A brief study to understand the physical driving forces concerning ice growth. These forces are mathematically described in a basic prediction model. Further complexity to the model is added through adopting various assumptions about the physical factors involved.

Nocturnal weather can have significant impact on society. Dangerous conditions can occur after clear sky nights with weak winds. The lack of turbulent mixing during these nights can cause the near-surface temperature to rapidly decrease. A single short-lived turbulent event can renew turbulent mixing and hamper/stop the formation of fog or severe frost, averting dangerous conditions for e.g. Short-lived turbulence has been studied using single tower observations, but the spatial scale of this phenomenon is not yet known. The goal of this study is to fill that gap and estimate the spatial scale of short-lived turbulent structures in the nocturnal boundary layer.

Short-lived turbulent structures are studied using wind speed and near-surface temperature observations from existing observational networks in the Netherlands. Nights with short-lived turbulence are selected using an automated filter and manual selection. Visual analysis is first used to identify that the turbulent structures span multiple stations of the network. If so, the correlation between pairs of stations is quantified as a function of their spatial separation. Next, an exponentially decaying fit is applied to this data. The e-folding distance of this fit serves as the estimation of the spatial scale of the turbulent structure.

A dense network is found by using the observational masts at and around Schiphol Airport. Visual analysis shows that this network is sufficiently dense for the turbulent structures to span multiple stations. Using the aforementioned methodology, the scales of the short-lived turbulent structures are estimated to range between 1.5 km and 24 km, with a median of 9 km. The spatial scale of short-lived turbulent structures is therefore estimates at 9 km.
This study should be considered to be a first-order estimate. In fact, one of the weaknesses of the method used is that it only takes the distance between stations into account (without incorporating directional effects). As such, it assumes the turbulent event to be spatially isotropic. At present, limitations by the network do not allow for a clear assessment of the spatial structure of the events.
To accurately study both the dynamics and the spatial coherence of short-lived turbulent structures, a dense and uniformly distributed network is recommended. Ideally, such set-up would consist of a nested fine-scale network within a courser network which needs to span a distance of 25 km minimum to capture turbulent events such as the ones studied here.

In addition to turbulent events, regional temperature variations are found. Near-surface temperature differences up to 9 degrees over a distance of 40 km are observed during nighttime. The daytime temperature at these stations is nearly identical, stressing the differences in atmospheric physics during day and night. These local variations may result from different local (soil) conditions, local turbulence and clouds. Artificial turbulence, caused by large airplanes taking off, is observed close to the airport’s runways, increasing the local near-surface temperature up to 4 degrees. It will be interesting to also study the impact of artificial turbulence on the near-surface temperature as a follow up of the current work.