Global high-resolution satellite records of rainfall only go back a decade or so – not long enough to monitor the comings and goings of droughts. However, there is a 30-year low-resolution satellite-based precipitation data set known as the Global Precipitation Climatology Project (GPCP). But GPCP has a time lag of around 12 to 18 months – not much help if you need to monitor droughts in real time.

To overcome these problems and create a global long-term and near real-time satellite record of rainfall, Amir AghaKouchak and Navid Nakhjiri from the University of California Irvine, US, combined these two very different satellite records using a Bayesian correction. The results are published in Environmental Research Letters (ERL).

"In statistics and probability theory, the Bayesian approach provides a way to update or correct an existing prediction given new or additional information, data, or evidence," explained AghaKouchak. "Here we correct the near real-time satellite data with historical observations from long-term satellite observations. In other words, using the overlap between the two data sets, we estimate the likely correct value of near real-time satellite data."

In recent years, regional and global climate models have been used extensively to study droughts and their causes. The new record is observation-driven and model-independent, so can be used to validate and verify models.

Already the new record has correctly identified several recent major events such as the 2011 Texas, 2010 Amazon and 2010 Horn of Africa droughts. AghaKouchak and Nakhjiri are confident that this combined satellite record will help to identify and monitor future droughts. "Since near real-time satellite observations are available within a few hours to days, one can monitor droughts across the globe in a timely manner," said AghaKouchak. "This product is particularly useful for remote regions and basins with few rain gauges."

In fact the researchers have already used the new record to identify areas that show a significant increase in drought frequency and they hope to publish these findings soon.

AghaKouchak and Nakhjiri are keen to ensure that their new combined record is useful and easy for others to understand. With this in mind they have ensured that their algorithm can be integrated into operational products, to support current activities of the World Meteorological Organization, and the World Climate Research Programme on global drought monitoring.