The first global terrestrial gridded data set of the daily average and range of temperature and daily precipitation has been developed, intended for use in terrestrial biospheric modeling. Data for the year 1987 are shown to illustrate our methodology. Daily station data, primarily from the World Meteorological Organization global synoptic surface network of stations, have been extensively quality checked and interpolated to a 1x1 degree grid by using a nearest neighbors interpolation scheme. Annual averages of the daily average temperatures have been compared with 1987 temperatures constructed from data supplied by P.D. Jones (personal communication, 1996). Agreement between these two data sets is good, except in some areas of the southern hemisphere where station coverage is poor. Monthly and annual totals of the daily precipitation data have been compared with the monthly 1987 data set produced by the Global Precipitation Climatology Centre. Agreement between the two data sets is good over much of the northern hemisphere and South America; however, large discrepancies are seen in east-central and south-central Africa and in most of Australia, primarily due to the poor station coverage there. Comparison of the time series from individual stations with those from the gridded data set indicate that the day-to-day variation of temperature and the fraction of wet days are preserved, except in the tropics where wet days are overestimated. Station densities have been tabulated in terms of total annual net primary productivity to identify countries where increases in station data will be most effective for terrestrial biospheric modeling.
