May Rainfall is Average – An Unusual Event

Our rainfall in May of this year was average. In fact, it was nearly exactly average: the 17-year average is 22.4 inches for our local area in western Panama; this May we had 22.7 inches.

On the following graph, the blue line is the 17-year average for each month of the year, and the red line represents the rain we’ve had so far this year. Months that have more than 10 inches of rain occur during our rainy season, thus our rainy season extends from May through November, a period of seven months.

March had rainfall pretty close to average, but February and April were way off the average line. Because April was so far above average (it qualified as a rainy season month this year) and because May was so on-the-nose average, I decided to look at just how often do we have average rainfall in any given month.

What is average?

Here’s a graph representing each of 17 years of rainfall, superimposed on the average (thick black line).

On seeing this graph, one person remarked that “It looks like there’s not a single average year on that graph!”

So I wondered – how many months actually have average rainfall? I created a table showing the deviation from average for each month. (You may have to click on the table to see it more clearly.)

I highlighted the months that differed from the average by 1 inch or less and then totaled those months in the last row. A total of 28 months during that 17-year period had within an inch of average rainfall. This is 14% of the months represented. We can therefore expect that we will have non-average monthly rainfall 86% of the time.

Over that 17-year period, August never had near-average rainfall. September had the greatest deviations from the average – more than two feet of rain (24.4 inches) above average in 1999 and 20.4 inches of rain below average in 2009.

Does El Niño have anything to do with this?

It turns out that the 2009/2010 El Niño dissipated in May of this year. You may have noticed in the table above that from June through September of 2009, rainfall was below average, usually several inches below average. This is the typical pattern of El Niño in Panama – below average rainfall in the rainy season – leading to droughts – and (slightly) above average rainfall in the dry season. With the dissipation of El Niño in May, our rainfall returned to exactly average.

If El Niños result in less rainfall in Panama during the rainy season, then La Niñas result in more rainfall during the rainy season. Note that the September 2-ft excess rainfall was in 1999 and that rainy season months August through November of that year all had higher than average rainfall. That year, 1999, was a La Niña year.

Back in graduate school, when I first learned about El Niño and atmospheric circulation as a basis for understanding the surface currents of the oceans, we learned that El Niño occurs more or less every 30 years. This gives you an idea about how long ago I was in graduate school! It seems these days that we’re either in an El Niño or a La Niña practically all the time. In fact, NOAA’s Climate Prediction Center states that “Conditions are favorable for a transition to La Niña conditions during June – August 2010.”

Grateful as we may be for average rainfall this May, it looks like average conditions won’t last long, and we may be in for some heavy rain this year.

Update: Carla Black has posted a comment giving the rainfall data from Volcan, which is to the west of us and uphill by a great deal, but still on the Pacific slope side of Volcan Baru. You’ll see that their May rainfall, unlike ours, was far below average.

A lesson from this observation is that local conditions can seem to overwhelm global conditions at any particular time. For instance, last month Carla sent me 13-year’s worth of rainfall data from Aprils in Volcan. This exchange was prompted because here in Potrerillos we had an extraordinarily high rainfall this April. When I plotted all 13 Aprils from Volcan vs. the same 13 Aprils in Potrerillos, I found that there was no correlation between rainfall in the two places whatsoever. Yes, there’s a slope to the line, but the r-squared value, which shows whether there is any significance to that slope, is about 0.2, which says that there is a 20% chance that the data are related. Or an 80% chance that they are not related.

So, if it’s raining in Potrerillos, we cannot predict at all whether it will be raining in Volcan. For example, she wrote that it had been raining for six hours in Volcan – at the moment I received her comment it had only begun to rain here and has now done so for about one hour.

Good lesson!