Post #G21-043: A model of a rain barrel system based on historical weather data

Posted on July 20, 2021


If you use rain barrels to water your garden, you will eventually realize that you don’t have enough of them. 

This post tells you that you never will.  But that’s OK.  Most of the benefit you will get from using rain barrels, you will get with the first few you install.

This post isn’t nearly as folksy as the introduction implies.  To write this, I first created as reasonably sophisticated model of rain barrels and gardens using four decades of daily weather data for Sterling, VA.  While the results of that model really only apply to my local climate, I think the more general lessons about rain barrels apply broadly to well-watered temperate climates such as the climate of Virginia.

You know the worth of water when the barrel runs dry.

I routinely water my garden using a set of six 55-gallon rain barrels.  These are former soda-syrup barrels that I converted to rain barrels decades ago.  The total usable capacity is about 300 gallons, and they are fed by about 850 square feet of roof surface.

I find that I can adequately water my roughly 300-square-foot garden about twice, with that 300 gallons of water.  And it seems like I manage to run those rain barrels dry during the gardening season, and end up using municipal (tap) water (plus a filter to remove chloramine) to water the vegetable garden.

When they run dry, they do so in a hurry.  Rain barrels are an inherently unstable water source for gardening.  A modest dry spell simultaneously increases your garden’s need for stored rain water, even as it cuts off the supply.   As a result, my water barrels remain partly or completely full for months on end.  Until I hit a dry spell, at which point they go from full to empty in about a week.

How many rain barrels do I need?

I never planned my rain barrel setup.  It just sort of grew over time.  And that raises an obvious question:  Am I running out of water every year because my rain barrels system is under-sized?  If so, how many rain barrels do I need, to avoid running out of water?  And more generally, how would that vary with type of garden and climate?

That seems like a straightforward question.  Any normal person would probably try to find some reasonable rule of thumb for that.  Surely somebody on the internet or on YouTube will offer some advice.  But the only rule I came across is to store up enough to provide an inch of water per per week, to your garden, for the length of your dry season (from this website).

And while that’s plausible, it’s distinctly unhelpful.

First, we don’t have a “dry season” here in Virginia.  We have a bit of drought now and then, during the summer.  It certainly varies from year to year.  So, for sure, the number of barrels has to depend on the summertime weather patterns in a locality.

Second, the question isn’t well-defined.  What do you mean by “need”?  Do you need to be sure that you never run out of rain water?  Is it enough just to ensure that the water bill for watering with municipal water is reasonable?  Or is there some other criterion?

Finally, it surely had to depend on what you’re growing.  A raised bed packed full of trellised plants has a lot of leaf area supported by each square foot of soil.  In hot weather, that’s going to require a lot more water than (say) a bed of herbs.

A micro-simulation model using NOAA weather data.

Failing to find a good answer on the internet, I did what any reasonable person would do and created a simple-but-detailed mathematical model of a water barrel system and garden.

  • It starts with four decades of daily weather data for Sterling, Virginia.  This gives the daily rainfall for my area.
  • You then tell it how many inches of water you think your garden needs, per week, month by month for the growing season.  For example, for July and August, I think I need about 1.5″ per week, not the standard 1″ that everyone seems to advise.
  • Pick a value for the saturation point of your soil, which determines the maximum amount of rain water your soil can hold.
  • Set the size of your garden and the size of your rain barrels.
  • Finally, set the amount of roof area that drains into your rain barrels.

With all those things in place, this little model steps through the days of the growing season (assumed May 1 to September 30), for each of the past four decades.  When it rains, that adds to the water in the garden bed and to the barrel.  If it rains too much — if you reach the saturation point of your garden soil or the capacity of your rain barrels — the excess is lost.  If it doesn’t rain enough,  you first allow the soil to dry out, then start withdrawing water from the barrels to meet the garden’s daily water needs, once the soil is dry.  Both the rain barrels and the garden are constrained to have no less than zero water, and no more than their capacity.  If the garden needs water, and the barrels are empty, municipal (tap) water is assumed to make up the difference.

Really, it’s just an accounting exercise in a spreadsheet.  But instead of money, I’m accounting for water.

The best place to get historical weather data quickly is probably the National Oceanographic and Atmospheric Administration (NOAA).  Starting from this page, I was eventually able to download a .csv file showing daily temperature and precipitation  for the nearby Sterling, VA national weather service office.  I got four decades of data, which I figured was adequate for this task.

I think I’ve set up this first version so that one could download a standardized NOAA file showing climatology in another region, overwrite the Sterling VA data, and simply re-run the model to get results for another part of the country.  FWIW, we get about 20″ of rain around here, during the May 1 to September 30 growing season.

