Drought in India, not Bangladesh

The government of Orissa is claiming drought conditions in 8 of its 17 districts, and the Indian national government is investigating those claims this week. I'd wager on the central government coming back and confirming crops did fail in most, if not all of those districts.


The above image shows the four critical months of the Indian monsoon: June, July, August and September. The circled area shows the location of Orissa. Looking at this you will note that three of the four months were poor, and two of the three bad months were abysmal. Only one month, June, was marginal.

Looking just to the north of Orissa, you see a dry pattern for most of these months. Drought looks wide spread across the nearby Indian states of West Bengal and Jharkhand and all of Bangladesh. Before we get carried away with failed crops across eastern India and Bangladesh, let's take a quick look at ground data. Notice that at the top of most graphics on this site you frequently see "Estimated Rainfall". This refers to the fact that for the purpose of maps, I rely on satellite base precipitation estimates. So, before declaring drought, lets take a look at some ground based data.

Let's turn to the US Climate Prediction Center, who has a great data set for this kind of analysis. In their data set, Jharkhand is still united with Bihar, so we will have to look at those states together:

The top graph here is the important one, the normal and observed precipitation are both plotted. The brown on the graph shows the area were the observed precipitation is not enough to rise above the normal precipitation, and signifies rainfall deficits. There is absolutely cause for concern in one or both of these states. West Bengal, however, looks to be alright. Rainfall was nearly perfectly distributed across the season:

The green area in the upper graph shows the rainfall surplus that the wet season ended with. The minimal amount of brown also shows how well rainfall followed climatology throughout the season. Meanwhile, Bangladesh had season long surpluses, as shown in it's nationwide graph:

Flooding appears to be more of a problem nationally versus drought. So the obvious question becomes, why use satellite rainfall estimates at all if they can be so inaccurate? Most of the areas DevWeather monitors do not have nearly the rain gauge coverage of India, or even Bangladesh. Since a gauge only covers the location that it exists in, a dense network is needed to observe an areas that deviate from any regional tends. Satellites can, and do, cover the entire globe eliminating this problem. The best, most reliable products are those that merge the two ways of measuring rainfall together. Using satellites to monitor spatially, and gauges to check and calibrate the satellite output is likely to be the future of large scale weather monitoring in the developing world.

In case you thought I forgot, here is the Orissa graph. Looking at this, it seems quite possible that there could be problems in part of the state, and is well worth the investigation being done by the Indian government.