Prediction of undisturbed ground temperature using analytical and numerical modeling. Part III: Experimental validation of a world-wide dataset

Lu Xing*, Jeffrey D. Spitler, Arkasama Bandyopadhyay

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)


Undisturbed ground temperatures are often required for engineering applications, such as analyses of building heating and cooling load calculations and design of ground heat exchangers. In two companion articles, an automated procedure for generating a world-wide dataset of typical year ground temperatures has been developed and validated against measured results at 19 Soil Climate Analysis Network sites. This procedure relies on a two-harmonic relationship and five weather-related constants—annual average ground temperature, two temperature amplitudes, and two phase lags—to predict the undisturbed ground temperatures. These constant values generated for 4092 sites world-wide covered by short and tall grass are summarized in Appendix A and B in this article. In the United States, a one-harmonic model which relies on three constant values to estimate the ground temperatures is commonly used. The ASHRAE Handbook and ASHRAE District Heating Guide provide these constant values in North America and in a world-wide range, respectively. The estimated typical year ground temperatures using the Xing and Spitler procedure are compared to the measured results at the 19 Soil Climate Analysis Network sites, it shows the mean root mean square errors of all sites are within 1.3°C–1.6°C (2.3°F–2.9°F) at 5–100 cm (2–40) depths, which are 0.7°C–1.4°C (1.3°F–2.5°F) lower than the root mean square errors given by the commonly used procedures in United States. Estimations of the peak (maximum/minimum) ground temperature over multiple years are also important. Thus, a correction factor β is introduced in the two-harmonic model, the newly developed simplified design model could now be used to predict typical year ground temperatures and peak ground temperatures of multiple years. For a typical year, the correction factor is set to be 1. It is suggested that the correction factor of 1.6 to be chosen, so as to use the simplified design model to estimated maximum/minimum ground temperatures of multiple years.

Original languageEnglish
Pages (from-to)826-842
Number of pages17
JournalScience and Technology for the Built Environment
Issue number5
Publication statusPublished - 4 Jul 2017
Externally publishedYes


Dive into the research topics of 'Prediction of undisturbed ground temperature using analytical and numerical modeling. Part III: Experimental validation of a world-wide dataset'. Together they form a unique fingerprint.

Cite this