Description
This is a benchmark dataset for comparing a number of methods of spatial prediction, found in: A comparison of spatial predictors when datasets could be very large by Jonathan R. Bradley, Noel Cressie, and Tao Shi, which can be found here. This dataset reports level-2 mid-tropospheric CO_2 values at a 17.6 km × 17.6 km spatial resolution, which is obtained from Atmospheric Infrared Sounder (AIRS) data retrieved from 1–9 February 2010. AIRS is a remote sensing instrument on board the Aqua satellite administered by the National Aeronautics and Space Administration (NASA). Among other measurements, it collects CO_2 measurements in the form
of spectra (level 1) that are then converted to mid-tropospheric CO_2 values (level 2) given in units of parts per
million (ppm).
This dataset is in the form given by Bradley et al. (2016) and is freely available under the Creative Commons Attribution 4.0 Australia License.
The ZIP file contains three folders, "Small," "Large," and "VeryLarge," the data in these folders are used in a comparison study in Section 4 of Bradley et al. (2016).
In each folder, there are two excel csv files, respectively for the training and the validation datasets. In each excel file, the first two columns are the latitude and longitude, respectively. The third column is mid-tropospheric CO_2 in ppm.
Reference
Bradley, J.R., Cressie, N., and Shi, T. (2016). A comparison of spatial predictors when datasets could be very large. Statistics Surveys, 10, 100-131.