Daily systematic meteorological observations made by George Marcgrave in Recife from 1640 to 1642 have been retrieved. These observations were published in 1658 by Piso in “De Indiae utriusque re naturali et medica”. Specifically, in chapter two “De Aeris temperie atque Anni tempestatibus” of the first book entitled “Tractatus topographicus et meteorologicus Brasiliae, cum observatione eclipsis Solaris”. Marcgrave recorded meteorological observations during six years with sub-daily resolution, nevertheless Piso only published the daily observations relative to the years of 1640, 1641 and 1642. The observations are recorded in three tables, one per year, that show the wind direction and the presence of rainfall. Rainy days are identified with a “p.” (from pluvia), about the meaning of a rainy day for Marcgrave, he writes “In these tables, we have noted with the letter P all the days on which even a very fine and brief rain fell so that no one thinks that during those days it rained continuously”. The wind direction is registered in 24 directions using letters according to the Franco names e.g. N (Nordoni), E (Ostroni), S (Sundroni), W (Vuestroni). This is known thanks to a Marcgrave's annotation “We have annotated the winds with letters: S.O. designates the Euro, N.O. the Boreas or Aquilon”. SO are the initials of Sundostroni and correspond with the SE direction in which the Euro wind blows. NO are the initials of Nordostroni, which corresponds with the NE direction in which the Aquilon or Boreas blow. The chapter shows not tabulated information in text format among tables, this informs about the days with fog, thunders, and lightning. All the meteorological information has been digitized by 'key entry'. These observations are the first systematic meteorological observations from the Americas and from the southern hemisphere. Although the observations retrieved cover only three years, they are very interesting due to the singular climate forcing situation i.e. one solar cycle before (1635-1645) the Maunder minimum (1645-1715) and two eruptions with a clear impact on the climatic system in 1640 (Parker, Philippines and Hokkaido-Komagatake, Japan).

Empty cells mean that the observer does not mention the presence of rain/thunder/lightning/fog occurred on that day.This dataset has emerged as part of project 'VARIABILIDAD DE LA SEQUÍA EN LA PENÍNSULA IBÉRICA EN LOS ÚLTIMOS 500 AÑOS, UNA APROXIMACIÓN METODOLÓGICA', Award PID2019-108589RA-I00.