WorldClim: a new high-resolution global climate database
Robert Hijmans, Susan Cameron, and Juan Parra
Climate data at a high spatial resolution are needed for a large number of applications in ecology and resource management. We developed a ~1 km2 resolution global climate database for monthly average temperature (minimum, maximum and mean); monthly precipitation, and derived bioclimatic variables. The data are available for free from http://biogeo.berkeley.edu We used ANUSPLIN software to interpolate data from ~46,000 stations for precipitation, ~25,000 for mean temperature, and ~14,800 for minimum and maximum temperature. The recently released SRTM30 elevation dataset was used as a co-variable in the interpolations. There was strong spatial variation in station density, reflecting population density as well as differences between countries in terms of investment in meteorological observations and access to these data. Station density strongly affects uncertainty in the climate layers, and such uncertainty can be quantified with ANUSPLIN and should be taken into account when the climate layers are used. We had much data for Latin America, thanks to specialized regional databases from CIAT and R-Hydronet, yet there are still considerable gaps in the data, particularly in the Amazon and eastern slopes of the Andes. The international databases we used have resulted from a tremendous amount of long-term efforts and collaboration, both in terms running individual climate stations, and compiling the data. Yet, given the large amount of error and imprecision that we discovered in the data (both in the location of the stations, and in the climate data reported), additional national and international efforts in data collection and verification are needed to fully benefit from these invaluable data.
|
webmaster@cria.org.br
