Accurate precipitation data is vital to sustainable municipal water management, efficient agricultural and corporate supply chains, and gaining a better understanding of global environmental change. Although there are numerous global precipitation data sets, they generally come in 3 types: 1) gridded interpolated data sets produced from point estimates of a global collection of rain gauges, 2) remotely sensed products that infer precipitation from either visible or infrared spectrums, passive sensors, or active sensors, and 3) reanalysis products that synthesize multiple geo-physical and climatological sources of data to produce high resolution and spatially uniform global precipitation estimates and forecasts.

Despite their seemingly crude technology, gridded rain gauge data products remain extremely popular with academic research, production workflows, and resource management. Their greatest advantage over remotely sensed data is a large historical record. Popular rain gauge data sets have monthly records dating back to 1900, whereas even the oldest satellite precipitation data is limited to 1979. A detailed historical record paramount to global environmental change research and establishing baselines for practitioners and resource managers. In comparison to satellite data, rain gauges are sufficiently accurate and cost effective, however, to leverage their full potential systems must be in place to centralize their data and perform quality control. With approximately 100,000-250,000 rain gauges in existence this can be an extraordinary logistical challenge.¹ The World Meteorlogical Organization (WMO) is the primary source for rain gauge data organization through maintenance of the WMO Global Telecommunication System and co-sponsorship of the Global Climate Observing System.

Three of the most widely employed and heavily cited rain gauge precipitation data products are the University of Delaware’s Terrestrial Precipitation Gridded Time Series (UDEL-TS),² the Climate Research Unit of the University of East Anglia’s Gridded Time Series dataset (CRU-TS v4.04),³ and the Global Precipitation Climatology Centre (GPCC).⁴ GPCC provides global monthly data from 1891-2016 at 0.5 degree resolution in addition to multiple other precipitation statistics. GPCC data has recently been featured in investigations of the effects of anthropogenic warming on California droughts,⁵ high resolutions climatologies for the earth’s land surfaces,⁶ and the effects of pervasive drought in forest ecosystems.⁷

The GPCC and CRU-TS provide nearly identical temporal coverage and resolution, although they rely more on national meteorological agencies, the WMO, and the Food and Agriculture Organization for rain gauge sources. The UDEL-TS utilizes gauges from the Global Historical Climatology Network, Daily Global Historical Climatology Network, Atmospheric Environment Service archive, Hydrometeorological Institute, GC-Net, the Global Surface Summary of Day, and several other regional and global rain gauge networks. In contrast to UDEL-TS and CRU-TS that implement cross validation and outlier detection, GPCC requires a minimum of 10 uninterrupted years for each station to be included in the dataset.

It’s important to review spatial and temporal coverage when deciding which gridded rain gauge data product to incorporate into your research or other workflow. A recent review of precipitation data sets found that UDEL-TS, CRU-TS, and GPCC generally exhibit consistent inter-annual variability, however, differences can be as great as 100mm.⁸ The three popular gauge data sets track exceptionally well in tropical zones, but provide divergent estimates in areas with low population and complex topography such as northern Africa, northern North America, eastern Russia. Conversely, differences in seasonal precipitation estimates between UDEL-TS, CRU-TS, and GPCC were negligible.