Search

Many upland headwaters of the UK drain areas of blanket peat, much of which has been degraded through atmospheric deposition of pollutants, vegetation change, peat extraction, artificial drainage and erosion. These areas are increasingly the focus of interventions to restore some of the multiple-benefits lost through degradation. Understanding their runoff generation processes underpins analysis of their wider benefits including their potential to mitigate downstream flooding.Using a series of multivariate analysis techniques we examine controls on storm runoff in ten blanket peat catchments of 0.2-3.9 hectares all within 5 km of one another. We find that: 1) rainfall intensity is the dominant hydro-meteorological driver for both magnitude and timing of peak discharge for all ten catchments, with antecedent rainfall only relevant in small storms; 2) most of the inter-catchment variability in discharge predictability from rainfall can be explained by catchment characteristics, particularly catchment area; 3) runoff responses, particularly in small storms, are sensitive to scale even in an apparently homogenous and saturation-excess overland flow dominated peatland landscape; 4) peak discharge in large storms is strongly controlled by attenuation processes associated with the travel time distribution, and thus drainage network geometry; 5) peak discharge in smaller storms underlines the importance of hydrological connectivity at scales Together these results suggest a switching in rainfall-runoff behavior within these catchments where peak discharge is controlled by: catchment storage, connectivity and antecedent conditions in small storms; but runoff attenuation, travel time and thus and network structure and scale in larger storms. In the context of Natural Flood Management, our findings suggest that enhancing depression storage by creating distributed shallow peatland pools in addition to existing restoration methods could raise the threshold storm size below which catchment storage, antecedent conditions and connectivity remain important. However, changes in surface roughness and other measures that target runoff velocities are likely to be more effective in the largest (and thus most flood relevant) storms.

Upland blanket peat is widespread in the headwaters of UK catchments, but much of it has been degraded through atmospheric pollution, vegetation change and erosion. Runoff generation in these headwaters is an important element of downstream flood risk and these areas are increasingly the focus of interventions to restore the peat ecosystem and to potentially mitigate downstream flooding. Here we use a series of multivariate analysis techniques to examine controls on storm runoff behavior within and between ten blanket peat catchments all within 5 km of one another and ranging in size from 0.2-3.9 hectares. We find that: 1) for all 10 catchments, rainfall intensity is the dominant driver for both magnitude and timing of peak discharge, and that total and antecedent rainfall is important for peak discharge only in small storms; 2) there is considerable inter-catchment variability in: runoff coefficient, lag time, peak runoff, and their predictability from rainfall; however, 3) a significant fraction of the inter-catchment variability can be explained by catchment characteristics, particularly catchment area; and 4) catchment controls on peak discharge and runoff coefficient for small storms highlight the importance of storage and connectivity while those for large events suggest that surface flow attenuation dominates. Together these results suggest a switching rainfall-runoff behavior where catchment storage, connectivity and antecedent conditions control small discharge peaks but become increasingly irrelevant for larger storms. Our results suggest that, in the context of Natural Flood Management potential, expanding depression storage (e.g. distributed shallow water pools) in addition to existing restoration methods could increase the range of storms within which connectivity and storage remain important and that for larger storms measures which target surface runoff velocities are likely to be important.

The record of Pleistocene glaciation in the Mediterranean remains incomplete, restricting our understanding of spatial variation in their extent and timing. Digital Elevation Models (DEM) could offer a basis for identifying glacial geomorphology and therefore address this issue. Currently, free DEM availability is limited to global datasets such as the 30m Shuttle Radar Topography Mission (SRTM), which are too low in spatial resolution for such applications. However, Sentinel-1 Interferometric Synthetic Aperture Radar (InSAR) has been investigated to offer a new higher spatial resolution DEM. This study examines the potential of Sentinel-1 InSAR derived DEM by investigating evidence of glaciation at Rumija, Montenegro. Interferometry of Sentinel-1 SAR imagery is used to generate a DEM covering Rumija. Root Mean Square Error (RMSE) indicates a high vertical inaccuracy of 67.5m compared to the SRTM. However, visual comparison shows the Sentinel-1 DEM better preserves topography than SRTM at elevations exceeding 400m. Geomorphological mapping is carried out using several DEM products and supported with satellite imagery. A range of karstic and glacial landforms are identified through the landscape, including a cirque on the north side of Rumija (1596m). Glacier reconstruction with Accumulation Area Ratio (AAR) is based on cirque morphology. A 0.15km² Pleistocene glacier is identified, with a DEM error adjusted Equilibrium Line Altitude (ELA) of 1332m. This correlates to the ELA range of the MIS-12 glaciation at Orjen, 75km north west. The Sentinel-1 InSAR derived DEM is demonstrated to be an effective free data source for the identification of glacial geomorphology.

Comparison against 19 days of national Incident Recording System (IRS) fire statistics for England in spring 2011 showed that MODIS Rapid Response hotspot data missed 92% of significant vegetation fires attended by Fire Services due to a combination of cloud, smoke, small fire size and short fire duration. Despite this, MODIS provides a near-real time, if only very partial, national picture of the geography of large fires in Great Britain, for operational fire managers, which is currently not available from IRS due to delays in collation and spatially aggregated reporting. Fire managers require faster, spatially-specific delivery of IRS data on wildfires (and an agreed definition of ‘wildfire’ using IRS data fields).IRS is limited as a validation source for remote sensing because fire location is recorded as a single point with estimated burned area. Standardising this point and/or recording fire perimeter would assist validation of remote sensing fire products and facilitate spatial analysis of GB’s fire regime.