IMPROVING ESTIMATES OF ROOT ZONE SOIL MOISTURE TO UNDERSTAND ECOSYSTEM DROUGHT SENSITIVITY
Open Access
- Author:
- Baldwin, Douglas C
- Graduate Program:
- Geography
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- December 11, 2015
- Committee Members:
- Erica A H Smithwick, Dissertation Advisor/Co-Advisor
Erica A H Smithwick, Committee Chair/Co-Chair
Alan MacEachren, Committee Member
Kenneth Davis, Committee Member
Klaus Keller, Outside Member - Keywords:
- Soil moisture
Bayesian
MCMC
Root zone
Remote sensing
Kalman filter
Drought
Carbon - Abstract:
- Extreme droughts can drastically inhibit the growth of terrestrial forest ecosystems, limiting their ability to assimilate and sequester atmospheric carbon dioxide, an important ecosystem service. Temperate forests in particular are an important component of Earth’s terrestrial carbon ‘sink,’ but ecosystem carbon (C) flux monitoring since the 1990s shows that C uptake in temperate forests is more vulnerable to drought than previously estimated. Climate models typically project that drought events will become more frequent and intense across temperate regions in coming decades. As a result, soil moisture is an increasingly crucial variable to monitor for diagnosing and forecasting drought occurrence and severity. Advancements in satellite technology now facilitate more precise and accurate daily retrievals of global near-surface soil moisture content but satellite data must be downscaled to finer resolutions before it can be used in hydrologic models to estimate subsurface root zone soil moisture content over broad spatial extents. This dissertation explores how a combination of ground based measurements, satellite data, and hydrologic modeling can be used to generate high-resolution predictions of root zone soil moisture content. Root zone soil moisture observations are also used to evaluate the sensitivity of temperate forest growth, estimated by eddy flux towers, to drought in multiple locations in northern temperate forests in the United States. The opening chapter reviews research that connects root zone soil moisture to plant water stress and the development of hydrologic models that estimate root zone soil moisture across large areas using satellite soil moisture data. Chapter two demonstrates how an ensemble Kalman filter (EnKF) data assimilation algorithm can be integrated with a physically based hydrologic model that predicts root zone soil moisture with satellite data. This model system can produce root zone soil moisture predictions close to observed in situ measurements. The model is also rigorously tested in a ‘perfect model test’ experiment before its application with in situ data. The spatial extension of this model from point locations to a watershed-scale gridded surface was developed and tested with field observations at the Shale Hills Critical Zone Observatory (Chapter 3). The integration of a distance decay function with distance metrics calculated with high resolution LiDAR data into the algorithm approximated a spatial pattern of root zone soil moisture content that was comparable with the geostastical regression kriging approach. Ecosystem C flux modeling in Chapter 4 identifies potential critical points where temperate forests may become net carbon sources in high growth conditions under a range of temperate and root zone soil moisture states. There is evidence that soil physical properties and hydric regime (i.e., percentage of wetlands) influence these critical points. This dissertation uncovers new research challenges associated with large-scale root zone soil moisture mapping and drought limitations on temperate forest growth. Soil survey and terrain data are crucial for high precision estimation of satellite soil moisture, but the classification of soil-landscape units and subsequent root zone soil moisture mapping at other watersheds and regions should be investigated. The analysis in Chapter 4 shows that relatively old temperate forests with coarse textured soils can be sensitive to drought, but growth rates can also drop exponentially in relatively young forests beyond a temperature and root zone soil moisture threshold. This analysis should be extended to other temperate forests to gain a greater understanding of their drought limitations, which can then be used to improve C sequestration processes in global carbon cycling models.