Small-scale climate maps: A sensitivity analysis of some common assumptions associated with grid-point interpolation and contouring

From Shepard’s (1968) local-search method, algorithms are developed for contouring on spherical surfaces and in Cartesian two-space. These algorithms are used to investigate errors on small-scale climate maps caused by the common practice of interpolating-from irregularly- spaced data points to regular-lattice nodes-and contouring in Cartesian two-space. Using mean annual air temperatures drawn from 100 irregularly-spaced weather stations, the annual air- temperature field over the western half of the northern hemisphere is estimated both on the sphere (assumed to be correct) and in Cartesian two-space. When these fields are mapped and compared, error magnitudes as large as 5° to 10° C appear in the air-temperature field approximated in Cartesian two-space.