1. Introduction
1.6 Erosion and sediment dynamics under different land cover and land use
Suspended sediment concentration (SSC) data in the streams of undisturbed rain-dominated forested catchments can be divided into two periods. During the storms, discharge (Q) and SSC show wide fluctuations, while SSC are relatively low and steady between the storm periods. In snowmelt-dominated regions, these fluctuations are lower (Thomas 1985).
Besides the flooding periods, forest management activities can lead to the increase of soil loss by water erosion, the complexity of SSC-patterns and the suspended sediment yield (SSY) through addition of woody debris to the stream, triggering mass movements, increasing and/or redirecting surface runoff and modifying channel geometry (Lewis 1998). Figure 1.5 demonstrates the conceptual diagram how the different logging activities influence the stream sediment transport.
Figure 1.5. Conceptual flow chart of the major pathways through which forestry activities affect the stream sediment delivery (from Lewis 1998)
Removing trees reduces evapotranspiration and rainfall interception, leading to increased surface runoff. Soils become vulnerable to surface erosion without vegetation cover.
Uprooting or root decay as well as road cuts may decrease slope stability and cause mass movements. Heavy equipment can compact soils during roadwork, log skidding, construction
32
and use of roads and landings which decreases infiltration, increases and concentrates overland flow. Linear structures alter drainage paths and redirect water to more erodible areas.
The construction of stream crossings and roads, due to the low permeability of the road surface, are one of the primary sources of erosion in forested catchments (Lewis 1998, Chang 2006). logging activities, where the average annual sediment yield (SY) became 1.4-times higher in a 25% patch-cutted creek and 2.8-times higher in a 82% clear-cutted catchment, and the change is obvious compared to an untreated adjacent basin as well. However, the statistically significant changes in SSC are not confirmed in all cases and for all watersheds, because of its large annual variation for a given sampling point in a basin (Brown & Krygier 1971).
Nevertheless, according to a 30-years investigation, Grant & Wolff (1991) also proves the suspended sediment increase after timber harvesting. Moreover, annual differences of sediment delivery are well demonstrated within a catchment and between the treated watersheds. In the “catchment 1”, SSY increased rapidly and remained elevated in the first 10 years after the forestry activities. The average SSY started to decline after 7 years following an exponential trend. In the other treated catchment, where the suspended sediment form are more dominated partly due to the debris flows, the early SSY-increase was even more rapidly, and the highest load was produced by a storm event within several hours after 3 years of the treatment.
Debris slides and associated debris flows are special types of fine material supply in forested basins. On those catchments where debris-flow-prone gullies are connected to the streams, episodic fine material inlet may contribute to the SSC-variability in compliance with the flow threshold and the sediment replenishment-exhaustion phases (Nistor & Church 2005).
Besides of the fine woody debris, the impact of the large woody debris (LWD) on the stream sediment delivery is also a specific characteristic of forested catchments. Natural processes of the forest ecology and forest operations can also add LWD to the streams directly. The sediment retention effect of these debris (including branches, leaves and twigs) and log jams (Figure 1.6) accounts for the changed flow and sediment characters, likewise the alteration of channel morphology such as widening, lateral shifts and downcutting of channels (Lisle &
Napolitano 1998, Nakamura & Swanson 1993). As Megahan (1982) reported, 3.6 obstructions were found per 30 m of channel with 0.8 m3 of sediment storage per obstruction, averaging every years and streams. Although the number of obstructions and stored sediment volumes varied year by year and between streams, data emphasize the importance of channel-sediment storage in small forest streams: channel-sediment amount stored behind the obstructions was an average of 15-times higher than the annual average SY transported to the stream mouth.
Similarly, the role of channel-sediment storage is pointed out in the study of Beschta (1979).
The author wrote that a debris removal activity led to the exhaustion of 3800 t of sediment
33
along 100 m of channel, where the greatest increase in turbidity was measured during the first autumn and winter storms.
Figure 1.6. Sediment retention behind log jam in a forested catchment (Sopron Hills)
Unpaved forest roads and skid trails, which may redirect and accelerate the surface runoff, can be the primary place of soil erosion and may contribute to the stream sediment budget (Figure 1.7). Certainly, produced sediment flux depends on several factors, such as length, slope, soil texture, vegetation cover, road maintenance-disturbance of the road surface and the traffic (Luce & Black 1999). During the experiments of MacDonald et al. (2001), runoff from the road surface started in response to most storms with at least 6 mm of rainfall depth, and the runoff ratio exceeded the runoff from undisturbed hillslopes by least an order of a magnitude.
