Estimation of frost occurrence and duration of frost for a short-grass surface

Abstract

The estimation of frost duration (FD) was investigated using dielectric, infrared surface temperature and grass temperature subhourly measurements. Near real-time FD data and information displays and alerts were also made available via a web-based system. FD was estimated using a dielectric leaf wetness sensor (LWS) method, for which the sensor voltage was between 274 and 284 mV with a voltage rate of change less than 10 mV h⁻¹ for a 4 min period, and two temperature methods for which infrared thermometer (IRT) and grass temperatures were compared with 0 °C. FD estimation using the LWS method ensured that most of the transitional dry-to-wet and wet-to-dry events were not included in the FD count. Generally, the IRT method yielded the largest estimate of FD, grass temperature method lower and LWS method lowest. Micrometeorological measurements showed consistent air temperature gradients of 2.25 °C m⁻¹ for cloudless nocturnal frosted conditions with few air temperature measurements at 1 m and none above indicating frost occurrence. At the very least, automatic weather station systems should contain a grass thermometer or preferably an IRT for determination of FD with near real-time data and graphics displayed, including timeous alerts of frost occurrence and FD, using the Internet.