The aim of this thesis was to improve our understanding of plant – flow interactions and to develop new remote sensing techniques that would allow a marsh scale assessment of flow modification due to the presence of salt marsh vegetation. The limitations of current approaches which improve our understanding and prediction of tidal flows centre around poor assessments of canopy structure and mechanical properties. The validity of such simplified and reductionist assessments of canopy structure were tested and found to contribute considerable error in estimations of canopy frontal area and canopy drag. New metrics to assess canopy structure were tested as part of a flume study using two salt marsh species with varying form and architecture. Results from this experiment found that biomass located immediately below the water surface are important for determining fluid momentum losses in salt marsh canopies. These results led to the development of a new empirical based model using vertical measures of biomass and approach (incident) velocity which can accurately (R2 0.71) predict flow momentum losses. This suggests that the characteristic vegetation parameter and the drag coefficient may be substituted with vertical canopy biomass and an empirical coefficient. This may lead to more accurate assessments of canopy structure and thus comparable results across the literature as well as potentially apriori assignment of parameters in the force drag model. Vertical canopy biomass (3D biomass) was then estimated at the marsh scale using a combined remote sensing approach and an empirical model. Accurate assessments of the marsh surface are critical for hydrodynamic models and important if we are to determine vertical changes in canopy structure. The approach first identified marsh surface returns by operating a moving average smoothing filter on Airborne Laser Scanning (ALS) data. The automated procedure detected vegetated and non–vegetated surfaces using aerial NDVI which calibrated the filter and ensured ALS returns were representative of marsh surface elevation. Using the marsh surface DEM, vegetation was reconstructed at 0.2 m grid cells. Terrestrial Laser Scanning (TLS) was found to accurately quantify maximum canopy height (RMSE 0.14m) whilst a regression model using aerial NDVI and spatial coordinates gave reasonable predictions (RMSE 0.08kg/m2) of total plot canopy biomass within each 0.2 m cell across a ~20,000 m2 area of marsh. Ground measurements found the vertical distribution of canopy biomass followed a power law increase with elevation from the marsh bed. Combining all the approaches allowed the creation of a 3D assessment of canopy biomass with an average error of 30% of the mean amongst plots exhibiting larger canopy biomass ( > 0.4 kg/m2). This vertical measure of biomass can be combined with the flow momentum loss model generated in the flume experiment to assess hydrodynamic canopy drag potential at the marsh scale. Roughness coefficients can also be calculated using this approach which can be easily fed into commercially available numerical flow models.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:725589 |
| Date | January 2017 |
| Creators | Tempest, James Alexander |
| Contributors | Möller, Iris ; Spencer, Tom |
| Publisher | University of Cambridge |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | https://www.repository.cam.ac.uk/handle/1810/267914 |
Page generated in 0.0021 seconds