Using Tree Damage For Analyzing Wind Intensity

The 2017 Hurricane Maria transpired into the most devastating tropical cyclone in Puerto Rico. The storm moved diagonally northwestward from the southern tip of the island at a reported speed of 250 km/h at landfall. Consequently, Maria’s strong wind destroyed nearly all weather stations thus voiding ground wind speed information as the storm passed through the island. The notion of this study is to further use the broken trees on the island as sensors to estimate the wind speed of Hurricane Maria and its distribution through wind-tree breakage mechanics calculation. The notion of this study is to further use the broken trees on the island as sensors to estimate the wind speed of Hurricane Maria and its distribution through wind-tree breakage mechanics calculation. Island-wide pre- and post-LiDAR point clouds were processed into canopy height models for individual tree segmentation. More than 150 million trees were delineated with each stem diameter at breast height (DBH) modeled from canopy height and crown size. Overall, the island’s forest height remains lower a year after the hurricane, and trees that are taller and have smaller stem diameters experienced greater height loss. The difference in canopy height shows the eastern part of the island having a greater percentage of broken trees with areas toward the landfall in the southeast experiencing the largest portion of breakages up to 60% broken trees. The resulting wind speed at 10 m above tree surface (U10) derived from critical wind speeds of individual trees are then aggregated by a 115 km2 range in a gradient from 100 km/h on the southwestern coast to nearly 260 km/h on the southeastern coast at the landfall. Areas near the track have significantly higher wind speeds above 200 km/h. A comparison of the U10 speed to the remaining NOAA maximum gust readings shows a mean difference of 15.4 km/h with the largest difference at Culebra and landfall. However, the modeled U10 shows a wind speed of 257 km/h at the landfall area, which is close to the 250 km/h speed described in the NOAA official report. While such discrepancy highlights the importance of ground wind-speed information, the U10 wind speed estimated from the LiDAR-based tree model demonstrated the possibility of further utilizing high-resolution remote sensing in analyzing the complex relationship between hurricanes and the earth’s surface which is useful for future scenarios.