Student Research Talks
Soundscape Mapping: Understanding Geospatial Patterns in Soundscapes
I will briefly discuss the field of ecoacoustics and focus on research from my dissertation. This includes applying convolutional neural networks to identify ecologically meaningful sounds, applying soundscape metrics in a species distribution modeling framework to understand their spatiotemporal relationships with environmental characteristics, and exploring biotic soundscape responses following wildfires.
Opportunities Using Thermal Imaging
Leaf temperature is central to every canopy-atmosphere exchange. Yet the availability of high-resolution thermal imagery has been limited, inhibiting a clear understanding of these processes. My research aims to analyze thermal data across various scales, from regional (remote sensing using ECOSTRESS) to local (flux tower site). To aid in this effort, we have been collecting and processing two long-term thermal imaging datasets: a mixed temperate forest in Massachusetts (Harvard Forest) and a subalpine conifer forest in Colorado (NIWOT Ridge), which has the world's longest in-situ thermal record. Additionally, we have begun to explore the potential use of more readily available flux data, such as longwave radiation, as a proxy for leaf temperature. We have developed equations that account for tower-based thermal interferences and can apply them to estimate surface temperature using flux longwave radiation. We aim to use these datasets and alternative methods to improve model constraints and better understand the complex interplay between leaf temperature and canopy-atmosphere exchange.
Predicting tree high temperature thresholds in southwest US and the tropics
Identifying Potential Sources of Model Uncertainty Using Functional Benchmarks