An algorithm for the construction of spatial coverage designs with implementation in SPLUS
Computers & Geosciences
K-means clustering via principal component analysis
ICML '04 Proceedings of the twenty-first international conference on Machine learning
Least squares quantization in PCM
IEEE Transactions on Information Theory
Design of computer experiments: space filling and beyond
Statistics and Computing
Value of information and mobility constraints for sampling with mobile sensors
Computers & Geosciences
Optimal reduction of a spatial monitoring grid: Proposals and applications in process control
Computational Statistics & Data Analysis
| Hi-index | 0.00 |
Both for mapping and for estimating spatial means of an environmental variable, the accuracy of the result will usually be increased by dispersing the sample locations so that they cover the study area as uniformly as possible. We developed a new R package for designing spatial coverage samples for mapping, and for random sampling from compact geographical strata for estimating spatial means. The mean squared shortest distance (MSSD) was chosen as objective function, which can be minimized by k-means clustering. Two k-means algorithms are described, one for unequal area and one for equal area partitioning. The R package is illustrated with three examples: (1) subsampling of square and circular sampling plots commonly used in surveys of soil, vegetation, forest, etc.; (2) sampling of agricultural fields for soil testing; and (3) infill sampling of climate stations for mainland Australia and Tasmania. The algorithms give satisfactory results within reasonable computing time.