The following model simulates hourly temperature fluctuations at 6 Kansas stations: \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \[T_{h}=\frac{(T_{x}-T_{n})}{2}\left[\mathrm{exp}\left(\frac{0.693h}{DL_{M}}\right)-1\right]+T_{n};{\ }0{\leq}h{\leq}DL_{M}\] \end{document} \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \[T_{h}=\frac{(T_{x}-T_{n})}{2}\left[1+\mathrm{sin}\frac{{\pi}(h-DL_{M})}{2(23-DL_{M})}\right]+T_{n};{\ }DL_{M}{\leq}h{\leq}23\] \end{document} where h = time (hours after sunrise), DLM = 20.6 - 0.6 * daylength (DL), Th = temperature at time h, and TX and Tn = maximum and minimum temperature, respectively. Required inputs are daily TX and Tn and site latitude (for the calculation of DL). Whereas other models have been derived by fitting equations to chronological temperatures, this model was derived by daily fitting of hourly temperatures sorted by amplitude. Errors from this model are generally lower, and less seasonally biased, than those from other models tested.