Percentile Method

Each category on the U.S. Drought Monitor has a percentile range associated with it. For example, D4 is in the first‐to‐second percentile range. Percentiles help to place the drought into a historical context.

Let’s look at this another way using 100 years of precipitation data from a station in Nebraska. Percentiles are determined by ranking the data from largest to smallest. In this example, 2007 had the highest precipitation with just over 39 inches of rainfall and 2012 had the least with about 11‐and‐a‐half inches. Additional years falling in the top 5 wettest and driest are indicated by raindrops.

Let’s rank the precipitation amounts in order, from the year with the most precipitation to the year with the least precipitation. The average, or mean, precipitation for this location is 24.21. This exact amount only occurred in 2 years out of 100, 1962 and 1995.

Now, let’s apply the percentile categories for the U.S. Drought Monitor. Exceptional drought, D4, corresponds to the lowest 2 precipitation values. This is the most severe drought, with the worst conditions on record. It would only be expected to occur once or twice within a 100‐year period.

Extreme drought, D3, occupies positions 3 through 5. These conditions are still among the worst on record and would be expected to occur once every 20 to 50 years.

Severe drought, D2, occupies positions 6 through 10. This type of drought would be expected to occur once every 10 to 20 years.

Moderate drought, D1, occupies positions 11 through 20. This type of drought would be expected to occur about once every 5 to 10 years.

Abnormally Dry conditions, D0, occupy positions 21 through 30. We would expect these types of conditions once every 3 to 5 years.

Rainfall that falls within the top 70 positions do not get a designation on the U.S. Drought Monitor and show up as white areas on the map. As you can see, drought is truly a rare event and extreme and exceptional droughts are even rarer.

Convergence of Evidence

The authors of the U.S. Drought Monitor use Geographic Information System software, called ArcMap, which lets the authors overlay percentile maps for multiple types of data at multiple time scales to determine where the data is converging ‐‐ that is, what is the D‐x level of dryness or drought that most of the data indicate. A unique thing about the U.S. Drought Monitor is that over 400 experts around the country provide feedback and ground truth for the map. This provides for a built‐in validation step, which ties the indicators to the reality of local conditions and impacts that are occurring.

What do percentiles and the convergence of evidence mean for the map? A location that is designated with Exceptional Drought (D4) is experiencing a drought that would only be expected to occur once or twice within a 100‐year period AND that is evident through a strong consensus across multiple types of indicators.

These figures shows some sample data, put into percentiles and colored with the appropriate U.S. Drought Monitor categories. The solid lines on each figure are an overlay of the previous week’s drought categories. Keep in mind that this is only a small subset of the several dozen maps that the authors look at each week. Notice that the indicators do not line up nicely and tell the story as to what drought category the location is in. In cases such as this, the author’s must use a combination of the physical data, feedback from the local experts, impact reports, and experience to determine how the evidence converges to a final drought category.

Prev Next