The posts on this blog so far have focused mostly on weather and not really on climate. (Not counting the talk at Open Data Camp). Today’s post seeks to rectify that.

When people talk of climate, it usually refers to averages and numbers from long periods. “How long?” is a fairly nebulous question, but the convention today is that it ought to be for at least about 30 years.

Below you’ll find a monthly rainfall profile of Bangalore, which is a result of averaging monthly rainfall amounts from a 100 years of data, from 1901-2000. This monthly data is actually publicly available at an old IMD page which was discovered with the help of Anand, so feel free to use it yourselves as well. Unlike other sources, this is a 100 years of station data, not gridded, not massaged, so it’s a valuable resource. I know that I’ve made strong statements in favour of looking at daily data, but monthly numbers do have their uses, and climate information straight from the weather stations is ideal, if you can get your hands on them.

I’ve chosen the full 1901-2000 period above in order to capture whatever long-term patterns exist, and the graph that you see above can change a little when you take an average over say the more recent period.

The hundred year span of data, however, is quite a luxury most of the times. Bangalore was a seat of a British cantonment and also a prominent city in the Kingdom of Mysore and was fortunate enough to have a weather station that was set up late in the 19th century. Thus, for standardization, meteorological departments choose a common 30 year period during which they have continuous data for the most number of weather stations and rain gauges across the country. One standard followed by the Indian Met Department is the period 1951-80. The graph from then is below.

Note that there is only minimal deviation from the previous graph here. Rainfall in March, May, September, October, and November are a bit higher, the rest barely changed. What would happen if we instead try to look at the climate of the most recent 30 years for which we have data, 1971-2000?

Turns out that things change quite a bit! September rains are way higher at 233mm, small increases in June and August, losses in October, May and November. The line graph below has all three together.

Now what does this mean? Does it mean that Bangalore’s climate is changing?  The annual totals for the three periods (1901-2000; 51-80; 71-00) come up to 930mm, 963mm and 968mm. Since the annual totals are not very different, does it mean that the distribution of rainfall in Bangalore is changing rapidly?

The questions can be asked rather easily, but answers are harder to find. The short answer is: no, not quite. A lot of things change when you look at only averages, but these don’t necessarily imply underlying trends. Or even if they do, whatever is changing is not always obvious. Over the next few posts, I’ll be taking a closer look Bangalore’s rains and try to examine how it might be changing.

Written by Pavan Srinath. This post was a result of conversations on twitter and email with @zenrainman. The monthly rainfall data is taken from an old page on the IMD website.