.