Reconstructing the past and modeling the future of wetland dynamics under climate change
Author
Halabisky, Meghan Andrea
Metadata
Show full item recordAbstract
Abstract Reconstructing the past and modeling the future of wetland dynamics under climate change Meghan Halabisky Chair of Supervisory Committee: Dr. L. Monika Moskal School of Environmental and Forest Sciences Wetland ecosystems are widely considered to be highly sensitive to climate change. However, scientific capacity to model climate impacts to wetlands has been hampered by the lack of accurate maps showing the spatial distribution of wetlands and data on their historical hydrological dynamics. Though these data may exist for particular wetlands, there are no broad scale datasets of wetland location and long-term hydrological dynamics. Remote sensing has been an important vehicle for mapping change to wetlands, but generally at spatial or temporal scales that do not capture the variability necessary for linking climate to wetland hydrodynamics. This data limitation and lack of methods have restricted research on how changes in climate will impact wetland hydrology to explorations of limited scope. The goal of this PhD was to characterize and model historic and future climate impacts to dynamics of wetland hydrology (i.e. inundation quantity, frequency, timing and duration) across the Columbia Plateau ecoregion. To achieve this goal, I developed new remote sensing methods to map and reconstruct wetland dynamics for thousands of individual wetlands at finer temporal and spatial resolutions than previously available (Chapter 1 and 2). In Chapter 1, I combined high-resolution aerial photographic imagery and a time series of Landsat satellite imagery to reconstruct wetland inundation patterns for individual wetlands from 1984 – 2011 in Douglas County, WA, USA. A key component of this method was the ability to measure fine scale changes (<30m) in surface water area using a sub-pixel technique called spectral mixture analysis. In Chapter 2, I adapted these methods so they could be scaled up to large extents without the computer processing requirements and technical challenges of using aerial imagery. In order to do this, I identified wetlands, not from the spectral and spatial characteristics one can derive out of aerial imagery as in Chapter 1, but instead using their temporal pattern of flooding and drying derived from the time series of Landsat satellite imagery. Using the methods developed in Chapter 1 and Chapter 2, I mapped and reconstructed wetland hydrodynamics for wetlands in the Columbia Plateau ecoregion, far surpassing any existing measurements of wetland hydrology in sample size (n= 5,382), temporal richness (~ 23 days), and temporal extent (27 years). Finally, in Chapter 3 I used this novel dataset to map changes in wetland hydrology across the Columbia Plateau identifying areas undergoing change. Additionally, I developed wetland-specific regression models to understand the relationship between climate and wetland hydrology, which I used to forecast changes to wetland hydrology under climate change. Beyond the technical analyses, an additional important part of the process for Chapter 3 was working with wetland practitioners from start to finish to ensure the data developed is both useful and used. The findings of this research suggest that wetlands in the Columbia Plateau are hydrologically variable with each wetland falling along a continuum from those driven primarily by surface water (i.e. precipitation, evaporation, and surface runoff) to those driven primarily by deep groundwater sources. The location of each wetland along this continuum, which I was able to approximate, varies greatly throughout the region, but follows a defined spatial pattern related to underlying geologic processes. Where a wetland falls along the groundwater to surface water continuum largely determined historical changes in inundation levels and how a wetland will respond in the future under climate change. In general, water levels in groundwater driven wetlands have typically decreased since 1984, whereas water levels in surface water driven wetlands have increased or stayed at similar levels over the same period. However, under the climate change scenario selected (ECHAM5 A1B) the results from the wetland-specific regression models suggest that groundwater driven wetlands will increase in water levels and dry less frequently. On the other end of the wetland continuum, surface water wetlands will decrease in surface water levels, dry more frequently, dry earlier in the season, or have little change. The results of this PhD provide an example of how remote sensing can deliver the fine scale detail and broad temporal and spatial extent necessary to model complex ecosystem dynamics. This knowledge is being used to inform the development of strategies to conserve the biodiversity supported by these systems, and prioritize and help stratify wetlands for further study and conservation action in the Columbia Plateau.
Collections
- Forestry [391]