CTF Europe

Controlled Traffic Farming across Europe

The Impact of Upland Land Management on Flooding: results from an improved pasture hillslope

*This Agri Topic has been made public as an introductory offer*

Flooding in the UK has long been a topical subject. For example, in 2004 abnormal rain intensity caused some of the worst flooding ever observed in Boscastle, Cornwall. In 2007, flooding occurred at the other end of the country, primarily affecting Scotland, East Yorkshire and the Midlands. The flooding even made it into Richard Hammond’s autobiography where he had to abandon his Porsche and run the 16 miles home to see his daughter on her birthday. Both of these devastating floods occurred in the summer. However, winter flooding has been just as prevalent throughout the country and was particularly bad in 2015-16 wen York and other towns were devastated by large volumes of water with nowhere else to go.

The reasons for these floods and others are increasingly associated with farming and land management practices. However, rather than speculate, it is useful to look at particular studies that attempt to quantify the impacts that farming and/or land management practices have on water movement so that realistic, and potentially minor, changes could be implemented to alleviate flooding downstream.

The study

A study by Marshall et al. (2009) aimed to do exactly this – to establish the impact of different land management practices on runoff. They completed their study over three scales, plot (~100 m2), hillslope (~0.1 km2) and catchment (~10 km2) at one site in Pontbren, Wales. The hillslope was located on artificially-drained, agriculturally improved pasture. In their desk survey, they found that over 50% of the uplands in the UK are grassland, rising to 70% in Wales with primarily sheep production. They also found that, due to government incentives to increase production, field sizes have increased, hedgerows removed and fields drained. Over 40 years, sheep numbers have increased by 104% and so studies into their impact soil compaction and the impact on soil physical properties have increased (e.g. Rauzi and Smith, 1972; Langlands and Bennett, 1973; Gifford and Hawkins, 1978; Willatt and Pullar, 1984; Greenwood et al., 1997; Nguyen et al., 1998; Drewry and Paton, 2005). However, throughout these studies it has been difficult to isolate land use change from the impacts of climate change and catchment scale variability (Sullivan et al. 2004; Orr and Carling, 2006). Rather than make assumptions that land management is only to blame for increased flooding downstream in fact, we should consider flood alleviation as one of the greatest challenges for farming at the moment.

How they did it

Marshall et al. (2009) located their experiment in the Pontbren, a headwater catchment of the Upper Severn in mid-Wales. This area is particularly well known, at the time of writing in 2017, due to a farmer-driven initiative to address the flooding occurring in their catchment. The Pontbren is largely affected by warm, moist conditions and three yearly rainfall was measured as 1501 mm, relatively low compared to the highlands of Scotland (another "hilly" area) that regularly see this annually, if not higher. Due to this low rainfall, the rivers in Pontbren are largely perennial, i.e. they dry up over the summer. Despite this, farmers are cited as observing increases in channel erosion and overland flow since the 1960s. The Pontbren catchment is largely clay-rich with soils that remain water-logged for long periods of time. As a consequence, soil quality has seen structural degradation at varying scales across the catchment.

The hillslope experiment, as shown in Figure 1, aimed to quantify the overland flow (OLF) and drainflow from the land in two distinct areas – under a tree shelter belt and the open area bowl. Monitoring of the flow was by v-notch weirs and tipping buckets for high and low flow respectively.

Figure 1 Aerial photograph and diagram of experimental design of the instrumented hillslope (adapted from Marshall et al. 2009)

Along with monitoring OLF and drainflow, tensiometers were installed across the site (denoted Hn in Figure 1) to measure the soil water potential every 10 minutes at three depths (10 cm, 30 cm, and 50 cm). Groundwater levels were also measured at nearby boreholes, denoted by a grey triangle in Figure 1.

