AbstractRecent climate predictions for the United Kingdom expect a nation-wide shift towards drier and warmer summers, increasing the risk of more frequent and severe drought and fire events. Such shifts in weather patterns have been identified to cause challenges for peatland functioning globally, including for blanket bogs, a globally rare peatland types particularly abundant in Scotland, where they hold up nearly a quarter of the UK’s soil carbon.
This thesis combines the use of ground-based empirical data and UAV-derived cover type estimates to assess the effects of climate extremes on Europe’s largest expanse of blanket bog, the Flow Country. Carbon dioxide (CO2) flux from dominant peatland plant functional types (Sphagnum, graminoids, ericoids) and other cover types (pools and bare peat) were measured and compared across upland and low-lying blanket bog margins and centres, immediately before and during a summer drought in 2018, and over the subsequent year. During that period, most sites acted as net sources of CO2 to the atmosphere. Net Ecosystem Exchange (NEE) was limited by water availability, with ericoid shrubs showing the highest drought resilience, followed by graminoids and Sphagnum mosses.
An exhaustive optimal vegetation index combination method based on ground-based spectral signatures showed that seasonally changing edaphic and phenospectral conditions can decrease classification accuracies between blanket bog cover types, such as Sphagnum spp. and graminoids. Integrating these effects helped achieve high classification accuracies of blanket bog cover using consumer-grade UAV imagery, processed with open-source software. Applying blanket bog cover as proxies to upscale annual NEE showed that during droughts, current emission factors used in the UK for IPCC reporting overestimate carbon uptake in near-natural blanket bogs.
The image processing and classification developed in the thesis was applied to monitor vegetation recovery 3- and 15-months after a large wildfire in 2019. Analysis of UAVs imagery revealed that about 40 % of the surveyed area burnt and that different microtopographic positions were evenly impacted initially. However, after 15 months, there was a faster vegetation recovery in the lower microtopographic positions. Repeat surveys will enable to confirm whether blanket bog vascular plant cover from the burn site will return within 2 or 3 years after the fire, as the models predict.
The thesis demonstrates that consumer-grade UAVs have a place in peatland monitoring and could be used to detect potential shifts in blanket bog vegetation in response to climate change, but also restoration and management intervention. Going forward, co-locating and co-developing monitoring programmes that tie in ground-based measurements, UAVs and Earth Observation will help support future protection of blanket bogs.
|Date of Award||24 Jan 2022|
|Supervisor||Roxane Andersen (Supervisor) & Jason McIlvenny (Supervisor)|