A Global Assessment of Bimodality Underlying Normalized Drought Indices (SPI and SPEI)
This global assessment confirms that the unimodal distribution assumption underlying normalized drought indices like SPI and SPEI is valid for most locations and applications, with bimodality being a rare phenomenon primarily limited to short accumulation periods in arid, highly seasonal regions where it often stems from data artifacts rather than true climatic processes.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
For decades, meteorologists and water managers have relied on a specific way of measuring drought to decide when a region is in trouble. They do not just look at how much rain fell yesterday; they look at how much rain has fallen over a period of time, such as three months or a year, and compare that total to what is normal for that specific time of year. This comparison creates a score that tells us if a place is wetter or drier than usual. To turn these raw numbers into a score, scientists use a mathematical tool that assumes the rain patterns in any given place follow a single, smooth curve. Imagine a hill with one peak in the middle; this shape suggests that most years are average, with fewer years being very wet or very dry, and that the transition between these states is gradual. This assumption has been the foundation of global drought monitoring, but until now, no one had thoroughly checked if the real world actually fits this single-hill shape everywhere.
A team of researchers set out to test this long-held assumption across the entire planet. They wanted to know if the rain data used to calculate drought scores ever forms two distinct peaks instead of one. In the real world, a two-peaked pattern would mean a place has two very different types of weather years: perhaps a cluster of very dry years and a separate cluster of very wet years, with very few years in between. If this "double-hill" pattern exists, the standard mathematical tools used to measure drought might be misleading, potentially hiding the true severity of dry spells or misidentifying wet ones. The researchers examined global climate records, looking at precipitation data from 1901 to 2021, to see if this double-peaked behavior was a common feature of our climate or a rare exception.
The study, which analyzed data from two major global climate datasets, found that the worry about a double-peaked pattern is largely unfounded for most of the world. The researchers discovered that the single-hill assumption holds true in the vast majority of locations and for almost all seasons. Bimodality, the scientific term for having two peaks, was found to be a rare event. When it did appear, it was almost exclusively in hot, dry regions like the edges of the Sahara, the Arabian Peninsula, and parts of Australia and South America. Even in these dry places, the double-peaked pattern was only visible when looking at very short time windows, such as a three-month accumulation of rain. As the researchers looked at longer periods, like six months or a full year, the double peaks disappeared, merging back into a single, smooth curve. This suggests that while a specific month in a dry desert might have two distinct weather modes, the climate over a longer stretch of time smooths out these differences.
The researchers also investigated why some data appeared to show two peaks when it shouldn't. They found that many instances of apparent double-peaked patterns were not caused by nature at all, but by errors in the data itself. In some locations, the records contained long stretches of identical numbers, likely because missing data was filled in with average values without proper flags. In other places, the data was rounded into a few specific numbers, creating artificial spikes that looked like a second peak. Once these data errors were removed, the number of locations showing a true double-peaked pattern dropped significantly. The study also looked at how the inclusion of days with zero rain affected the results. While including zero-rain days did increase the appearance of double peaks in dry areas, the researchers noted that standard drought calculations usually handle these zeros separately, so this is not a major practical problem for most users.
When the researchers compared the standard precipitation index with a more complex version that also accounts for evaporation and plant water use, the results remained consistent. The complex version showed even fewer instances of double peaks, suggesting that adding more physical details to the calculation actually makes the data look more like a single, smooth curve. The study concluded that the fear of bimodality violating the rules of drought monitoring is overstated. For the overwhelming majority of the globe, and for the time scales most commonly used by drought monitors, the single-hill model is accurate. The few places where the model might struggle are limited to specific, strongly seasonal dry climates between 30 degrees north and 30 degrees south latitude, and only when looking at very short time frames.
For the scientists and officials who track drought, this finding is a relief. It means they can continue to use the established, reliable methods for calculating drought scores without needing to overhaul their systems for most of the world. However, the study does offer a specific warning for those working in arid regions with sharp wet and dry seasons. In these specific zones, particularly when using short three-month windows, it is wise to take a quick look at the data to ensure it does not have two distinct peaks. If a double peak is found, there are more advanced mathematical tools available to handle it, but these are needed only in a tiny fraction of cases. The research confirms that the simple, single-curve approach is robust, handling the complexity of global climate with surprising accuracy, and that the rare exceptions are usually just quirks of the data rather than a fundamental flaw in the science.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.