Hamid Dashti Dissertation Defense

Title: Characterizing dryland ecosystems using remote sensing and dynamic global vegetation modeling

Abstract: Drylands include all terrestrial regions where the production of crops, forage, wood and other ecosystem services are limited by water. These ecosystems cover approximately 40% of the earth terrestrial surface and accommodate more than 2 billion people (Millennium Ecosystem Assessment, 2005). These fragile ecosystems are prone to ecological disturbances that are within the range of conditions naturally experienced by the ecosystem (e.g. fire, extreme climatic events), and also stressors which are outside these ranges and typically have an anthropogenic origin (e.g. fire regime alteration, land use change). Understanding the dynamics of such ecosystems is significant for assessing the potential for and impacts of disturbances and mitigation planning and a necessary step toward enhancing the economic and social well-being of dryland communities in a sustainable manner (Global Drylands: A UN system-wide response, 2011). In this research, a combination of remote sensing, field data collection, and ecosystem modeling were used to establish an integrated framework for semi-arid ecosystems dynamics monitoring.

Foliar nitrogen (N) plays an important role in vegetation processes such as photosynthesis and there is wide interest in retrieving this variable from hyperspectral remote sensing data. Since foliar N is not represented in canopy radiative transfer models, empirical methods based on machine learning techniques is common in N retrieval. In this study, I used the theory of canopy spectral invariants (AKA p-theory) to understand the role of canopy structure and soil in the retrieval of foliar N from hyperspectral data. The results of this study showed the inconsistencies among different machine learning techniques used for estimating N. Using p-theory, I demonstrated that soil can contribute up to 95% to the total radiation budget of the canopy. I suggested an alternative approach to study photosynthesis is the use of dynamic global vegetation models (DGVMs). Gross primary production (GPP) is the apparent ecosystem scale photosynthesis that can be estimated using DGVMs. In this study, I performed a thorough sensitivity analysis and calibrated the Ecosystem Demography (EDv2.2) model along an elevation gradient in a dryland study area. I investigated the GPP capacity and activity by comparing the EDv2.2 GPP with flux towers and remote sensing products. The overall results showed that EDv2.2 performed well in capturing GPP capacity and its long term trend at lower elevation sites within the study area; whereas the model performed worse at higher elevations likely due to the change in vegetation community. I discuss that adding more heterogeneity and modifying ecosystem processes in ED.v2.2 will improve its application to higher elevation ecosystems where there is more vegetation production. And finally, I developed an integrated hyperspectral-lidar framework for regional mapping of xeric and mesic vegetation in the study area. I showed that by considering spectral shape and magnitude, canopy structure and landscape features (riparian zone), we can develop a straightforward algorithm for vegetation mapping in drylands. This framework is simple, easy to interpret and consistent with our ecological understanding of vegetation distribution in drylands over large areas.   Collectively, the results I present in this dissertation demonstrate the potential for advanced remote sensing and modeling to help us better understand ecosystem processes.

 

 

Where: Riverfront Hall 101
When: Thursday, October 3rd
Time: 3:30pm
Directions: Map