Precipitation is a key factor determining agricultural potential, especially in drylands, wet and dry tropics but also in many regions with temperate climate. Where rainfed agriculture is more or less settled, a high degree of precipitation variability can be a major constraint, given the risk of drought.

Precipitation statistics from historical data are essential for assessing climatic variability. Analyzing these statistics helps identify deviations from the average over selected periods, probabilities of exceeding specific totals, and the frequency and intensity of extreme events like droughts or floods. These statistics can be derived from individual precipitation stations and extrapolated spatially to grid formats, such as the GPCC dataset.

A useful way to quantify these fluctuations and concommitant risk is by expressing precipitation not in terms of averages, but as probabilities of exceedance or non-exceedance over different time intervals. Thus, precipitation can be expressed as either the probability of not exceeding a fixed amount (e.g. on annual basis: 100, 500, or 2000 mm) or as the amount likely to be exceeded (e.g. in three out of four, four out of five, or nine out of 10 years). The assessment of precipitation probabilities on either annual or seasonal basis helps to overcome the influence of the shorter-term fluctuations. Stating precipitation in these terms helps to define ‘reliable’ rainfall and can thus serve as a simple adaptive capacity tool.

Precipitation probabilities can be obtained through the cumulative Gamma distribution, which for a long enough record, allows an excellent fit to ranked historical observations.

NOTES:
Pub. 109, Section 3.2.1.3 provides formulas for the cumulative Gamma distribution. An example of calculating probabilities in Excel, using 522 data points from West Asia, is available here.
Examples of interpolated grids showing annual precipitation probabilities for parts of West Asia are available here.