UC Berkeley

Wind disturbance, along with heavy precipitation and strong wind, caused extensive impact on forests, including leaf falling, tree snapping and uprooting. In this dissertation I investigate the impact of wind disturbance, including cyclones and mesoscale convective system, on temperate and tropical forests, from local to landscape scales. I quantified the wind disturbance impact on forests using remote sensing images, and focus on four primary questions: (1) What are the geophysical and biological variables that affect the forest disturbance intensity caused by hurricane Maria? (2) How do these variables vary across different regions that affect disturbance caused by multiple cyclones? (3) How do climate variables affect windthrow disturbance? (4) How does windthrow disturbance respond to climate warming?In the first chapter, I reviewed 8 remote sensing indices and methods that can measure the windstorms’ impact on forest. I summarized the methods to measure the disturbance intensity and compared their advantages and weaknesses. I also explored the environmental factors in literature that drive the spatial variation of the wind-related forest disturbance. In the second chapter, I measured hurricane Maria’s impact on Puerto Rican forests and developed a GEE remote sensing data analysis tool for rapidly quantifying spatial variability of forest disturbance following a cyclone landfall. I also studied the landscape factors which affect the patterns and severity of forest disturbance intensity. Results show that significant disturbance was found in Luquillo Mountains, the mid-west forested area, the southeastern area, and the northeastern coastal area. Forest type, elevation, green vegetation ratio in the preceding year, distance to hurricane landfall, and distance to hurricane track were the most important variables explaining the variance in forest disturbance caused by hurricane Maria. In the third chapter, I quantified the disturbance intensity caused by multiple cyclones in different regions and explored the feature importance in various cyclone studies. Results showed that climate variables, terrain features, and forest properties were significant tin predicting tree damage caused by cyclones, and wind elevation, and the pre-disturbance vegetation condition were strong predictors. Little consistency was found on the variables among studies of different cyclones, and we believed that the complexities of cyclone effects coupled with landscape, soils, states of affected systems, climate change making the existence of an omnipotent cyclone model very difficult. Moving from cyclones over the world to extreme wind events over Amazon region, in the fourth chapter, I explored 38 case studies of windthrow on the land surface and the corresponding mesoscale convective systems (MCS) in the atmosphere. I studied the relationship between windthrow size and features of MCS, including the duration, coldest cloud top temperature, and rainfall. Results showed that there was a positive linear relationship between the duration of MCS lifespan and the size of windthrows. Deeper convective storms resulted in large windthrows. In the fifth chapter, I investigated the mechanism of windthrow over Amazonia. The results demonstrated that CAPE was an appropriate proxy for estimating the density of windthrow events in the Amazon, and a significant increase in windthrow density over this century was projected under a warming climate. Taken together, these results quantified the impact of extreme storms on forests over the tropical and sub-tropical regions and improved the understanding in the mechanism of land-atmosphere interaction, thus contributing to rapid assessment of extreme wind events impacts and the development of mitigation and forest management strategies under climate change.