INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. 25: 1965���1978 (2005) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/joc.1276 VERY HIGH RESOLUTION INTERPOLATED CLIMATE SURFACES FOR GLOBAL LAND AREAS ROBERT J. HIJMANS,a,* SUSAN E. CAMERON,a,b JUAN L. PARRA,a PETER G. JONESc and ANDY JARVISc,d a Museum of Vertebrate Zoology, University of California, 3101 Valley Life Sciences Building, Berkeley, CA, USA b Department of Environmental Science and Policy, University of California, Davis, CA, USA and Rainforest Cooperative Research Centre, University of Queensland, Australia c International Center for Tropical Agriculture, Cali, Colombia d International Plant Genetic Resources Institute, Cali, Colombia Received 18 November 2004 Revised 25 May 2005 Accepted 6 September 2005 ABSTRACT We developed interpolated climate surfaces for global land areas (excluding Antarctica) at a spatial resolution of 30 arc s (often referred to as 1-km spatial resolution). The climate elements considered were monthly precipitation and mean, minimum, and maximum temperature. Input data were gathered from a variety of sources and, where possible, were restricted to records from the 1950���2000 period. We used the thin-plate smoothing spline algorithm implemented in the ANUSPLIN package for interpolation, using latitude, longitude, and elevation as independent variables. We quantified uncertainty arising from the input data and the interpolation by mapping weather station density, elevation bias in the weather stations, and elevation variation within grid cells and through data partitioning and cross validation. Elevation bias tended to be negative (stations lower than expected) at high latitudes but positive in the tropics. Uncertainty is highest in mountainous and in poorly sampled areas. Data partitioning showed high uncertainty of the surfaces on isolated islands, e.g. in the Pacific. Aggregating the elevation and climate data to 10 arc min resolution showed an enormous variation within grid cells, illustrating the value of high-resolution surfaces. A comparison with an existing data set at 10 arc min resolution showed overall agreement, but with significant variation in some regions. A comparison with two high-resolution data sets for the United States also identified areas with large local differences, particularly in mountainous areas. Compared to previous global climatologies, ours has the following advantages: the data are at a higher spatial resolution (400 times greater or more) more weather station records were used improved elevation data were used and more information about spatial patterns of uncertainty in the data is available. Owing to the overall low density of available climate stations, our surfaces do not capture of all variation that may occur at a resolution of 1 km, particularly of precipitation in mountainous areas. In future work, such variation might be captured through knowledge- based methods and inclusion of additional co-variates, particularly layers obtained through remote sensing. Copyright ��� 2005 Royal Meteorological Society. KEY WORDS: ANUSPLIN climate error GIS interpolation temperature precipitation uncertainty 1. INTRODUCTION Spatially interpolated climate data on grids, here referred to as ���climate surfaces���, are used in many applications, particularly in environmental, agricultural and biological sciences (Hijmans et al., 2003 Jones and Gladkov, 2003 Parra et al., 2004). The spatial resolution of the climate surfaces used in a particular study depends on the needs for that application and on the data available. For many applications, data at a fine (���1 km2) spatial resolution are necessary to capture environmental variability that can be partly lost at lower resolutions, particularly in mountainous and other areas with steep climate gradients. However, such high-resolution data are only available for limited parts of the world, e.g. the Daymet database for the United * Correspondence to: Robert J. Hijmans, Museum of Vertebrate Zoology, University of California, 3101 Valley Life Sciences Building, Berkeley, CA, USA e-mail: rhijmans@berkeley.edu Copyright ��� 2005 Royal Meteorological Society

1966 R. J. HIJMANS ET AL. States (http://www.daymet.org/ Thornton et al., 1997). The highest resolution interpolated climate database for global land areas (excluding Antarctica) has a 10 arc min spatial resolution (18.5 km at the equator New et al., 2002). Leemans and Cramer (1991) and New et al. (1999) created important earlier data sets, at a spatial resolution of 0.5�� (55.6 km at the equator). We compiled monthly averages of climate as measured at weather stations from a large number of global, regional, national, and local sources, mostly for the 1950���2000 period. We interpolated these data using the thin-plate smoothing spline algorithm implemented in ANUSPLIN (Hutchinson, 2004) and created global (land areas only, excluding Antarctica) climate surfaces for monthly precipitation and minimum, mean, and maximum temperature. Our surfaces have a 30 arc s spatial resolution this is equivalent to about 0.86 km2 at the equator and less elsewhere and commonly referred to as ���1-km��� resolution. The data are referred to as the ���WorldClim��� database and are available for download from http://www.worldclim.org. Here, we describe the methods used to compile and interpolate the climate data. We present several analyses to illustrate the uncertainty in both the input data and in the climate surfaces: we describe the elevation bias in the weather stations, quantify within-grid cell variation in elevation, and analyze results from data partitioning and cross validation. We aggregated the climate surfaces to a 10 arc min resolution to illustrate the benefits of higher resolution surfaces and to compare our results to those of New et al. (2002). We also compared our surfaces to two high-resolution data sets that cover the conterminous United States. 2. METHODS 2.1. Climate data compilation and processing Weather station data were assembled from a large number of sources: (1) The Global Historical Climate Network Dataset (GHCN) version 2. (http://www.ncdc.noaa.gov/pub/ data/ghcn/v2 Peterson and Vose, 1997). GHCN reports data by year and month, and we calculated monthly means for the 1950���2000 period. GHCN has data for precipitation (20 590 stations), mean temperature (7280 stations), and minimum and maximum temperature (4966 stations). For precipitation and mean temperature, GHCN has global coverage, but there are large gaps in the geographic distribution of stations with minimum and maximum temperature data. For some stations, ���adjusted��� data that had been through homogeneity control procedures were used (Peterson and Easterling, 1994 Easterling and Peterson, 1995). We included the adjusted data where available and used unadjusted data for the remainder of the stations. (2) The WMO climatological normals (CLINO) for 1961���1990 (WMO, 1996). This database includes monthly mean (3084 stations), minimum and maximum (both 2504 stations) temperature and precipitation (4261 stations). WMO did extensive quality control on these data (WMO, 1996). (3) The FAOCLIM 2.0 global climate database (FAO, 2001). This database contains monthly data. It has precipitation data for 27 372 stations, mean temperature data for 20 825 stations, and minimum and maximum temperature data for 11 543 stations. This database includes long-term averages (1960���1990) as well as time series data for temperature and precipitation. (4) A database assembled by Peter G. Jones and collaborators at the International Center for Tropical Agriculture (CIAT) in Colombia. It includes mean monthly data for precipitation (18 895 stations), mean temperature (13 842 stations), and minimum and maximum temperature (5321 stations). This database has data for the (sub)tropics only and is particularly data rich for Latin America. (5) Additional regional databases for Latin America and the Caribbean (R-Hydronet http://www.r-hydronet.sr. unh.edu/english/), the Altiplano in Peru and Bolivia (INTECSA, 1993), the ���Nordic Countries��� in Europe (Nordklim, http://www.smhi.se/hfa coord/nordklim/), Australia (BOM, 2003), New Zealand (http://www.metservice.co.nz/), and Madagascar (database accompanying Oldeman, 1988). The GHCN data set has undergone the most explicit quality control (Peterson et al., 1997). For this reason, additional records from other databases were only added if they were further than 5 km away from stations already in that database and were added in the order as listed above. In this way, most duplicate records Copyright ��� 2005 Royal Meteorological Society Int. J. Climatol. 25: 1965���1978 (2005)