Model parameters and results

Here’s where I’m coming from.  It’s roughly 300 square feet of raised beds.  I’d like to have 1.5″ of water per week in the heat of the summer, but less at other times.  The rest of the data are as shown below.

Given that, all I have to do is add the actual historical weather data for the past 42 years and track the water during the May 1 to September 30 growing season.  The results will show the averages for the past 42 growing seasons.

For my convenience, I modeled this day-by-day, as if I were to make up each day’s water deficit in the garden by watering a little bit each day.  That was a lot easier than modeling a week at a time (which is closer to how you would actually water your garden), and I think it has no material impact on the results.

Here’s the result of those 42-year simulations, where each line of the table below represents a different number of water barrels.  I find the results to be informative.

First, the overwhelming lesson from this is that it’s a game of diminishing marginal returns.  My current setup is six barrels, 300 gallons total storage.  With that, I supply just under 2000 gallons to the garden during a season.  If I bumped that up to 20 barrels, 1000 gallons total storage (which would look ridiculous in my back yard), I would end up supplying just over 2000 gallons to the garden during a season.

In going from six water barrels to twenty water barrels, the total savings on my town water bill would amount to ($6.79 – $0.35 = ) $6.44 per year. 

In fact, half the savings I would ever be able to get on my water bill, I would get with just two barrels and a total storage capacity of 100 gallons.  Which tells you that maybe putting in a couple of water barrels isn’t such a bad idea, even if they run dry frequently during the year.

The second lesson is that, practically speaking, it’s impossible to guarantee that you’ll never run your water barrels dry.  Even if I had 20 of them, for my 300 square feet of raised beds, I’d still run dry in five percent of years.

The third lesson is that these rain barrels have essentially no impact on the runoff from my suburban house lot.  I’ve done a quick-and-dirty calculation here, and simply shown how many gallons of water fall onto my lot in a typical growing season.  (Which is not the same as runoff, because if it’s a light rain, a large fraction will be absorbed by the ground.  But I think this is close enough to show the point).  Even a 1000-gallon system would only be diverting a fraction of a percent of all the water that falls on my half-acre out here in the ‘burbs.

And if you were to look at the underlying weather data, that would all make sense.  Rainfall in this area tends to come in big chunks.  We have frequent 1″ and 2″ rainfall events, which in the summertime are often intense thunderstorms.  A 2″ event would generate more than 1000 gallons from the 850 square feet of roof that feeds my rain barrels.  With any modest-sized system that is partly full under most conditions, nearly all the rainfall during those intense events runs off and is lost.

Even year-by-year, rather than event-by-event, wet growing seasons mean that most of the water that might fill the barrels is simply lost, because those barrels are already full.  Conversely, in dry years, those barrels are going to run dry no matter what.

Finally, the monetary value of the collected rainwater (or avoided tap water, if you prefer) is minimal.  From a sheer cost/benefit standpoint, if you buy expensive off-the-shelf rain barrels, the payback period is enormous.

I was lucky enough to pick up a bunch of used soda syrup containers years ago.  (Well, actually, my wife picked them up.  But I drilled the holes and such.)  If you can pick up those food-grade barrels for next-to-nothing, this pays for itself in short order.  But given the labor involved, it’s not something you’d do based on economic motivation alone.


I made my rain barrels years ago, when I believed it was The Right Thing to Do.  And if you keep them painted, and drain them in the winter, these big, thick HDPE soda syrup barrels will last a lifetime.  Possibly longer.  My point being, I already own them, and I’m used to using them.

To be honest, If I’d run that model 25 years ago, I’m not so sure I’d have made the effort to use rain barrels.  Particularly now that filters for removal of chloramine from city water are widely and cheaply available.

There’s only one other service that these barrels provide.  They work very well as mosquito control devices.  The standard advice is to screen the openings tightly so that mosquitoes can’t breed in the water.  I find that it’s much more effective to leave the openings unscreened and add a quarter of a mosquito dunk every month.  As far as I can observe, this serves as an effective mosquito trap.  And as the season dries up, and these barrels become the only source of standing water in my yard, the effectiveness increases.  All at the cost of being sure to add those dunks each month.

Otherwise, I guess the message of this is that you should use water barrels if you want to.  They won’t save you a lot of money.  They really won’t divert much runoff from entering local streams.  They can help a bit with mosquito control.  And, plausibly, the quality of the water is somewhat better than out local tap water.  But that’s about it.

Sometimes the tree of knowledge is not the tree of happiness.  Even if that tree is irrigated with rain water.