Understanding LA's Rainfall Variability
Los Angeles gets about 15 inches of rain per year on average. That number means almost nothing if you live here. I moved here in 2008 expecting predictable seasonal patterns. I learned quickly that the averages are based on data stretching back to 1877 at the official downtown gauge, and they smooth over conditions that would ruin your plans if you actually relied on them. The water year runs from October 1st through September 30th. This matters because California stores its water in snowpack and reservoirs during the wet season. The state's hydrologic models track whether you're in a wet, above-normal, below-normal, or drought category using the Palmer Drought Severity Index and standardized precipitation indices. I spent three years working with municipal water data and the first thing I learned was that the official rain gauge network has significant gaps in the mountains and eastern valleys where most runoff actually originates. Most people looking at LA rainfall data pull from the downtown recording gauge. It sits at 290 feet elevation near First Street and Temple. The problem is that this gauge captures maybe 60 percent of what falls across the greater basin. When I tried to model flooding for a stormwater project in the San Fernando Valley, the downtown numbers made my peaks look half as severe as they actually were. I ended up cross-referencing with gauges at Burbank, Glendale, and the Sepulveda Basin. The valley stations routinely logged 40 percent more rainfall during the same events.
Here is what the actual distribution looks like. Roughly 86 percent of annual precipitation falls between November and March. That leaves five months where you might get a trace or nothing at all. The variability within those months is extreme. I have seen months where the gauge recorded zero rain for three consecutive weeks, then 4.2 inches in a single afternoon during El Niño conditions. That happened in January 2023 when atmospheric river storm Tyrus hit. The downtown gauge logged 2.87 inches in 24 hours, but nearby Santa Clarita station recorded over six inches in the same window. The extremes are where this data becomes useful or dangerous depending on who is reading it. The wettest calendar year on record at the downtown gauge was 1940-1941 with 35.36 inches. The driest was 1923-1924 with just 3.72 inches. That is a ten-fold difference between years. If you are planning landscape irrigation, roof drainage, or anything that depends on reliable rainfall estimates, you need to understand which decade of data you are referencing. The period from 1977 to 2020 showed a slight upward trend in total annual precipitation, but the trend is noise compared to interannual variability driven by ENSO cycles. El Niño years typically deliver 20 to 40 percent above normal rainfall to the LA basin. La Niña years tend toward 20 to 30 percent below normal. The 1997-1998 El Niño dumped 41.93 inches at downtown. The 2013-2014 La Niña year produced only 5.67 inches. These are not theoretical scenarios. Both happened within my lifetime. The 2011-2017 drought period saw five consecutive years below normal precipitation, which dropped reservoir levels to crisis points and triggered mandatory conservation that lasted two years.
When you decode these patterns for practical use, start with the Climate Data Center at UC Berkeley. Their station network dataset gives you daily totals going back decades for multiple locations. Download the CSV files for downtown, Van Nuys, and Santa Clarita. Load them into a spreadsheet or Python script. Calculate rolling three-year averages to smooth out individual storm years. Flag any year above the 90th percentile or below the 10th percentile for separate analysis. This gives you a baseline that accounts for the oscillation between wet and dry epochs without relying on the single annual average number that everyone cites. The biggest pitfall I see is treating monthly averages as predictive. The November through March wet season accounts for most rain, but the exact timing shifts every year. Atmospheric rivers can arrive in October or drift into April. I once scheduled a construction crew for late February based on historical averages. We got shut down by a February 28th storm that dropped 3.1 inches in one day. The historical average for that week was 0.12 inches. Averages do not schedule your work. Percentile ranges might. If you need this data for engineering or permitting, use the NOAA Atlas 14 point precipitation frequency estimates. They provide adjusted rainfall depths for various durations and return periods. The 100-year 24-hour storm for downtown LA is approximately 5.8 inches. For the mountains above 2000 feet elevation, that number climbs past 10 inches. Designing drainage without accounting for elevation differences will get your plans rejected by the Department of Public Works every time.
Get the Full Details
The data is solid. The limitations are in how you apply it. LA rain does not follow a normal distribution. It clusters into intense bursts separated by long dry periods. Any model that assumes steady distribution will underestimate peak loads and overestimate soil moisture recharge between events. I stopped trying to force linear regression on this dataset years ago. Seasonal breakdowns with percentile bins work better. You will still get surprised by individual storms, but you will be less surprised than you would have been using the averages alone.
