Using Airborne Lidar Data to Link Local Surface Features to Observed Wind Statistics

Small scale surface features can have profound effects on wind observations. These features are poorly resolved or unresolvable by numerical weather models and can result in large biases when verifying forecasts. Cao and Fovell 2018 proposed Gust Factor (GF), the ratio between observed gust and mean wind speeds, could be used as a simple numeric to quantify the relative exposure of a site and account for the influence of local features. However, correlating GF to significant surface features can be difficult due to lack of information about the environment surrounding a site. Satellite photography lacks accurate measurements of size and height, while direct site surveys are rare and only provide information in the immediate area. Luckily, airborne lidar presents a unique opportunity to visualize a high resolution three dimensional environment around an observation site. Lidar scans generate highly dense three dimensional point clouds with resolution of approximately 1m.

The New York State Mesonet (NYSM) is a recently established high quality network of 126 stations that are sited across NYS in a wide variety of environments. Wind observations from the NYSM are coupled with airborne lidar scans to analyze the relationship between obstructions and wind speeds, gust speeds, and GFs. A technique from Wolfson and Fujita 1988 for relating wind observations to panoramic photos around sites is adapted. Obstruction angles of objects are calculated from the lidar point clouds for both obstacles and terrain, smoothed azimuthally, and correlated to wind observations. Mean wind speeds exponentially decay with obstruction angle, whereas gust speeds exhibit only a small negative relationship with obstruction angle. GFs are well correlated to obstructions at both the individual station and network average level. This relationship is dominated by prominent obstacles within 200m of the station or when obstacles are sparse, terrain patterns up to 2km away.