As for the SY, again the unpaved roads were the dominant sediment sources (annual SY: 8-15 kg·m-2), even though they occupy a very small portion of the catchment area. In contrast, analysed the spatial distribution of fully connected pathways to the stream, Croke et al. (2005) drew the attention to that total sediment yield (TSY) from these pathways are relatively low, and the highest mass of sediment reach the channel at a few locations (e.g. at direct stream crossings and disperse pathways with large contributing area and small available hillslope length).
Notwithstanding, mass movements (Figure 1.7) due to high intensity rainfall events can also provide abundant SY from the steep hillslopes (MacDonald et al. 2001). However, as Beschta (1983) reported, sediment from mass failures were carried downslope only to be deposited on alluvial fans, in topographic depressions or on river terraces in many cases, and thus had little impact on stream SY. This fact also indicates well the complexities of hillslope sediment delivery component, floodplain storage and channel routing which lead to inaccurate SY estimations and increasing sediment variability.
34
Figure 1.7. Erosion processes in forested catchments: soil loss from an unpaved forest road (left) and mass movement (right) (Sopron Hills)
1.6.2 Impact of the land cover alterations on the suspended sediment dynamics
In the case of forest land cover, different land use types can be distinguished which influence differently the fine material availability in the stream channels. Research of Surfleet & Ziemer (1996) confirm that large organic debris input depends on the tree harvest method: a watershed with 60% clear-cut and riparian buffer strips showed slower recruitment from large organic debris than a second basin where 60% of the timber volume had been selectively harvested. Stott et al. (2001) investigated the effects of site-specific harvesting techniques (e.g. machine never drove on bare, unprotected ground, minimized number of stream crossings) to the SSC, and found that SSC-increase show a delay compared to the traditional forestry methods. Although in the case of an agricultural field, but Mosimann et al. (2007) pointed out that soil loss from tractor tracks can be reduced by up to 80% due to intermittent planting. In addition, conservation tillage provided sufficient soil protection up to 5% slope.
Hossain et al. (2002) made a comparison between the suspended sediment exports of three adjacent subcatchments, concluding on clear SSY response to the different ratio of forest coverage. Two of the subcatchments (22% and 42% forest) produced > 93% of the SSY during the study period, while sediment flux from the third subcatchment (75% forest) remained low due to its extensive forest cover and relatively flat terrain. Different land cover and topography induced different SSC-Q relationships as well: clockwise hysteresis loops were obtained in the subcatchments with lower forest coverage, and anti-clockwise hysteresis appeared in the third basin, in the case of studied floods. To criticize this study, the authors would have to separately analyse the effect of forest cover and different topography. The role of different ratio of a land cover type is also emphasized by Bartley et al. (2010), who pointed out that pasture rehabilitation through reduced utilisation and resting led to the increase of ground cover, and thus to the decline of SY where the vegetation cover exceeded 10%. In short term, surface runoff was less sensitive to land use change compared to erosion rate.
Sorriso-Valvo et al. (1995) detected that experimental plots with well-developed grass cover produced negligible runoff and SY, and plot with Eucalyptus forest and 100% litter cover generated rapidly high rate of runoff but virtually no erosion. In contrast, based on a
12-years-35
long experiments performed in three small (< 2 ha) control catchments with different aspect, slope and vegetation but uniform lithology, Iovino & Puglisi (1991, in Sorriso-Valvo et al.
1995) observed highest annual runoff in the grassed catchment and lowest in the undisturbed forested catchments. Highest erosion rate was measured in the logged and undisturbed forested catchment, querying numerous erosion studies which neglect the soil loss in forest regions.
García-Ruiz et al. (2008) gave a good example of different rainfall-runoff and erosion processes at small catchments (<3 km2) with different land cover (dense forest, agricultural land and catchment consisting in part of active badlands). As for the average SSC in the three experimental catchments, the highest value was obtained from the badlands which exceeded 200-times the average SSC in the forested catchment. Similarly to the above cited studies, land cover determined the seasonality and intensity of floods, and the annual volume of Q as well.
Considering the consequences of the changes in land cover, researches related to soil erosion and sediment delivery processes have to be even more stressed due to the expected climate change and forest decline.
36