Geometric shapes of coherent structures such as ramp or cliff like signals, step changes and waves, are commonly observed in meteorological temporal series and dominate the turbulent energy and mass exchange between the atmospheric surface layer and the layers above, and also relate with low-dimensional chaotic systems. In this work a simple linear technique to extract geometrical shapes has been applied at a dataset which was obtained at a location experiencing a number of different mesoscale modes. It was found that the temperature field appears much better organized than the wind field, and that cliff-ramp structures are dominant in the temperature time series. The occurrence of structural shapes was related with the dominant flow patterns and the status of the flow field. Temperature positive cliff-ramps and ramp-cliffs appear mainly during night time and under weak flow field, while temperature step and sine structures do not show a clear preference for the period of day, flow or temperature pattern. Uniformly stable, weak flow conditions dominate across all the wind speed structures. A detailed analysis of the flow field during two case studies revealed that structural shapes might be part of larger flow structures, such as a sea-breeze front or down-slope winds. During stagnant conditions structural shapes that were associated with deceleration of the flow were observed, whilst during ventilation conditions shapes related with the acceleration of the flow.