12 days of rain and implications for cistern design

I saw this really interesting article in an email update I received from the magazine sister publication for the journal Science, url: https://www.sciencemag.org/news/2018/11/just-12-days-world-gets-half-its-annual-rainfall?utm_campaign=news_weekly_2018-11-16&et_rid=300642751&et_cid=2493410

Geophysical researchers sampled 185 weather stations all over the world (they kept their observations to sites within 50 degrees of equator for satellite accuracy/calibration reasons I don't understand, but that captures where almost all of us live anyhow). Interestingly, even though we think of some locations/climates receiving rain much more evenly throughout the year (e.g. here where I am in NJ and in the northeast US generally), versus more seasonal variation (Oct-March constant drizzle in the Northwest, or summer monsoons in the Southwest), the researchers found that consistently, on average across all sites, 50% of the entire year's rainfall occurs on just 12 days. This still does vary by region, where from the maps (the study doesn't show the data for each individual site, unfortunately) it looks like the SW gets half in maybe 6 or so days, while here in NJ I probably get half in about 16 - 20 days. This might not be a big finding for folks in dry climates, who are used to designing their systems to catch large, one-off rains, but might be more surprising for folks who design for a more even distribution of rain.

There could be a good "when it rains, it poors" (WIRIP) rule of thumb here for selecting cistern size, where you'd divide your annual rainfall by 2, then divide that product by the number of days to get to half for your area. So for my area, 44 inches/2 = 22 inches. 22 in./16 days = 1.375 inches/day. So I'd design for about a 1.4 inch rainfall. Of course, if you want to catch every drop, the paper also found that, on average across all sites, the rainiest single day of the year accounted for 8% of rainfall. For me that'd be 44*0.08 = 3.52 inches. But that's probably over-engineering for me. Still, catching optimal amounts of rainfall can be important, especially if you're doing Jeavons-style biointensive raised bed growing that requires more irrigation.

Another important thing would be knowing how the rainy days actually cluster. It could be very possible, based on observations of weather in my area, that several of these rainiest 12 days occur within a day or two of each other. If folks are interested, I could reach out to the researchers to see if they'd provide their cleaned data, but it will probably require statistical software to view/process it.

This is very interesting. I have always heard about the 80/20 rule.

20%(73) of the days of the year gives us 80% of the rain that we receive.
20% of that 73 will give us 15days and 80% of 80% gives us 64%
So 15days give us 64% of the rain that we receive.

I wish there was an app/website where I could pull the 15 wettest day and tally it then compare it to the total amount of rain per year.

Looking at my own airport weather station data.
I see that only 1/3 of the days actually have precipitation.

For small rain events alot of water is wasted by 1st flush/evaporation/infiltration/etc.
So with bigger rain event less water is wasted.
On a different note plants only get 50% of there water from rain/snow. The rest is from condensation/dew/fog/etc

I size my tanks for longest days between rains. If you use 2,000 gallons a month and "average" longest between rains is 3 months. Then 6,000 gallon capacity is your minimum storage .

In my area we generally get good rain around memorial day and labor day, with little in between.

For most locations you can find rain fall history data.  Rather than using generic rules, or world wide averages, etc.  I think it's best to dig up that data and design your system around rain fall patterns for YOUR specific area.

The things I looked for were things like:  longest period without ANY measurable rain (1 year in my area), the most rain in any one hour, and daily rain patterns.

With this information I've decided to build tanks large enough to hold one years worth of water for my household, 15,000 will just cover household use,  20,000 would  give me enough to keep my fruit and nut trees alive (but probably not productive)
Note: I'm capturing grey water for use in irrigation

The 1 hour max tells me how large the pipes  need to be from gutters to tanks, you don't really need to build to the one hour max unless it frequently rains that hard, but if you can come close then you won't lose any significant amounts of rain from overflowing.

With a spreadsheet, you can use your average daily consumption, average daily rainfall, and tank size to determine how much capture area you need to keep the tank adequately full. With that information I've determined that I need at least another 500 sq ft of roof (hmm, covered parking)

I'm not building for worst case, just good enough for 99% of the time.  Worse case I can buy water from one of my neighbors who has a well, or pay to have it hauled in.