Modelling impacts of agricultural practice on flood peaks in upland catchments: An application of the distributed TOPMODEL

Abstract

Upland agricultural land management activities such as grazing, vegetation burning and bare ground restoration impact hydrological elements of headwater catchments, many of which may be important for downstream flood peaks (e.g. overland flow and soil water storage). However, there is poor understanding of how these management practices affect river flow peaks during high magnitude rainfall events. Using the distributed TOPMODEL, spatial configurations of land management were modelled to predict flood response in an upland catchment which contains different regions operating subsidised agricultural stewardship schemes. Heavy grazing leading to soil compaction and loss of vegetation cover in stewardship regions covering 79.8% of the catchment gave a 42 min earlier flow peak which was 82.2% higher (under a 1-hr 15 mm storm) than the current simulated hydrograph. Light grazing over the same regions of the catchment had much less influence on river flow peaks (18 min earlier and 32.9% increase). Rotational burning (covering 8.8% of the catchment), most of which is located in the headwater areas, increased the peak by 3.2% in the same rainfall event. Vegetation restoration with either Eriophorum or Sphagnum (higher density) in bare areas (5.8%) of the catchment provided a reduction of flood peak (3.9% and 5.2% in the 15 mm storm event); while, the same total area revegetated with Sphagnum in riparian regions delivered a much larger decrease (15.0%) in river flow peaks. We show that changes of vegetation cover in highly-sensitive areas (e.g. near-stream zones) generate large impacts on flood peaks. Thus it is possible to design spatially distributed management systems for upland catchments which This article is protected by copyright. All rights reserved. reduce flood peaks while at the same time ensuring economic viability for upland farmers.

The effect of interactions between rainfall patterns and land-cover change on flood peaks in upland peatlands

“…However, there is a paucity of evidence for evaluating such naturebased solutions for their flood-attenuation performance under a range of rainfall patterns (Dadson et al, 2017;Rogger et al, 2017). For upland blanket peatlands, as overland flow is common, even when peat has been disturbed by drainage (Holden et al, 2006), then potentially one of the most effective ways of delaying streamflow using natural flood management is to create a rough, well-vegetated surface (Ballard et al, 2011;Gao et al, 2016;Gao et al, 2017;Lane and Milledge, 2013). Such rough revegetation most commonly occurs for disturbed peatlands where mosses are encouraged to re-establish in bare areas or to act as an understorey to existing sedge and shrub cover (Parry et al, 2014).…”

“…The modelling studies by Gao et al (2016) and Gao et al (2017) indicated that the same land-cover change in 'sensitive' areas of upland peat catchments (e.g. riparian zones and gentle slope areas) could have three times the impact on river flow peaks as those same land-cover changes in 'insensitive' areas such as headwater regions and steep slopes.…”

Erosion in peatlands: Recent research progress and future directions

“…A spatially-distributed version of TOPMODEL developed by Gao et al (2015) simulated how restoration and the associated land-cover change impact river peak flow. They reported that a catchment with a cover of Eriophorum and Sphagnum had much lower peak flows than that with bare peat (Gao et al, 2015;Gao et al, 2016;Gao et al, 2017).…”

“…Numerous studies have examined the techniques available for restoring degraded blanket peatlands (Armstrong et al, 2009;Crowe et al, 2008;Holden et al, 2008b;Parry et al, 2014), and the role of conservation techniques on stream peak flow (Gao et al, 2015;Gao et al, 2016;Gao et al, 2017;Grayson et al, 2010;Lane and Milledge, 2013), water table and hydrological processes (Allott et al, 2009;Holden et al, 2011;Wilson et al, 2010;Worrall et al, 2007b) and sediment and particulate organic carbon (Holden et al, 2007b;Holden et al, 2008a;Ramchunder et al, 2012;Shuttleworth et al, 2015;Wilson et al, 2011). Restoration practices that result in stabilisation and revegetation are recommended as vegetation cover is capable of reducing erosion by: i) significantly reducing overland flow velocity by 32-70% (Holden et al, 2008a); ii) reducing hydrological connectivity (Gao et al, 2015;Gao et al, 2016;Gao et al, 2017) and sediment connectivity (Evans and Warburton, 2007;Evans et al, 2006); iii) protecting peat surfaces from the effects of rainsplash (Li et al, 2018b), freeze-thaw action and desiccation (Brown et al, 2015;Li et al, 2016b); and iv) enhancing the organic matter and microbiological function of peat. In turn, areas with enhanced peat erosion and good hydrological connectivity would make it more difficult for the peat to host vegetation as seeds or small plants would be readily washed away during rainfall events (Holden, 2005b).…”

Identification of potential conservation practices and hydrologic modeling of the upper Iowa watershed

“…HEC-HMS, a lumper parameter surface water model was used by Drake (2014) to assess the impact of increased infiltration due to soil improvements and land use change, and increased storage from flood control ponds for the 4364 km 2 Upper Cedar watershed in northeast Iowa. TOPMODEL, a distributed physically-based model was used by Gao et al (2017) to simulate the impact to flood peaks of grazing, vegetation burning, and bare ground restoration for the 84.0 km 2 Coverdale catchment of the United Kingdom. HydroGeoSphere, a coupled surface-subsurface distributed model was used by Thomas (2015) to simulate the hydrologic impacts of distributed flood mitigation wetlands, terraces, and drainage tile in the 44 km 2 Bear Creek basin in northeast Iowa.…”