As an aid to interpolatory estimation, normals of monthly mean, mean maximum, and mean minimum recorded at 206 climatological stations have been graduated by least squares formulae constructed from polynomials of moderate degree in latitude, longitude, and altitude coordinates. Seasonal variation was evident in all computed temperature gradients, indicated elevation effects being maximal in summer and those associated with latitude and longitude most pronounced in winter and spring. These empirical formulae did not account completely for all systematic features of the data, but resulted nevertheless in correlation coefficients of r ≥ 0.95 between recorded and calculated mean temperatures for all months in this relatively homogeneous geographical region. An illustration is given of their possible application to approximate isotherm location by inverse interpolation.

You do not currently have access to this article.