Faculty Seminar Series: March 10
Resilient and resistant urban forests: the role of statistical models in understanding forest structure and response to disturbance
Dr. Christina Staudhammer
University of Alabama
March 10, 12 – 1 pm
FSC Caseroom #2916
Abstract
Hurricanes can result in extensive damage to urban forests, with uprooted and downed trees as well as severe defoliation and limb loss. Previous research suggests spatial heterogeneity in tree damage, even when considering differences in wind speed, with patterns related to landscape context, species composition, and community socioeconomic factors. Urban forests provide a wide variety of ecosystem services, including decreased pollution, flood risk and climate mitigation, and carbon sequestration. Moreover, they are an ideal system to study disturbance impacts as a proxy for future climate conditions.
Prior to Hurricane Irma (2017), urban tree inventories were conducted in three cities in Florida, with tree- and plot-level measurements such as tree species, diameter at breast height (DBH), tree height, tree density and ground covers collected. After Irma, a subset (215) of the originally inventoried plots were revisited and reassessed, noting branch damage as well as damage to the main stem. While our primary aim was to identify the significant drivers of windstorm damage across the three urban systems using plot- and tree-level factors, a secondary aim was to develop analytical methods to recognize uncertainty in damage assessments, as lag times between the storm and post-storm ground sampling add additional uncertainty to estimates of observed tree damage. Using analysis methods more commonly utilized in wildlife ecology, we can test methods that explicitly include uncertainty in responses traditionally presumed as fixed and binary. Exploring alternatives to logistic regression models, we investigate the impact of observational uncertainty to analyze the relationship between wind damage and plot and tree level factors. In particular, binary models were more sensitive, indicating a stronger damage response related to tree size and significant differences in damage by land use across the study sites. Regardless of model type, we found no difference in tree failure rates between non-native and native species. Lack of native tree resistance across these systems may be a consequence of the inherent stressors of urban environments, where few remnants of a potentially protective natural forest structure remain. Additionally, we found that both native and non-native varieties of oaks (Quercus) failed at higher rates than would be expected in all three cities, refuting previous research in this area. While this research contributes to our understanding of urban tree resistance, it also indicates a need to consider uncertainty in predicting tree failure due in part to the complexity of urban systems.
Bio
Christina Staudhammer’s research covers a broad range of topics across forest inventory and analysis, as well as ecological statistics, urban trees and urban forests. Her expertise is in statistical models, design and analysis of experiments, and model uncertainty across a range of topics in urban forestry and natural resources. Her research aims to promote better management of natural resources by developing a better understanding of the dynamics of these systems. Her research over the last decade has been focused in three main areas: 1) enhancing models of urban forest structure and growth, and quantifying how trees respond differently to disturbances and along urban to rural gradients, 2) advancing models of carbon fluxes derived from eddy covariance (EC) studies, and 3) developing models characterizing the population dynamics of Bertholletia excelsa (Brazil nut).
N.B. The Forest Biometrics lab is hosting Dr. Staudhammer’s visit. We are scheduling meetings with her on March 10 and 11. If you are interested, please reserve a spot here >> or contact Nguyet Anh.
We are also hosting a pizza dinner on campus on Wednesday, March 11, starting at 5:00pm. If you are interested in attending the dinner, please contact Nguyet Anh for details.