The findings

Measuring soil water potential through the soil profile showed that the soil remained largely saturated throughout the winter due to low intensity but frequent rain. The study also identified the importance of macro-pores that create “preferential pathways”, which explained some of the anomalies in the study, i.e. that water was able to move laterally, by-passing much of the soil profile. The important message was that the permeability of the soil A horizon was strongly influenced by the saturation of the B horizon. In addition, Marshall et al. (2009) found that the dominant runoff process on the hillslope was drainflow – which makes sense as that is their purpose. However, the efficacy of the drains is also largely dependent on the presence of the macropore structure (hence the backfill of gravel in most drainage systems).

When considering a rain event, OLF was found to be the predominant waterflow pathway but was largely influenced by the antecedent soil moisture conditions. Figure 2 shows how the rainfall event in Figure 2c did not incur such a large flow peak in OLF as Figure 2a, despite the similar intensity of rainfall. This may be intuitive to agriculturalists who work with their land every day, but having a study that shows this specifically, reinforces the need to improve our soil structure and our natural drainage systems. As the rainfall event shown in Figure 2a occurred in January and the one in 2c was in July, antecedent soil moisture conditions are likely to have played a large part in this variation.

Figure 2 Bowl runoff data and corresponding pore water pressure ψ from tensiometer array H12

Considering soil moisture then, the study also asked whether trees can reduce local surface runoff. It found that the loss of soil water was greater under trees compared to the A horizon under improved pasture. This, essentially, is a consequence of the type of soil where clays located in the pasture have a stronger holding capacity of water and release water slowly or not at all. The largest difference, however, was at soil saturation where it took more water to saturate the soil under trees than under improved pasture. So while it appears that water is lost from soil under trees more easily, it is important to account for where the water goes (i.e. taken up by plants). Interestingly, there was little difference in soil water availability in the B horizon under both types of land use. The important message from this part of the experiment, according to Marshall et al. (2009), is that there is likely to be greater storage of water in the B horizon under trees which helps to reduce the loss of water causing flooding in local rivers and downstream.

And finally

The main outcome of this paper, however, is that it highlights the need for further similar work. This is a single site, in a particular area of the UK and so results are likely to differ across the British Isles (including Ireland) dependent on soil type, land use and topography.

*This is an introductory piece to the larger review of upland farming and flooding which is due in Sept 2017.


Drewry, J.J. and Paton, R.J. 2005. Effects of sheep treading on soil physical properties and pasture yield of newly sown pastures. New Zealand Journal of Agricultural Research, 48 (1),  Mar, pp.39-46.

Greenwood KL, MacLeod DA, Hutchinson KJ. 1997. Long-term stocking rate effects on soil physical properties. Australian Journal of Experimental Agriculture 37: 413–419.

Gifford GF, Hawkins RH. 1978. Hydrologic impact of grazing infiltration: A critical review. Water Resources Research 14: 305–313.

Langlands JP, Bennett IL. 1973. Stocking density and pastoral production: I. Changes in the soil and vegetation of a sown pasture grazed by sheep at different stocking rates. Journal of Agricultural Sciences 81: 193–204.

Marshall, M.R., Francis, O.J., Frogbrook, Z.L., Jackson, B.M., McIntyre, N., Reynolds, B., Solloway, I., Wheater, H.S., and Chell, J. 2009. The impact of upland land management on flooding: results from an improved pasture hillslope. Hydrological Processes, 23 (3): 464-475.

Nguyen ML, Sheath GW, Smith CM, Cooper AD. 1998. Impact of cattle treading on hill land 2. Soil physical properties and contaminant runoff. New Zealand Journal of Agricultural Research 41: 279–290.

Rauzi F, Smith FM. 1972. Infiltration Rates: Three soils with three grazing levels in northeastern Colorado. Journal of Range Management 26: 126–129.

Willatt ST, Pullar DM. 1984. Changes in soil physical properties under grazed pastures. Australian Journal of Soil Research 22: 343–348.

UK: Tim Chamen
+44 7714 206 048

Netherlands: Sander Bernaerts

Powered by Wild Apricot Membership Software