LAUSR

Gridded climate datasets are among the most used datasets in the study of weather and climate. Given the large range of such products currently available, investigators must consider their strengths and weaknesses. This research evaluated the comparative performance of two reanalysis datasets (ERA‐Interim, and Agri4Cast [e.g. MARS‐STAT]) and one reprocessing dataset (E‐OBS) in replicating selected spatiotemporal characteristics of precipitation and drought at 20 wine production regions in Greece during 1981–2012. The results highlighted the abilities of E‐OBS and Agri4Cast, whose performance varied with season and the specific rainfall characteristic in question. The former product: (1) reproduced better the annual decreasing trends in spring, summer, and autumn, (2) captured a larger portion of the observed monthly variability in spring and summer, associated better and showed the lowest errors with observations, and (3) computed better or equal to Agri4Cast different skill scores applied on the daily time series, replicated better the probabilities of wet and dry days and equally well with Agri4Cast the extreme precipitation indices. Agri4Cast replicated better the monthly cycle of precipitation (underestimating though the station averages) and the increasing trend from SE to NW, on spatial basis. With regards to drought monitoring, frequency and detection of very and extremely wet and dry spells, SPI was superior to SC‐PDSI (the self‐calibrated version of Palmer's index), particularly when E‐OBS data was used. Agri4Cast and E‐OBS are the most appropriate products for use in climate/atmospheric‐related research over Greece.