Radiation fog forecasting using a 1-dimensional model

Abstract

The importance of fog forecasting to the aviation community, to road transportation and to the public at large is irrefutable. The deadliest aviation accident in history was in fact partly a result of fog back on 27 March 1977. This has, along with numerous less dramatic examples, helped focus many meteorological efforts into trying to forecast this phenomenon as accurately as possible. Until recently, methods of fog forecasting have relied primarily on the forecaster's ability to recognize surface weather patterns known to be favorable for producing fog and once it has formed, to state that it will persist unless the pattern changes. Unfortunately, while such methods have shown some success, many times they have led weather forecasters astray with regards to the onset and dissipation of the phenomenon. Fortunately, now with computers becoming ever-increasingly powerful, numerical models have been utilized to attempt to more accurately deal with the fog forecasting problem. This study uses a 1 dimensional model called COBEL to simulate several past fog cases in the hopes of mimicking its actual occurrence and determining what weather parameters the fog is most sensitive to. The goal is to create a technique where the weather forecaster will be able to run several fog forecasts with the model each time with different initial conditions representing the uncertain weather conditions. In this way, the forecaster will be able to use his expertise to choose the most likely scenario. Results indicate that COBEL is able to simulate the fog cases quite well. Issues remain with the model's handling of the gravitational settling rate, the fact that it currently does not include any vegetation, and its coupling process with the soil model. Nevertheless, simulations and sensitivity tests indicate that soil temperature, soil moisture, low-level winds, initial relative humidity, dew deposition and surface emissivity are the weather parameters that affect fog the most. These parameters will be prime candidates for the 1 dimensional ensemble (ODEP) technique described above.